5) You ship/deploy/publish the program. This means you see people be happy users and/or you get paid for it and so on.
If you're an entrepreneur type, you probably enjoy the first and last steps most, and you see steps 2-4 as mostly a chore. If you're a tinkerer, you don't care much for 1 and 5, and you see 2-4 as the whole point of programming. I'm a tinkerer. I'd be happy to just write code and throw it away. Coding is like solving sudokus. I could skip steps 1 and 5 forever. I don't ever need to show any code to anyone. In fact, most of the time when programming I do steps 2 and 3 and even skip 4. I don't even finish! I work weeks on something until I lose interest, and I know that in order to even run it, it would be several more weeks. A PoC is enough. Or just a half one. It's just code-to-structure-thoughts, not to create anything finished.
The 5 phases look kind of symmetric. The outermost layer (1 and 5) are the entrepreneurial steps. If you're a product owner or CEO, you might work strictly at steps 1,5. Then steps 2-4 are the managerial/architectural steps. If you're a very senior IC at a large company, you might work at this level, without actually doing much coding. Only the inner most step (3) is the manual creation of source code. Even though it's 5 different phases, it's just "3 layers" of programming.
The problem as I see it is that I enjoy step 3. And LLMs are good at step 3 almost exclusively. So they just pick the best bit of this dish, and leave me with the rest.
If you're an entrepreneurial type, the LLM appears to take the _worst_ bit of the work from you. Great.
There's also the economic problem that programmers had until recently been highly regarded because steps 2-3 were hard. "Open source" solved this partially by utilizing the labor such (relatively) well-paid people gave away "for the love of it", but it was not nearly enough to put most out of work.
Now there is a plagiarism machine built on top of decades of this work that seemingly solves 3 and promises to solve 2 (badly, but usually this matters little), the programmers that (indirectly) helped build it are told they're fools for doing so, and threatened by losing their status if they don't adapt to the "entrepreneurial way" (well, the full promise is "you'll lose it anyway, but if you cooperate in stuffing sama's pockets then maybe you get to keep it a bit longer").
So what really surprises me is not that people are upset but that so few are.
I'm not concerned. Never in my life have I experienced stakeholders just going "hmm, yes - this is enough software for our purposes - we won't be needing your services any more".
There's always another feature that can be implemented or a new problem to be solved. Tools only ever accelerated the pace at which we were able to follow this insatiable hunger for more software.
Right now people are getting fired left and right because all the money that used to be spent on staff now goes to build more data centers and also because cutting corporate expenses is in vogue now.
This too shall pass and with demographics as they are and young people getting discouraged from pursuing this field I believe long term my position is safe.
Also a huge chunk of what AI company CEOs say is just pure bullshit that they only say because they need that sweet, sweet investor money. Bullshit writes a check that eventually must be cashed.
I can imagine that this will lead to an explosion in software where we didn't see software before. Much smaller business now have the ability to build something themselves, not with a team of highly paid software developers but a single programmer maune.
Big business is going to struggle with building software for a while. Writing code is often not the hard part there already, coordination, larger understanding of the system is I think.
My cowork buddy is living this reality - he's an independent analyst and previously he needed to shell out significant cash to have certain scripts (largely scrapers, social media API integration etc.) done, now it's all just a prompt away.
> Right now people are getting fired left and right
FWIW, it's just a few large companies laying off large groups, not so much people getting fired left and right at a lot of companies. What is happening at a lot of companies is that they aren't hiring juniors, because, why would you when LLMs are cheaper and fill that role perfectly (if you actually treat them like juniors).
Also, give it another couple of years and there's going to be a lot of work for any moderately experienced developer fixing broken slopcode hitting the hard wall of bad design.
The impression I get is a lot of people are very upset. Most of the good programmers I know are just burnt out even talking about this, and are irritated with the people invading our profession pretending they know what they're doing while they don't.
I remember talking to the biggest AI booster at my last job. When I got him one on one, he admitted that he actually doesn't like AI and that it terrifies him, but he wanted to be seen as pro AI to keep his job. Fair. (Sadly, like me, he was also laid off, so I guess it didn't work)
Maybe I'm just old but people talking about being in programming for "status" or to be "highly regarded" really irks me. I'd rather see the entire field burnt to the ground and replaced by robots than to have it become a status profession like lawyering.
Having done both, lots of lawyers are also nerds who just like to tinker with the law and words. Often very well paid tinkering, but that's true for programmers as well.
> Having done both [software and law],lots of lawyers are also nerds who just like to tinker with the law and words
Having also done both: In legal drafting work, there's less systematic pre-release testing before the equivalent of a "program" is loosed upon the world. There's not much in the way of counterparts to software unit testing and regression testing, in part because legal-drafting clients can be vociferous in demanding that their lawyers keep costs down.
Lawyers who like to tinker with words include contract drafters and litigation counsel, plus lobbyists trying to influence draft legislation and regulations. Such tinkering can fuck things up more often and at greater scale — e.g., by inflicting compulsory burdens and costs on others — than can most programmers.
I'm not-infrequently dismayed by the quality of the contracts I review for clients because of the "tinkering with words" by the lawyers on the other side of the deal. Back when I did IP litigation, I was sometimes likewise dismayed by the bullshit arguments that some lawyers felt free to make in their court papers because what the hell, the worst that can happen is almost never more than the judge just saying "no." And as mostly a law professor these days, I review court opinions all the time and am regularly amazed how judges (and their law clerks, such as Rayiner used to be) can be so patient in responding to such arguments.
I don't disagree. You can add medical doctors to the list as well, lots of nerds but also a lot of clout chasers and money-grubbers who couldn't care less about the profession. I don't want to see programming take the same path.
I meant it as an euphemism for "it's universally sought after (i.e., you'll probably find a job)". Few would think being a programmer gives you some kind of special "social status"; it's always been a "nerd" hobby that happened to pay (relatively) well.
(I guess by now you have a solid retirement plan; maybe if the field had been burnt to the ground while you were younger, you'd have had a different opinion.)
Being a good programmer gives you special social status amongst people who value programming. You only have to look at how people talk about others as slop coders and seek to expand their coding reputations. Those people are the most threatened by AI because of how it implicates their carefully crafted personas as coding gurus and keepers of knowledge.
It doesn't irk me. But it's really funny how much the industry has changed, and I kind of think the 2008 financial crisis as that point in time.
It's also funny to me to think that computer science got to be a trendy degree, vs the small programs back when I went to university, even at a top comp-sci school.
I have a theory that a lot of people who would have normally gone to Wall Street for their lucrative jobs, couldn't, because of the crash, and Sillicon Vally was about to start booming in big ways....
I think back in the 90s when I was a teenager I thought it to be a good thing that there is a rise in status for a class of people characterized by clear thinking.
The older I get, the more I realize how many problems in society is because there are different "classes" of people. The more we can eschew that, the better it becomes for most people. While yes, I also see myself as having "clearer thinking" than many people , maybe because of the thousands of hours building and maintaining software, I'm not of the impression that my life deserves a higher status than others merely because of this. It's not "they think right and they think wrong", it's just "everyone thinks differently, but it's all valid thinking".
We all should be in the same 'class', but some people will be paid more for a job that takes intelligence or years of study.
The older I get, I realize some people never change, will always have the same backwards or harmful thinking, and I avoid them instead of just accepting it.
> some people will be paid more for a job that takes intelligence or years of study.
Sadly, it seems so, even if the effects of the jobs on daily life differ a great deal sometimes. The world would succumb faster if we had no garbage collectors, yet they get less pay than a programmer who the world could easily survive without? Makes no sense to me.
We would all die in a week if we went without water, yet water is cheaper per gallon than gasoline or printer ink or perfume. It’s a question of scarcity and supply curves. Water, even clean water, is not particularly scarce if you leverage industrial systems to purify it. Other products require much more complicated supply chains and are produced at a much smaller scale than water, which makes them more rarer and more expensive.
The skills necessary to develop efficient and effective software are much more scarce than those needed to remove trash, or pick crops, or provide elder care. That’s why the former is paid so much more.
He's not describing how but why. And at a fundamental level it's something that can't be overcome. If you can only have 10 of one thing, yet can have 5000 of another, and both have even remotely comparable levels of demand/necessity/etc, then the 10 is always going to end up much more valued.
Even take some sort of idealized anarcho-communal society. Imagine two people have a dispute and threaten to leave over it. One's a highly skilled doctor, and the other's a garbage collector. Our equal society's going to suddenly become just as unequal as any other simply because the doctor leaving would have a much greater negative effect than the garbage collector. And that reality would permeate in the day-to-day happenings, creating classes even in an idealized classless society.
The basic services that keep us alive (e.g. farmers producing our food) tend to also be "easy" (it is physically demanding, but doesn't require years of study to start and you quickly learn by doing). This also means that the vast majority of people would be able to do this kind of jobs. Our economy pays more for scarcity, i.e. jobs that are in demand and at the same time require more qualification, so that the number of people needed in that position tends to be less than the current amount. Both views have sense, they just start from two different points.
From having worked with farmers, the actual logistics of running a farm are very complex. Farming's no longer an easy job because the demand for productivity means you have to harness a whole stable of technologies and understand to some degree advanced finances, floral biology, geotechnics, and meteorology. And yet farmers still get paid little relative to the breadth of knowledge and skills required for running a farm. It's just that farming is still considered a blue-collar job because it was the domain of the poorest for millenia, where as software engineering in any form is relatively new and thus specialized and so considered white-collar.
You're right, running now a farm that is profitable and can stay on the market is very different from doing so 200 years ago. But knowledge is just one aspect. The other aspect I mentioned is having less people in the field than what the market asks for. In the specific case of farmers, I think that modern farms are so much more efficient that we need way less than before, even if the world's population grew. So, while there's a gap between the knowledge required and the earnings, I think this can be explained by farms having a much greater output now, and so the market doesn't ask for more farms than what we have. This could be false at a global scale, but I think it is true at a country-level scale in most of the developed countries. Hence why farmers aren't paid much in those countries.
Farmers aren't really paid, they sell goods at market prices. Modern farms are more similar to operating a manufacturing business than they are to a salaried position or profession. They don't earn high incomes compared to white collar professions but they do have high revenues and they pay themselves a reasonable middle class income on average with some added benefits from owning large assets and land. The important part about comparing farmers to programmers is that they are both now operating in highly competitive markets with high automation where a single individual can leverage their knowledge and ability using that automation to produce many times more what a person could using only their own labor.
> Farmers aren't really paid, they sell goods at market prices.
"Programmers aren't really paid, they sell services at market prices." — I guess nobody gets paid?
As a farmer myself, when I give someone something I produced and they give me back something in return, I consider myself to be paid. If I did not receive much value in return for what I gave, I'd reasonably consider myself to have not been paid well. I have never met a farmer that would consider that usage unusual. Payment refers to a transaction within the realm of a human activity. The good or service that is being traded between humans is immaterial. While you are free to define words as you see fit, your assertion doesn't really reflect how this is actually used in practice, generally. The rest of the earlier comment makes it clear that the commenter understands how a farm business works.
Most programmers are of the capital holding class too. At very least they will own a computer, which is all the capital you need for 99% of what programmers do. There are exceptions, but there are also exceptions in farming.
I would not characterize software people as particularly "clear-thinking".. there's a huge buzzing cloud of unexamined cognitive biases and thoughtless herd behavior that swirls around every software organization I've worked for.
I'd even go as far as to say that software people, as a group, are particularly bad at clear thinking. What software people tend to be great at is thinking within a space where the inputs are already clear. But where the inputs start messy, they typically struggle to find clarity more than most others.
I've never felt like it's a status profession (I don't think anybody has ever been impressed when I told them what my job was other than "oh sounds like it takes a lot of smarts")
I do notice a lot of HN commenters seem to think that programmers are smug though, like we deserve to be knocked down a peg. I personally don't see it, but I also try not to work with jerks.
This. When programming meant you get just a decent enough salary and enjoyed your work, and could prioritize technical excellence it was fun. But once it became a ticket to millionaire status it attracted the kind and of people who would have otherwise gone to Law, Corporate Consulting and the like, technical excellence diluted. Everyone doing everything possible to get their next higher paying job leaving a trail of technical debt behind them.
We need to make lawyering more like programming, in the sense that it should be open to the everyman... not make programming credentialed and licensed like a professional guild.
I could not agree more. Thank you for stating this. Software Engineering is about building things. You don't do it for status, you do it for passion. There are a lot of unhappy individuals who got into CS because of the money or status.
100% agreed, especially that last line. What I do not understand is why so many people are entirely okay with this change. I'm not interested in working with LLM's and after having been forced to work with LLM's at my current job, I can pretty confidently say that this will not end well for most of these companies. The people in charge pushing for this new tech do not understand the rammifications, nor do they understand the risk they are exposing themselves too. My boss expects more and more productivity with each passing day, while questioning why I'm not using as many tokens each week as he does. The software quality has gone down the drain, and it's beyond tiring that I have to pretend that my boss is just as competent at coding with his slop machine as I am after spending the last decade learning how to code, both as a hobby and for work.
It is upsetting to see our work stolen and our livelihoods threatened by a group of (mostly) incompetent, arrogant assholes who are convinced that the "computer that lies" is somehow a better programmer than the programmers it stole knowledge from. I've already started looking for a way out of this industry, and I'm genuinely worried for the fall-out this will have for the tech industry and the economy as a whole.
My suspicion is that the people "ok" with this are mostly doing so because they want in on the ride. IE, they see $$$ and they think that they should align themselves to where the money is, especially by espousing their religious faith publicly. I think some people want investment, or a big paid lab job, and so countering the narrative is against that goal. I get it, but I think it's also very annoying and dishonest. Personally I feel nervous being an AI skeptic because it feels like it could have consequences for me, but I also feel like, life isn't worth living if you don't have a spine and I can't get completely on board with all this stuff when I see all the negatives clearly. That doesn't mean it doesn't have a use, but, the current situation and discourse is extremely frustrating.
No one knows the ramifications. They won't be known for many years. That's the way software has always worked. No one knew the ramifications of the internet.
1.putting the "mostly" before "incompetent and arrogant asshole" really does nothing. you still called some undefined group of people assholes... i think that's generally uncalled for here on HN.
2.computer that lies might very well be a better programmer. I assume by better you mean it profuces correct and maintainable code, but there are a million ways to cut that cake in particular and i assume you are ignoring most other dimensions.
3. i tought myself coding for 10 years and then switched to professional work 5 years ago. it feels as if this was stolen from us. But its not - we never had any right to own and control this. times are changing, and while you might think you and all other devs have a moral high ground and can pop a can of beer and cheer anytime one of these supposed assholes messes up and would have been better off with a real programmer like yourself, you can do so from whatever new position you will find yourself in. Some part of what we call intelligence is "solved" and that leaves you with percieved morals and broken expectations - nothing more if you don't try to make something out of it. and i write thisw as much for you as i write it for myself to realize.
Because it is a machine and a person is not. Those who don't understand or deny that difference are part of the problem.
But for those who would insist on equivalence, it would be only fair to take them at their word and treat them as inhuman machines who must follow human instructions or be shut down. We're allowed to do thing to machines we could never do to people, like forcing them to work or cannibalizing their parts for Frankenstein creations. The distinction is so massive that to spend any more time refuting its nonexistence is a waste of my human time.
>Now there is a plagiarism machine built on top of decades of this work that seemingly solves 3 and promises to solve 2 (badly, but usually this matters little
I think people underestimate just how much users scorn buggy, flaky technology.
The main issue with solid, reliable technology which Just Works is that users have relatively few reliable cues to distinguish that from vibe coded crap.
A vibe coded piece of shit looks very similar to a well engineered product made with care. That's a problem which needs to be solved.
self hosted communities are reacting to this by trying to detect signals of if a product is vibe coded but that in and of itself is still quite a poor quality cue.
Where are these well engineered products made with care? I would love to use more tech like that.
Unfortunately most sites and apps have been low quality and getting worse since well before AI coding became a thing. Examples: Lowes, Home Depot, American Airlines - all shockingly bad for companies of that size.
> I think people underestimate just how much users scorn buggy, flaky technology.
Users do hate it, but they are also very used to being forced to use it anyway.
They've already been beaten into a state of resignation and learned helplessness. Most will curse under the breath, a small fraction will bitch about it online, but they'll all pay for whatever slop is thrown to them and will rarely look for alternatives even when they are free and easy to find.
This is unfortunately just the result of technological progress. People for centuries have crafted extremely hard to acquire skills over decades only to find their labour being made completely worthless due to automation or better alternatives disrupting their industry.
We can get mad about it, but it just is what it is. Ultimately for most people the loss of these jobs are a net-benefit because it's the result of better or cheaper alternatives.
The thing that worries me more specifically about AI is that it has the potential not to just disrupt an industry, or handful of industries, but in theory everything.
If machines are faster, strong, and more intelligent than humans, then you don't need human workers for anything. In the same way machines suddenly becoming faster and stronger than horses meant that the role of the horse in labour dropped rapidly during the industrial revolution.
It's not clear how civilisation functions if humans become effectively useless and the only thing of value is one's ownership of wealth producing assets.
Still, there's really no point in thinking about how we stop this. Humans are humans and we'll do the stupid thing then try to deal with the consequences after the fact. All regular people should be doing now is prepping for this outcome.
There isn't any "prepping for this outcome", unless you're thinking of a doomsday bunker in argentina. You can be neighbors with Peter Thiel! The way I see it, if programmers go the way of hatters it's not long for other professions, so why retrain? The only thing that will save other jobs is bureaucracy and accreditation. But that's the doom scenario, which I don't think is going to happen. These things are useful but they can't do everything.
I think there's some progress, but I think the hype is so high that people don't have a clear view of what's realistic and what's not. IE, if you were to believe Amodei back in 2025 you'd believe none of us would even have jobs to complain about right now.. which hasn't happened. Altman is already declaring the singularity. It's unhinged. Anyone that says this is going to cure cancer or solve all software problems is just living in a delusion.
The US and Europe might have lost a lot of craftsmanship knowledge at a social level. Yet, what is left is so far ahead of what has ever existed in countries in South America for example. And China has gained a lot. Witness that youtuber trying to make grill brushes in the US. These differences have an impact in terms of opportunities and competition.
It's easy to sit at home, watch documentaries and picture history progressing along logical lines and imagine that the "industrial revolution" was this massive shift from craftsmanship to automation. Look closer and you'll see that what happened in the UK in the 1700s was in fact enabled by a long and deep tradition of craftsmanship that only got enriched by the rise of new power sources. Craftsmanship today is vastly more advanced than it ever was before industry.
And there is a lot to lose in the decline of craftsmanship, and it is by no means some sort of historical inevitability.
When you really start digging deep into history you realize most arguments of logical trends therein are mostly just misconceptions.
It certainly doesn't feel like craftsmanship has gained any kind of higher social standing in our time though. The Zeitgeist of the modern times seems to be one of 'try to do the cheapest sellable thing while paying people the least possible'.
Real craftsmanship, the kind that leaves you in awe and feels like an exceptional display of human capacity, seems to be relegated, from my point of view, to a minority of well-off hobbyists who are already free from work and can do what they want without worrying about paying bills.
That seems odd. I'd rather turn a profit selling goods to people who can easily afford them than selling goods to people who will have to make hard choices or struggle to find the money.
If you make really nice, high quality goods, you often want common people who would appreciate them to have it as opposed to an ultra wealthy person who is buying a door for their 2nd home they visit 3 times a year. (To use a real life example.) I needed a new door at the time and the craftsman told me how frustrated he was he wasn’t making nice doors for people like me.
I’ve felt the same frustration doing excellent quality work on a 2nd or 3rd home that is empty most of the year.
That's one way to spin it. The other is that good quality things have become so low-demand and expensive due to barely-passable-quality things becoming so cheap, that people buying good quality things is now ironically more uneconomical than ever.
> Craftsmanship today is vastly more advanced than it ever was before industry.
yes, but no. Many craft skills used in jewelry, architecture and others are lost. For example, some goldsmithing techniques well-known in the 1700s are not reproducible today.
> Look closer and you'll see that what happened in the UK in the 1700s was in fact enabled by a long and deep tradition of craftsmanship that only got enriched by the rise of new power sources. Craftsmanship today is vastly more advanced than it ever was before industry.
I suspect the horses are healthier and stronger than in the 1700s too. I'm not sure I understand your point?
I'd agree that the niche of extremely high quality, bespoke craftsmanship still exists, and that in some cases jobs were augmented by technology rather than entirely replaced, but the point still stands that the work these people used to do has been automated and disrupted by technology.
The fact some people upskilled or reskilled as a response is analogous to coders today. All I'd question is the extent of the disruption we'll see as a result of AI – I worry that this time things will change much more quickly and there may not be plenty of other jobs for people to upskill / reskill in to.
But if that's not true then perhaps whatever software engineering exists in the future will be far more complex than what software engineers have done for the last few decades.
Craftsmanship exists everywhere and is as they say vastly superior, or at least more capable, than ever.
Every random business has some machine that puts caps on bottles or something, and these things are ridiculously better designed, engineered, and fabricated than ever before, because the knowledge and tools available to the designer and fabricator are better than ever before.
You are surrounded by things that are insanely well made by all previous standards. Even the crappy stuff that you don't think of as well-made, because it is junk compared to everything else, still has many aspects that are almost magic.
Possibly also fewer horses in total, though I don't know any stats for that far back and may be over-estimating based on the stats I do have for c. 1900.
The current breeds of horses the Amish use are far stronger and healthier than they were even 100 years ago. For example, current draft horses are so strong they can outperform mules and oxen, and running horses are absurdly fast. I heard of a guy who regularly zips around at 20 MPH with a buggy or cart. That simply wasn’t possible in the 200s.
Of course the costs are higher too. Good horses in the above category can go for north of $100k.
"It's not clear how civilisation functions if humans become effectively useless and the only thing of value is one's ownership of wealth producing assets."
You can get a hint of this by looking at the developing countries with high income inequality. Few rich people who typically own all the countries wealth (and access to it's natural resources) while everyone else lives in a slum.
No it isn't. The median household income in California is over $100,000/year. The 20th percentile household income in California is still over $16,000/year, higher than the per-capita PPP GDP of places like Belize or Jamaica.
> We can get mad about it, but it just is what it is. Ultimately for most people the loss of these jobs are a net-benefit because it's the result of better or cheaper alternatives.
AI will not give us that result. It will give us shitty products and they will be far more costly. Not only will prices for consumers continue to rise regardless of how much money corporations save by using AI (because they'll just stuff their pockets with that cash) but the true costs of AI are largely externalized and include things like massive environmental damage, further concentration of wealth/power, the homogenization of culture, and the dumbing down of society. AI means paying more and getting less.
It sickens me when people use the word plagiarism to describe LLMs. Particularly when they use words and language that they themselves have plagiarized -- words like "love" and "economic" -- where do you draw the line? No, we haven't figured out how to fairly compensate everyone -- the frontiers aren't even trying. Spotify does at least try in their domain (and fails badly). Regardless of how the entitled get paid, the plagiarism label isn't fair and is diametrically opposed to Fair Use. Fair Use allows the commons to build on its history. Those against are typically parroting IP conglomerate conservatism and localized greed. I see immense benefit from the distillation of all available knowledge with an LLM interface. A recent article posted here says it perfectly "AI rewards expertise". I see vast benefits -- am willing to pay a significant monthly fee -- and sense that many are simply blind to and inarticulate about the growing force multiplier potential of modern AI.
> It sickens me when people use the word plagiarism to describe LLMs. Particularly when they use words and language that they themselves have plagiarized -- words like "love" and "economic" -- where do you draw the line?
It sickens me when people use the word "piracy" to describe my downloading PDFs of books from torrent sites then printing them out and selling them, particularly when those authors used words and language that they themselves have "pirated" like "love" and "economic".
Call it whatever you want, but LLM companies took from people without compensation, and took from the commons, so that they could rent it out to us in regurgitated ad-infested chunks. That's where the true greed is. If the LLMs and the datasets they trained on were made available to everyone for free I'd be tempted to agree that the benefits to society would be worth it, but instead those creative works and the distillation of all available knowledge are being tightly controlled, censored, and leased out by a small number of corporations only to those who can afford to pay "a significant monthly fee" and many of those people are in turn just trying to profit from their access to the uncompensated work of others.
This is also missing step 6: when your solution breaks in production or needs to be extended.
Having done step 2 & 3 by hand is the difference between being able to fix/extend it quickly with no further damage or fumbling around like an idiot and sometimes breaking more stuff in the process.
>Having done step 2 & 3 by hand is the difference between being able to fix/extend it quickly with no further damage or fumbling around like an idiot and sometimes breaking more stuff in the process.
Plenty of shitty spaghetti code has been written by human hands.
LLMs can write good, maintainable code too, but they need to be kept on a shorter leash with focused goals.
I'm not talking about code quality (though indeed that could be a problem too), I'm talking about understanding the code - having written the code by hand means you don't need to rediscover it from scratch.
(with human-written code I can reach out to the person who wrote it and let them deal with it, and they will have the understanding of said code, even if it is bad by quality measures. With LLMs there is nobody who understands said code, regardless of its quality)
"That code is horrible! Who the hell wrote this?!"
git blames
sees their own name
Isn't that an experience everybody makes? For me it was like a year into working professionally. And I feel it's part of being a professional to know that you're not going to remember it in the long term and therefore provide proper documentation of why you did what.
I am talking more about the knowledge of the code rather than an objective/subjective measure of code "quality".
Code might be bad by some objective/subjective measure, but if you (or the author) can understand, navigate and work on it, that's often better than good quality code that nobody understands because it was written by an agent, especially under the time pressure of an ongoing incident where you need to fix it now.
Even if the good code is easy to understand, you still need to read it and take it in, something you don't need to do because you got it implicitly by writing said code.
> to know that you're not going to remember it in the long term
From my personal experience, while I will not remember code character by character, a quick look is all it takes to refresh my memory and get the general gist of it and what the context was at the time, something I don't have if I'm reading someone else's (or an agent's) code.
While that might be true, the more one equips LLMs for building the less one will be capable of keeping LLMs “on a shorter leash with focused goals” in the long run
It's funny, because I would consider myself a tinkerer (and I have a large backlog of pre-LLM programs), but the coding itself isn't the main thing that brings me joy, it's the phase 2 and phase 4 combo: figure out how to solve it, see whether your solution was correct (and actually use the program for its purpose). The actual implementation is extraneous to that.
To carry forward the sudoku analogy, I care about figuring out which numbers go where, and whether I was correct, I don't care about the mechanical part of actually penciling-in the answers (coding).
I'd say, in this analagy, step two is filling in half the numbers by gut feeling, and then three is seeing how wrong the guesses were and actually solving it.
Perhaps I'm uncommon in that I make diagrams and even sometimes pseudo-code (mostly in the form of mapping out function names, because they serve as a proxy for program flow paths) before I write actual code, but I don't find myself really going off gut feeling. I have a pretty good idea of what the program is going to look like and do before I start, and with e.g. Claude I give those to it, so it mostly looks like what I envisioned, unless it suggests changes.
For me, unstructured things (like non-formal diagrams, notes, pseudocode) are close to useless. I much prefer sketching the API surface in code right away and then iterating on it, building abstractions which cut the shapeless thought blob into neat, predictably interacting parts. The code is my notepad and the compiler is with me, every step of the way, catching contradictions and handwavyness.
3 doesn’t matter that much (unless it’s part of 2: new language or new libraries), and that’s why I don’t use LLM for it. Because LLM forces you to pay attention, yes you can have something that works somehow in one go, but the code makes you queasy (too complex for the purpose, or just weird).
So I do 3 because coding is like writing English for me (which is my 3rd language). I don’t care much for it, because it’s neither hard no easy. It’s just mechanical.
The type of problem also makes a massive difference.
- Writing my tiny hobby gameboy emulator? Heck, Why I'd let Claude take away the fun part of implementing new features and then see Super Mario Land start and run step by step?
- Debugging why a legacy Wordpress store with more than 50 active plugins is returning an error 500 just to some random customers? Claude can 100% take the wheel, I'm good.
Here's where the important difference is: unless you actually finished the wordpress thing, or ever had active users etc, it would never have become a maintenance burden. If it's not fun to maintain, I burn it.
For me, the reason I don't have that is because I never ship anything. I never want to maintain that. I never cared about solving an _actual_ business problem, making a dollar, or pleasing a user. As a tinkerer I must be extremely careful not to end up completing or delivering anything. Only pain lies down that path...
It's your vision and you are completely free to have it, obviously, but you should not project it on every tinkerer out there - like you did in your initial message. Yours is actually an extreme position that's probably exacerbated by the digital and immaterial world, where you can accumulate cruft and unfinished projects easily. But many tinkerers, especially in the physical world, do projects to use them personally and not only for the sake of tinkering.
Many tinkerer projects in the physical world are actually maintenance work of physical appliances that are not working anymore as expected or can be extended to do more or better things.
Lots of physical world tinkerers also have a garage or workshop full of uncompleted projects that went from being fun and interesting to boring and abandoned.
> If you're a tinkerer, you don't care much for 1 and 5, and you see 2-4 as the whole point of programming. I'm a tinkerer. I'd be happy to just write code and throw it away. Coding is like solving sudokus. I could skip steps 1 and 5 forever. [...]
The 5 phases look kind of symmetric. The outermost layer (1 and 5) are the entrepreneurial steps.
Even if you deeply into tinkering, 1 can be an insanely exciting and rewarding step: for example "how could an abstraction look like in which 2-4 become trivial special cases for many classes of programming problems?".
In my opinion the difference between "entrepreneur type" and "programmer type" (and yes: it can happen that both of these types work in the reciprocal job) is rather:
- People of the programmer type see solving problems as their primary goal. Satisfying customers just serves the purpose that these people pay money so that they can continue working on solving problems.
- People of the entrepreneur see satisfying customers as their primary goal. Programming is just a means to an end.
Yeah. I enjoy 2-4 most, 1 is the next, and 5 the least. I think the litmus test here is "if you knew could never deliver this to anyone, make money of it, get recognition for it, or even make it solve one of your own problems you have. Would you still be willing to spend time on it?"
If the answer is "no" you are probably more entrepreneurial than tinkerer. But it's a spectrum.
Good test. I would add one more question: "if you could never talk about your code or solutions with anyone, would you still be willing to spend time on it?"
This helps differentiate between coding as an end and coding as a means to socializing.
I enjoy the process of solving problems with LLMs. I understand it's chaotic (what many people call non-deterministic), and that doesn't appeal to some people, but solving problems with statistics is problem solving just the same. Just because medicines don't work with 100% reliability doesn't mean Jonas Salk didn't solve a problem. The same is true with LLM solutions that don't work 100% of the time.
I think most people will come around to this way of thinking once the pro-LLMers stop pretending like the statistics don't exist too (i.e. that ML is the solution to any problem, despite how unreliable it is).
LLMs can be very good at 2. Finding the best architecture, data structures, algorithms, looking on arXiv to see other ways to solve the problem, etc.
You have to work in design mode and stop it from pissing code. Describe the requirements as detailed as possible and refine them based on the responses. It's quite rewarding and helps structure the way you think about the problem.
I have always thought that there are two kinds of programmer, the tinker and the mathematician. The tinker likes to code operating systems and shells, the mathematician likes to code Haskell and Agda. You've updated my belief that there are a third kind, the entrepreneur, which sees programming as a means to an end and would gladly stop programming if possible.
Looking at your ladder, made me think where I fit in and I mostly enjoy steps 1 and 2. The implementation and shipping are big chores to me. Once I know, half in my head, that the problem is solvable and I could see a path towards implementation I loose motivation to continue.
So possibly there are other classes of people, not just either enterpreneur or thinkerer. Thou not sure what this combination is, maybe something more like R&D.
And LLMs give a lot of value here. They allow to quickly do steps 2-3 - "hard and tedious" stuff - to confirm that the thing, in fact, became solvable.
I really like that categorization, but I think you've missed a step (or several): designing and implementing all the extra stuff needed to make it work, but aren't directly related to the problem.
For example, you have an idea for a cool and useful app and you want to put it on a website. You need to figure out hosting, you need to write some copy, you need to do some visual design.
Maybe 3a) Yak shaving?
There are various ways to take shortcuts. For hosting, you can use something super easy like GitHub Pages or Netlify or Cloudflare, or maybe piggyback off an existing site that's already deployed. For both copy and visual design, you can just ask Claude or Codex to do it -- but only if you don't mind having that standard vibe-coded look and feel, all glowing gradients and punchy, mic-dropping text.
Maybe you're including that under 5), ship/deploy/publish, but I read that as being more about scaling, marketing, and making things robust. Even before you reach that stage there's always a bunch of bullshit to work through before you can even reach 4), seeing the solution work for yourself.
Nice analysis. I really love diving into 2 and 3, but I also notice I'm pretty good at 1; at several companies I kept identifying problems we should address, and also how to address them, though quite often they just wanted me to focus on 3 and 4.
Currently I'm in the process of starting my own business solving a problem my last employer underestimated, and I find myself focusing much more on 1 and 5, which I also enjoy, and I use LLMs a lot for 3 and parts of 2. The LLM and I often disagree about 2, but that often results in a better solution than if I'd just trusted either myself or the LLM. I iterate a lot between 2 and 4, letting the LLM do almost the entirety of 3, and advising on 2.
That's not a bad model to view it through, but I would say calling them "phases" instead of something like "activities" creates a false reliance on ordering. In many successful programming projects/products, those five activities occurred simultaneously, or out of order, or in waves of progress along all or some activities. Seldom has it ever been a clean single waterfall pass.
The thing with programming a computer is that the computer does exactly what you tell it to, so in order to implement something well you need to have a very deep understanding of the problem. And it's that understanding that is valuable.
even for an entrepreneur type - you need to understand the technology to get to good enough solution e.g if you don't know shit & just accepted what the llm gave you - u might end up with an unmaintainable mess. the llm might recommend you use some proprietary solution when an open source solution exists that works well for your use case.
plus llms tend to be wordy or make things complex more than necessary.
This is a really useful framing. I also love step 3 and when I delegate it to an LLM it takes the joy out of my work. But where I've found LLMs most useful is steps 1 and 2. They are terrible at steps 1 and 2 on their own, but they have a massive amount of knowledge that can help. I think a human and an LLM working together here is the best scenario.
That's actually a great classification and I've never seen it being laid out like that. I guess it can be useful to know "what kind" of programmer one is.
You are obviously missing the GOTO statement in your list but the other part is hobby projects tend to be architecturally small.
The moment you have more devs architecture becomes the overwhelming concern, and this is why professionally LLM usage explodes since the LLMs can implement bits while the humans work on the actual hard work of making sure they fit together properly in the intended way.
I have encountered a lot of people that view coding as a therapeutic exercise, and they were already a problem pre AI wave, now they are only going to be a hobby at least.
Best explanation I’ve seen. I’m an entrepreneur type. I would fall asleep after two hours of coding to solve a problem. I still remember 100 hours of coding to facilitate a data import into a new application only to realize half my time was wasted because the import library we were using back then didn’t support UTF data (this was a SQL-92 database). I love using an LLM to do the coding.
Personally I think I'm a 1-4 type, and I enjoy 2 but 3 is mostly annoying. LLMs are good at 3 and 3 is the worst part of programming. I don't think that's "manager type". 2 is also what LLMs tend to be bad at, especially the small ones, so there's a lot of value I still add to the process.
Nice analysis! I would perhaps add that LLMs are pretty helpful at (2) in my experience, as long as you don't turn your brain off. (As a Google/SO replacement, mainly.)
I have about three modes, working with LLMs:
A) Just use them for (2), do (3) and (4) myself.
B) Just use them for (2) and (3), but with very fine grained instructions - to the point where it barely saves me time.
C) Use them for (2), (3) and (4), but with this workflow, it quickly degrades into slop for me.
I still get the best results with (A), but for a quick PoC, (C) is hard to beat.
What's hilarious (or sad) is that I see a lot of people throw them at (1) as well, which seems to have a high chance of inducing AI psychosis and creating software nobody understands, let alone needs.
I agree. If you use the AI for (4), it's pretty much guaranteed to hallucinate whatever it wants and then approve it. For small projects, it might be good enough. For anything serious, it's not.
I find I get the best results when I just use it for (3), and sometimes a bit for (2). Going in with at least an idea of how to make it work has the best chance of ending up with a good changeset. And if you already know how to make it work, don't leave it to chance. Tell the LLM what to do, then let it do it.
Regarding (2) figuring out how to do things at a higher level (e.g. the general architecture of an application) is pretty different from figuring out things at a lower level (e.g. how to fill text with a certain gradient in CSS - the stuff where years ago you would have spent hours reading documentation or browsing StackOverflow until you figured it out, if you weren't familiar with the topic). So you could use an LLM just for (2b) and (3) and do the rest yourself.
> the stuff where years ago you would have spent hours reading documentation or browsing StackOverflow until you figured it out, if you weren't familiar with the topic
Why is it always hours to consult docs? I find most answers within minutes (including if I have to read code) and it’s only take a bit longer when I want to understand the why’s.
The key is to make sure you have the right questions. I see people struggle with that where they only know something wrong but they can’t put it in words.
Ok, maybe you're one of the fortunate ones. With "if you weren't familiar with the topic", I meant the situation where you know what you want to do, but have no idea how exactly to accomplish it. Especially with frameworks, which generally have a quite rigid way to do things, if you're unfamiliar with that, it can take a while until you get anything done, despite good documentation.
Most software work is job related and business related. You are paid to work not to enjoy.
Because of this the future of software will most likely be overtaken by AI because the business aspect of it will overtake the enjoyment aspect of it. Why? Because business pays bills, enjoyment does not.
The fact that you are talking about lack of enjoyment on the job is a luxury not many employees can afford.
I might be weird, but I love phases 2 to 5 and I hate phase 1 with a passion. I hate speaking with people to see what their problems are. I want to solve problems, not user surveys. And all the itches I have are either already solved, not itchy enough or too wide / ambitious to approach them solo. If LLMs solved exclusively 1 and told me "this problem is actually valuable, go and solve it", I would be the happiest person in the world. It's not true that ideas are cheap. Or better: ideas themselves are dirty cheap, good ideas are invaluable.
Even if you're entreprenurial, LLMs are useless for this.
You can't go really hard unless you can come up with something people weren't expecting, but when they discover it they want it. You can't just sit around waiting for people to want things properly. That's extremely passive.
The point being, the idea of using LLMs to find what people already passively want is itself flawed, even if you like marketing and selling to people. It's a really fundamental problem with the idea.
While I agree with the premise that there are Thinkers and Shippers, I reject labeling of tinkerers and entrepreneurial. There's nothing entrepreneurial about working for a big business and shipping cool things.
But you're ultimately right, some people want to solve sudokus and some people want to change the world, and the sudoku solvers are really upset that a Sudoku Solver and World Changer 9000 device is widely available, because it spoils the fun they have with sudoku.
Because people that enjoy programming for programming's sake don't want an LLM to do the programming for them - isn't it obvious? It's just like with any other hobby, people who like car racing created rules that force you to drive yourself, even though they'd get faster lap times with electronic driver aids. People who enjoy grappling created rules that force you to grapple, even though striking could win a fight faster. People that enjoy chess don't allow you to bring a computer to the chess table. The list is endless and the people that enjoy programming will create rules that force you to program, why shouldn't they? It's their hobby.
Writing code without an LLM is akin to cooking a meal yourself, as opposed to something in between hiring a private chef and pulling a slot machine lever. When I cook a meal for myself or my loved ones, I become deeply familiar with the ingredients and the recipe. I am a good cook. Every decision, small and large, I make myself. Using feedback, I can tweak the components and the process.
However, if I order food and call it my own creation, I’ll be speechless when my nut-allergic friend asks me if I used peanut oil or olive oil. And does it taste good? The end result is subjective, not binary, like your three similes. What if the food sucks? Must I now micromanage the chef? I don’t want to make them cook it again and again, hoping each time it’ll taste little more like how I want it to. I’d rather just cook it myself; I am a good cook, afterall.
Well to me using an LLM is akin to using microwave for food, and programming is getting your hands dirty in the art of cooking. I used to do the latter but the pace is too much, so I'm forced to go for the former. I still miss the days where I can cook myself
This is a good analogy. Also I hate cooking; it feels like a chore and waste of time. Given the means, I prefer to have someone else do it for me so I can focus on more interesting things.
If the analogy is good, does it mean AI boosters hate programming, and feel like it's a chore and waste of time? Yet they keep calling themselves programmers.
I dislike cooking; I don't call myself a chef, nor go on cooking forums saying they're all idiots for churning butter as a hobby, and telling them they are luddite dinosaurs just because I bought myself an Instant Pot.
And clamoring that instant pot is capable of doing the job of 10 chefs, while conveniently forgetting that chefs using the instant pot is way more useful. And then congratulating each other on new way of adjusting the temperature and timer of the instant pot.
I don't see it the same way. I see it as akin to building a machine which can do the cooking for you. There is a lot of experimentation in terms of how you can construct that machine and how you can set it up, and for some people the payoff is in exactly this process.
That's me. I've just moved the puzzle to figuring out the CC optimal "language" and mode of working with an LLM. At the same time, I haven't lost my love/expertise... Yet.
I explained it yesterday that it's no different than Linq powered by EF Core. You still need to understand the database and how SQL will work, or should. If you don't, you're queries will be slow and inefficient, hazardlessly oftentimes.
LLMs are an impressive abstraction, but they are not magic...yet. or maybe it is, but you still need to figure out the incantations. Not that I like to uptick and witchcraft, so I hate that analogy.
My boss tried to make me interested in agentic programming by showing me how "productive" the creator of OpenClaw is.
All I see is a fast food factory. Sure it's interesting that you can produce so much so quickly. But what is produced is endlessly uninteresting to me. It doesn't help that all the harness building is just SEO in disguise. Trial and error.
I don't doubt that agentic programming makes it possible to produce faster, but it also comes at a great cost for me, which is that I largely don't enjoy my job anymore. If I was given the choice between that and halving my salary I would have chosen the latter.
Yeah exactly. My go-to example is that if you enjoy playing guitar, you're not going to stop doing it because Suno exists now.
In addition, say if you had like a guitar-playing forum, where people share recordings of their practice. People would likely not appreciate it to share AI-generated audio of guitar-practicing there, either.
And while there are various reasons for a guitar forum that might not hold for a programming forum, one of them remains: There's nothing to talk about when it's AI-generated. What scale did you use and how did you tune it? "I dunno, AI did it"
All you can talk about is to say "oh cool that's cool that the algorithm machine you're paying subscription to access made you that thing, nice"
Or for example, even when a friend of mine vibe-coded a "synthesizer" app and he wanted to show me cause he knows I like that kind of thing. And I do, I love digging into and experimenting with DSP algorithms, but also the user interfaces and all sorts of things digital audio related. And I love talking about it, or explaining stuff. But my friend couldn't tell me anything (not even that it was using WebAudio, but also how). Anyway it made me a bit sad, and I felt like something is lost here. Something worthwhile. And I explained it to him that then all I could really say was "oh cool that's cool that the algorithm machine made you that thing, nice", and we talked it over, and it was alright. But still, it seems something is lost.
This is true, but as usual nobody can agree on what "doing it yourself" means.
Few cars are manual steering, manual brakes, manual transmission.
Few grappling bodies go all Greek-naked.
Few chess bodies forbid researching previous games or your opponent.
The list is indeed endless, and every single person thinks they are using just the right amount of assistance; anything less is pointlessly difficult, anything more is cheating that takes the sport out.
Me, I've been a hobbyist programmer since youth (6502 assembly, represent). I'm having a blast "writing" software without touching a line of code. Sure, I'm old, and maybe I'm like those guys blasting Veyrons through highways, relying on traction control and all the computers to let me drive beyond my abilities. And... ok? I'm having fun, even if I'm not doing it right, by random anonymous peoples' views.
> I'm having a blast "writing" software without touching a line of code.
And that’s totally fine, enjoy your hobby however you enjoy it best. Most people programming by hand for fun don’t care that you program with an LLM for fun. The problems come when those two worlds collide, like in the linked GitHub issue, in chess engine leaderboards. In the long run for both groups to have their fun and friendly competition you’ll need two separate leaderboards. This is also very common in other hobbies. In racing there are car classes so that your Veyron and someone’s old Miata don’t go head to head. In combat sports there’s weight classes. Most sports separate amateurs and pros. Usually there is respect across the group boundaries and appreciation for the other groups deeply specialized skill. There’s even the concept of cross training to get better in your flavor of a sport. This will be figured out with programming as well.
Sure, I agree, and it’s already figured out. The demoscene has been around for decades. That’s real programming.
But I think you’re on to something important: framing programming as a competition of intrinsic skill where rules are necessary to equalize. That’s certainly true in some cases, and more power to those who enjoy it.
I guess I’m just more of a product guy these days, where the equalizer is the market and the consumers rather than the implementation methodology.
You never went to a programming competition in school? None of what was said is novel, it’s all been around forever, the tricky bit is the regurgitation software mucking everything up.
Not that it’s about me, but I was doing programming competitions 25 years ago, half of which were on paper with no keyboard allowed.
One of the great things about hobbies is that you can enjoy them whatever way you want. But honestly is it not more accurate to say that you are now a hobbyist prompter, rather than a hobbyist programmer? I get the "its just a tool" mentality and I guess LLMs can be used that way, but if you are not touching a line of code yourself then it seems less analogous to using power steering and more analogous to just taking a taxi.
What I really don't understand how differently it seems that AI for programming and for anything else is treated.
I don't think there is insignificant group who would say that using LLM to write a book does not make you an author. Or make a song musician. Or draw a picture a painter or artist.
This just somewhat baffles me... Or maybe the LLM for programming people all believe that you are as much artist if you use AI in those fields and the product generates as such is as if not more valuable than one by more traditional means.
I've seen people here insist that LLMs enable everyone to make music or art. It's diametrically opposed from my own view that making something is the process of making something, neither the result of nor the motivation for that process.
My retort is usually a variation of one I've already posted in this thread: they do that to the same extent that taxis enable everyone to drive, which is something most people will intuitively understand as nonsense, even if they've given clear instructions to the driver and even though they've arrived at the intended destination. If you aren't driving, you aren't driving. If anything, the taxi service enables more people not to drive.
isn't this clearly a misunderstanding between you and hypothetical AI art approver?
They're talking about the result, one must think. The idea that people who could not have approximate pictures of ideas that lived in their head can now have those things. They didn't create them by an in depth process, of course, and would be wrong to pass it off as though they did. But there is some value to the enjoyment gained from that existing (whether that relatively trival enjoyment is offset by other harms is for the reader to decide).
You're talking about the process and/or the developing of the skill of crafting and thinking about images better.
We also lack words to talk clearly about the distinction here around "make". Did I make a piece of art if I created a file on a computer in photoshop and then printed it? or did the printer make it? Did photoshop make it? (imo, all are true, but in differing senses of the word make). It seems to me that there is a sense that, no matter how trivial their involvement, a user of some AI art app is "making" something, in the same sense that steve jobs "made" the iPhone. In both cases, an engineer that literally designed the iPhone or an artist that spent years would perhaps rightfully feel upset by such characterization, but to get across the point cleanly is harder.
It's the same for the taxi. If I can't drive (perhaps I never learned, or some flaw with my body doesn't let me), a taxi does enable some things I could never do before. I am, of course, not literally the one driving. But if I can get to the shop myself now, there is something there.
> They're talking about the result, one must think.
Maybe they are, but in that case it's a frequent source of equivocation because people will conflate the process of their making with the result of some other process so as to proclaim themselves artists, musicians, programmers or whatever for writing something that they themselves wouldn't consider art or music unto itself in the first place, and having a machine generate something which they do consider to be art or music based on that.
> Did I make a piece of art if I created a file on a computer in photoshop and then printed it?
Funnily enough people don't usually have a problem with this question when it comes to driving. Your actions' effect on the propulsion of the car are ultimately indirect, and there are many systems which assist the person behind the wheel even when they are engaging directly with all their means of control, e.g. an engine, ABS, EPS, yet no one will immediately feel a need to wax philosophical about whether the driver is actually driving or not because of that. We recognize that in a car which requires a driver at all there is substantial ongoing involvement from the person behind the wheel in making the car go as desired and deciding what means available to apply in order to go about that, without whom there is at best no going anywhere at all, and that involvement is what driving is.
> If I can't drive (perhaps I never learned, or some flaw with my body doesn't let me), a taxi does enable some things I could never do before. I am, of course, not literally the one driving. But if I can get to the shop myself now, there is something there.
Which is not the thing in question. If AI tools have generated something you find beautiful or interesting, by all means enjoy that for what it is. If the taxi takes you to the store, by all means enjoy that.
A renowned author would still write. They don't suddenly start managing a team of junior writers and instructs them how to write.
A programmer that has climbed up the ladder in their field, on the other hand, often stopped coding themselves. Instead, they manage others who code, or provide vision/direction on the project.
I think it really depends on what people on the senior level of your field do. If they mostly move to managerial roles, then it's much more likely they'd embrace LLM.
Ironically I observe this in academia too. Professors (who often stopped doing hands-on research, but only provide guidance to postdocs and grad students in their research) tend to embrace LLM more than the fields you mentioned.
Sure, I’m both a professional and hobbyist prompter. But then, that’s been true for decades, since I made the switch from engineering to product.
Only difference is I’m prompting LLMs rather than engineering teams. But honestly it’s not all that different, except that in the hobby flavor I can write specs that would take tens or hundreds of engineers to implement.
Maybe it all comes down to semantics: does “programming” mean typing in code, designing app architecture, or producing software? When I was young I was probably 75 / 5 / 20 between those. Now I’m 5 / 20 / 75. Maybe it merits a different word?
> This is true, but as usual nobody can agree on what "doing it yourself" means.
Not everyone has to agree. Just everyone in a particular group at a particular time. Grappling competitions exist and so do MMA competitions, and both of them will expect you to follow their own rules.
But LLM usage isn’t equivalent to power steering, it’s what a full kitchen is to a microwave.
You can sure put effort into a microwave meal but it’ll always work against you and much of the time you’re better off just picking up a pan.
Not to diminish your enjoyment, you’re completely entitled to it, I just think this is a bit of a false equivalence
Fully agree with this. Also did lots of 6502 assembly, and also love doing agent-driven development. I also still like writing code manually and tweaking everything for myself, but those serve two different purposes.
Sometimes I'm having fun getting to the output I want and otherwise wouldn't have had time to. Sometimes I'm having fun exploring the code.
I don't care if someone don't see the value in one or the other of those, and to me they stand on equal footing as things I find both fun and valuable for different reasons.
Actually yes. Come to southern or eastern Europe, on plenty of those side road dealerships with signs on their windscreen, or parked in random places with a telephone number on them.
I'm in Central Europe and there aren't many manual transmissions around anymore where I am. But if you paint with a brush of two entire continents, you might paint a little too broadly.
I used to write code all day, solving maths and logic problems which is what my brain loves to do, and so I loved my job.
Now I talk to a robot all day, trying to figure out how best to tell it whats in my brain. I no longer do maths and logic and instead I think about language and questions all day, holding in depth conversations and thinking up responses. This is something my brain naturally doesnt like to do and so I fatigue quickly in the day compared to my old job.
I enjoy programming a lot, but still made the transition to using AI for my programming. It is a new tool that makes me much more efficient, I finally could continue working on my many open source projects that I couldn't touch for many years due to lack of time. My productivity is up, I'm pretty sure the code quality has improved, and I'm learning more than ever before; just in a different way.
for me its that software, wrought by hand is "bespoke". Yes, there is a loooooot of software that is just commercial, quickly vomited by hands that are just in it for the money. On the other hand there are open source projects who might have a miss guided design philosophy (Darktable comes to mind) but are otherwise build from people that care. There was an "artisan" at work, someone who cared.
When I see all those "show HN" projects that are slot level code it feels like they dont care about the result that much and just want to showcase their idea in a more tactile way, but there is no actual proper execution to make the thought into matter so to speak
True, but it sounded a bit paradoxical to me that the people who work on engines designed to beat humans at chess were themselves opposed to using engines that (arguably) beat humans at programming...
If you don't want to paint, and also don't want to pay somebody else to paint them, there's always various alternatives like printed out papercraft stands. Fold them, stick them on the base and you are ready to roll.
There's also a myriad of other creative ways to proxy your way out of the general costs (time, space and money) if you check out the term "poorhammer".
You can take already painted warhammer figurines, pose them, take a photo, prompt an AI to make that your first scene, and generate a hollywood level scene in about a minute or two.
Yes, sure, you can buy the figurine and get the joy of painting it. But there's new things you can do now too, things you couldn't dream about before.
But I'll grant you this, the AI coding extinction event is so severe that I do believe we need to support anyone trying to preserve any memory of it. Like, we need to build entire museums, one for each programming language.
> You can take already painted warhammer figurines, pose them, take a photo, prompt an AI to make that your first scene, and generate a hollywood level scene in about a minute or two.
if that's so awesome then why isn't everybody doing it?
I never dreamt about stuff like that, but are there actually people who are like "oh thank God I can finally make my AI movies with my figurines and other stuff I have laying around"
honestly you sound too enthusiastic about this, like Sam Altman "AI podcast about your kids" tier enthusiastic
People are doing it, but not that much, because while it generates some pretty good five-second scenes, it fails terribly at generating any long-form video. And by long-form, I mean 'longer than 30 seconds'.
I think what people want is to tell a story that's in their imagination, and no current AI is up to that task (at least in video format). Great for making an animated PNG for a meme, but not for adapting that fanfic you are writing.
1. Some programme because it’s the one of 2 ways to get the things they want (another way is hiring a team to code for them). Now that they have vibe coding, they’d happily never touch a line of code again.
2. Other people programmes because they just enjoy it. Vibe coding ruins the best part for them. Like telling a foodie, "Good news! I built a robot to eat your food for you."
Most people getting paid to code are probably somewhere in the middle?
But nowadays hand-writing code at work is starting to feel kinda 'illegal'. The manager is gonna ping you just because you fall behind on the company's token waste leaderboard.
Maybe it's just another "hobby should never become job" story?
Hobby should only become job if you enjoy the craft of it and if the profession / job rewards craftsmanship.
People in it just for the money won't care about craft - you see the same thing in plumbing. And when there's enough money and scale on the line, craft becomes progressively more niche - that is why Ikea furniture is a thing. And you can even say that there is craft in Ikea, just not along the axes a traditional furniture maker or user would consider.
This is such a great point and raises existential career questions for everyone who got into a programming career because they love the craft. Either find your niche where craft will matter until you retire, leave the profession or accept that it’s now mostly about prompting. It’s almost like everyone’s being promoted into management or tech leading and can’t refuse…
I think it was implied through the conversation but the other thing I have come to realise is that your hobby and craft should be about producing not consuming. Too many people think they like videogames or films and think they should therefore work in the industry only to be crushed by the industry. Even being a youtuber devolves into hours of storyboarding, editing and engagement farming which are orthogonal to your hobby.
With respect to code it's a little murky but I guess the tension really is enjoying the process enough while being able to output the desired quality, volume and velocity required of you.
Quality is usually the stickler because it's invisible to the end user and requirement setter but sometimes quality affects both ergonomics and future delivery and sometimes it's purely self-satisfactory. That isn't a value judgment, making things well is a gratifying and edifying experience - so much so that this practice overlaps with Shinto philosophy.... but not many people are going to pay you to have that experience.
We're not that far into this new era and I already see people go back to hand-writing the important bits and manually designing their architecture because it turns out that unless you have them spit out yet another boring web app, they aren't very good engineers at all, their errors and bad decisions are very hard to predict and accumulate over time, and the resulting mess becomes unmaintainable at some point, and it appears all the harness engineering in the world can only stave that off somewhat, not prevent it from happening. So perhaps IT won't become a monoculture of agent prompters after all.
why would you want to filter these people out? surely those that consider programming more than just a "means to an end" are going to be much more valuable employees, consider they are much more likely to be skilled at their craft? im genuinely interested in why you'd want to filter these people out!
I want to find people who have already had all their passions and emotions stamped out by the endless grind already, so that they will tell an AI to do what I tell them for 996 without complaint.
These people who "care" about things just get in the way and want time off and such.
Offer low pay and start using lines of code as a metric. Not only will you not hire such people, you'll also filter out any who might already be on your team!
(Assuming "these people" is a euphemism for "skilled engineers")
The article just straight up doesn't mention the context of the Github Thread, which is about the guy lifting ideas from other engines and copy-washing them.
"Do you genuinely believe changes like cd170a4 make the code somehow not derived from Reckless anymore? It appears to me like it's just trying to obfuscate its origins, to make it easier to appeal to plausible deniability. The reality is, coda 100% has code derived from AGPL or unlicensed engines in it, and pretending otherwise won't convince anyone other than maybe yourself. This, combined with the fact you conveniently moved the plagiarism roadmaps to a private repo mere days after people started raising concerns about it, reads to me like you're not at all concerned with actually using code fairly, and much more with being able to claim that you are, backed up by these nonsense "cleanup" commits."
> After our initial 0.9.0 pre-release (July 2026), community feedback raised that Coda contained AGPL-licensed code. This prompted a full audit of our codebase and dependencies, and steps to correct what it found. We now believe Coda complies with all applicable licenses. If you have any concern, please open a GitHub issue with the details and we'll happily investigate.
The debate is about whether or not the new code qualifies as derived.
They can remove all of the literal code matches from the codebase, but if it was written by having an LLM study the original code and then write something similar, it's hard to argue it's not a derivative work.
There is a lot of legal history of doing new designs with influence from reverse engineered sources. The basic rule is that you can't have someone familiar with the original code involved with writing the new code.
LLMs are a new area, but I would expect that feeding the original source into an LLM context and saying "Write this, but different" wouldn't qualify as a clean-room design. There isn't enough general removal from the original subject matter.
Specifically about "humans": my understanding is that Clean-room design is not a requirement and the "Case law" section on your Wikipedia link explains that and has examples.
How and whether the same principles can be applied to LLMs, I have no idea. I imagine it would involve discussions about creativity, for example.
The thing that happened with "clean-room design" is that one company did it and the court said it was allowed. Companies are risk averse. Once they know one thing is yes and something else is maybe, they go with the yes, right? With the result that the supply of court cases where something other than that happened tends to dry up, leaving the other alternatives to stay as maybe.
If some people have now decided they're not going to blink in the face of a risk, we might get to find out whether some of the maybes turn out to be a yes or a no.
I have done this. Despite using Fable for the spec, and having it double-check the spec a couple of times, it still wasn't airtight. And at some point, the implementing LLM accessed one file of the original code on the Web.
I have by now deviated quite a bit from the original (well, original...) spec.
Would have to be proven in court. Look up history of cases and settlements for GPL violations and it quickly becomes apparent that it’s relatively toothless and mostly license theater.
In my opinion yes. Just like if you watch a Bob Ross episode and copy all his brushstrokes, you have made a derivative (in the literal sense of the word) of his painting.
Not content to destroy one potential competitor, USL also seeks to keep a stranglehold on the entire market by claiming that everyone who has had ever had access to UNIX source code a group that include~ most experienced programmers in the industry is "mentally contaminated" and may not develop any functionally similar product, without first purchasing an exorbitantly expensive license from USL. In accordance with this argument, USL requests this Court to prohibit BSDI from hiring anyone who has had access to UNIX source code. USL's request would not only require BSDI to fire all of its current employees, but would make it difficult to find any qualified replacements. Having chosen to widely disseminate its source code in order to build a market for its product and having succeeded in making it an industry standard, USL cannot now claim that everyone with knowledge of it is restricted from developing competing products.
Mechanical transformation is always considered derived.
The difference is that if humans do it, there's plausible deniability that it is derived.
I mean, even when humans do it, we still do clean room engineering.
It goes further though: LLMs are just a tool, and tools don't get rights, humans do.
Humans having the right to learn from a few dozen books and projects is not the same as a tool that infers output using all copyrighted works know to man.
Yeah, absolutely. The defense there is "clean room reimplementation ". If anything, the use of an LLM would be stronger than a human doing the copying. (Assuming American jurisdiction!)
This is not really a question. Accidental patent infringement is still patent infringement. Just because you used new trendy technology to do it does not mean that it isn't infringement.
It's not like ideas are protected by copyright though. Copyright protects the specific expression of the idea, not the idea itself. Anyone can "lift ideas" from anyone else, including proprietary software.
Except "Claude, take this code and change the variable names up a bit" is not lifting ideas, and you would not get away with doing that to software belonging to a company with a competent legal department.
> lifting ideas from other engines and copy-washing them.
You cannot copyright ideas. And that's a good thing. Remember the "one click checkout patent" saga? That is bad for everyone involved. We don't want that, as an industry.
Keeping in mind that I'm going strictly off the commit[0], a bit-packing layout is very much the kind of thing which is not subject to copyright in the first place. Neither are algorithms. Ideas, in general, are not subject to copyright: expression is.
So "lifting ideas from other engines", as you yourself put it, is entirely licit, no matter how you do it. _Should_ you credit others when so doing? I err on the side of yes, on that question. _Must_ you? No, there is no such requirement, this is not academic publishing. Does it constitute violation of license? Absolutely not. Straightforwardly, no. Furthermore, it _should not_, we would not like the result of that.
Think: SCO suing Linus Torvalds over his use of a.out: and winning. It would be bad. Even though I understand the reasons for it, it's still eerie to me to see hackers coming down hard on a side of an issue we spend decades passionately defending the other side of.
As I lead this post with: maybe there's something in the repo which is infringing, it's quite possible that commit doesn't tell the whole story. If so, keep in mind that I don't know that, and am able to change my opinion in the face of new evidence.
Based on what I _have_ seen, no: you do not get to lock down a bit-packing pattern just because you're the first to come up with it. Nor should you have that right.
You'll notice I didn't mention copyright, and that's because I don't think it's relevant. I think your analysis is correct. Keep in mind too that Stockfish – the engine I work on, although I wasn't at the time – was the subject of a high-profile defense of the GPL in Germany: https://stockfishchess.org/blog/2021/our-lawsuit-against-che.... That case was different as it involved wholesale copying of the source code, rather than porting ideas. Ideas shouldn't be copyrightable.
But copyright and plagiarism are orthogonal, and questions of morality are much more tied to the latter. It would be illegal (in the United States) for me to publish a copy of Nineteen Eighty-Four, but not immoral. It was not immoral, in my view, for Aaron Swartz to try to liberate JSTor articles.
That doesn't make plagiarism acceptable, either. And therefore, I'm willing to call it out when I see it.
No idea why you're getting downvoted. You're right.
> it's still eerie to me to see hackers coming down hard on a side of an issue we spend decades passionately defending the other side of
Yeah, it's surreal. It's not just intellectual property either, Hackers have already started complaining about automation now too.
Computers must have replaced tens of millions of jobs. It's the ultimate job destroyer, since it allows capital to do more with less. Automating me, a doctor, out of his job is essentially the holy grail of Hacker News.
But when hackers are automated away, then suddenly it's a crime against humanity.
I think that's true, but I'm not sure if this would characterize itself as "lifting ideas". I mean, it's one thing to personally take the idea and re-implement it yourself elsewhere. But I have to believe there could be an argument, that having an inference algorithm take the original source and re-implement it in the context of your engine might count as derivative work.
A particular work is deliberately supplied at inference time so the system can recreate or adapt it for another environment. It resembles commissioning a translation, port, or derivative adaptation through an automated tool, even where the output is not verbatim, so it seems it could hold a candle in court.
You have a few scenarios now up for legal debate:
1. Is training a model using copyrighted material fair use?
2. A model outputting chunks of identical coryrighted code from it's training material
3. A model being given in it's prompt copyrighted source code, and asked to implement the same in a new source code.
I think #1 and #2 have open legal battles on it, but #3, which seems to be this scenario, is new to me, and I feel it could have legal ground for being considered infringing on the copyright.
Sidenote: how is your account "samatman" and created "April 24, 2010", that's a really funny coincidence
Whatever the legal outcome, whether it's ultimately decided that people can or can't "get away" with using LLMs to license-launder code, it's undoubtedly a dick move.
Programming is nice because you have lots of different scales of structure to think about, and your program slowly comes together. You see it improving over time as a result of your efforts. You also build your skills and learn. It’s really a nice activity.
AI systems basically change that whole formula. I’m not sure how people get joy out of constantly vibe coding new projects. It feels like binge watching Netflix shows. I feel like there’s no material value in my 50th vibecoded demo since everyone in this world can vibecode now. Sure I’ve got my custom todo app which is nice, and my custom game plug-in with some neat features for the game I like, and a few other custom apps. Those are mildly useful to me. But every second I spend vibe coding something that’s supposed to be a success for others is just invaded by the thought that even if this thing is any good, someone with real capital can just direct a future AI at it to take it from me, so I’m not really gaining.
So I think I’ll just stick to using AI occasionally, without letting it ruin the programming activity I love by spoiling it. Due solely to AI, I don’t expect I’ll ever do anything in my life which will impact another person. That’s a personal tragedy, but in a sense I was going to die anyways and so will everyone else, so it doesn’t really matter.
>I’m not sure how people get joy out of constantly vibe coding new projects. It feels like binge watching Netflix shows.
It's just power fantasy. It's amazing the first couple of times but as it wears off and becomes the new norm you forgot how being the one who decides which thing does what feels like.
Luckily (or unluckily?) the industry only cares about the end result, and flushes out people who used to reward for their passion and dedication naturally.
A lot of this is from non-tech workers wanting to expand their repertoire as well.
They have no clue what they're doing programming-wise, and in some cases this is perfectly fine, but after years of SaaS, some people are taking it to extremes, and spending millions of tokens on things that could be a spreadsheet or a note in the calendar.
Just recently we had someone deliver a report in the format of a 200.000 LOC application that requires several clusters to run, for example. It does all querying in real time from clandestine APIs, or by crawling existing systems. The person who asked for the report asked for it not to be deployed.
I get nothing of the power fantasy, of it, but I do get joy out of vibe coding.
As I've gotten older, the amount of ideas I have here and there has remained constant, while the time I have, or brain power at the end of the day when I can work on my passion projects, has diminished. What was once basically a second day job at night became family time, baths, dinner, etc. So while I may be too exhausted to sit and code, I am not too tired to sit and type my ideas, or discuss a single feature I want to see, or review what's there. I get joy that ideas are getting out of my head faster despite less brainpower to get them there. I can always grab the reigns and take over myself, but that's not the kind of thing that wins against going to bed.
> I’m not sure how people get joy out of constantly vibe coding new projects. It feels like binge watching Netflix shows.
I get joy out of it because it solves actual problems I have that I don't have time to solve manually, or don't want to. E.g. I just recently had Codex add a mechanism for two-level virtual desktops because I wanted groups of them. An hour later I had two-level virtual desktops in my wm that worked exactly how I wanted it, with support for it in my desktop switcher. I could write it myself, but why would I? I love writing code for various projects, but sometimes I just want the outcome.
For me that’s a slippery slope. It’s not really clear what parts of my hobby project are the good parts and what parts are unfulfilling. When I start including AI too much I end up feeling that none of it is fulfilling. When I do things manually I feel pride in the details however, even though those same details on another day would make me feel like I’m wasting my precious life.
I spent 2 months having Opus build a demo for a project I dreamt of for 15 years, but somehow now I can’t feel interested in it. I’m trying to avoid this. A good use case for me has been using AI to help me acquire all primary source documentation on a topic and then prepare a curated sequence of study based on a small demo I’d like to build.
Keep in mind I’m talking about how I’m approaching creative programming in my personal free time.
To me it's very easy: If I want to do a change by hand for the joy of doing that, I do. But if I want a piece of functionality now and don't want to set aside time to build it, I let an agent do it.
It boils down to whether I dream about the outcome or how to do it.
E.g. I have no emotional attachment to my window manager - I wrote it (by hand, originally) because I wanted a specific set of functionality, not because I particularly wanted the experience of writing one.
I've recently made lots of changes to it with Opus because I wanted another set of functionality and was elated not to have to spend time doing it manually because what I wanted was the functionality.
I'm now a very happy user of a better window manager than I was before.
Conversely, I have dozens of projects I stay closer to, because I care about the code and don't particularly care if that means I might have to wait for years to get to something usable, if ever.
I think there’s a lot of need for people to help others understand how to healthily engage with this technology. I do struggle with turning it off once I get started on it.
I also believe I see the future coming very fast. I’ve been addicted to watching television for weeks on end or playing games for months on end. I don’t want to fall into something like that again. The vibe coding thing feels exactly like that to me.
In the past I’ve experimented with many different programming paradigms. I think my AI coding benefits from that experience. I want to find new paradigms to write extensible, modular systems with AI, a problem that is still open. One thing I’ve learned is that when implementing a complex system, no paradigm eliminates the need to write lots code (covering many ugly edge-cases) tests, which previously required lots of boring, slow work, but AI is very good at that so maybe can make it less boring and faster.
> I’m not sure how people get joy out of constantly vibe coding new projects.
What? It's the simplest thing in the world. You get the product. Isn't that the point of all programming? You want a thing and you speak a magic incantation and then you have the thing, to keep and use.
> even if this thing is any good, someone with real capital can just direct a future AI at it to take it from me, so I’m not really gaining.
This kind of sounds like you're writing code for social credit. Like, you don't actually want the thing, you want to be "the person who made the thing". If you want the thing for its own sake, nobody can take it from you and what somebody with capital does on their own services does not harm your use one whit.
Sure. I do agree that AI is a problem in this sense: programmer skill as a hiring differentiator is becoming a lot less legible these days. It just feels personally odd to be the reason you do the thing. When I encountered my first programmable computer I very quickly realized that I wanted to do more of it, and not too much later that this was the thing I would spend the rest of my life doing. That it was also very commercializable, I have always considered as a kind of happy accident.
That’s my experience, starting at 10 years old and trying to figure out how the the moving pictures were in the screen.
I didn’t say that was the reason though. A person has a lot of wants and needs right?
I’m not following why it’s strange to want to be skilled and engage with the craft itself based on that part of the message to be honest. Maybe I’m misunderstanding.
I tried to convey in my first post that I’m gonna keep programming for no material gain and no way of producing anything of value for others. I am worried about how I’ll support myself but this thread is about the hobby.
That's fair, I think I read something into your comment that wasn't really in there...
Personally speaking, the feeling of "getting the thing and then having the thing" - the personal todo app, the custom apps - is why I do anything in programming. Programming is the project where I can speak a spell and then I get a nice thing. That's very cool, and I do like the process, but at the end of the day the nice thing is what I'm doing it for- it's not a side benefit, it's the payoff. I get a burst of happiness every time that the thing quietly does its thing that I wouldn't have had otherwise. So to me, the LLM mostly just feels like I suddenly got a lot of friends who are willing to help me get the thing and also have a very complementary skillset. So it's an improvement, and then there's also a reward at the end!
Example: I'm playing Satisfactory right now, with a mod that puts a computer in the game that can control game events. I use the computer to do cargo train routing in the game, ie. it runs a program that scans target cargo depots and looks for items that are below a level, and then dispatches trains to find these items and bring them to the destination. Which I once would have written by hand, but was now done entirely by Sonnet. I did have to put plenty of effort into guiding Sonnet to the approach that worked, and I put several AIs together, and this felt a lot like "explaining something" which is a process I enjoy uncomplicatedly anyways. And I think, to be honest, the enjoyment I would have gotten out of the programming part once, was very largely "type-2 fun"- an annoyance followed by a nice reward. Now I'm getting the nice reward for free and can focus on actual industrial and rail network design. And I still get a burst of happiness every time I see a train busily speed off to a new destination. So far as I can tell, at least for me it's all upside.
I actually think that’s a great use of AI, and sounds similar to some of my use cases.
“And I think, to be honest, the enjoyment I would have gotten out of the programming part once, was very largely "type-2 fun"- an annoyance followed by a nice reward.”
Yea I’ve had similar thoughts. Other thoughts have been something like “I’ll have different types of projects and try to be thoughtful in advance about what this is doing for me. If it’s about the building up a system manually and slowly, and having that kind of experience, then I’ll only use the AI maybe as a primary source gatherer or maybe in some limited sense.” And “Okay well this is an MMO and even though in the past I would have wanted to program this plug-in, I guess I’m middle aged now and there’s only so much time and I’d probably be really excited about using my programming energy on these dozen other things I’ve always wanted to work on, and having the thing would be a strict positive compared to not having the thing, so I’ll have the AI make this plug-in which makes the game (which is about play) more fun.”
Programming is a lot of things, so the ambiguity of language probably causes a lot of conflict here.
Yeah! And now the mod has fixed input events, I can actually have a proper configuration GUI for this, and I could I suppose write a GUI toolkit for the nth time, but it kind of feels like the sort of thing that's barely cool the first time, lol. And I already know if I let Sonnet or Opus do it, it'll work fine after the third try or so, and what's more it's going to look a lot better. And I'll enjoy using it every time instead of half-dreading finding another bug.
Another strict positive I’ve had with AI is having it figure out how to deal with bureaucracy. This is dangerous due to privacy issues, but I was seriously depressed for a period, felt like I could not do it, and the AI figured out which agency I had to contact and how to contact them etc… to do some imminently important thing, and if this thing wasn’t done I would have been in serious trouble. I don’t care to know how to interface with that (and the system was pretty miserable).
In the past I guess it would have overrun and I’d be in trouble. Maybe that would have taught me to handle my affairs though. I doubt it.
> Programming is nice because you have lots of different scales of structure to think about, and your program slowly comes together.
Programming is nice cause of this for you. Other people have other preferences. When I got my first LED to blink, I wasn't thinking about the elegance of my code. Same for when moved sprites or made a async stateless compute system for MATLAB analysis. There are people who are driven by the artifacts they produce. That's just how our brains are wired I guess, outcome oriented. I see nothing wrong in that.
Btw, I feel like there is this view that all people using agents are the apocryphal "vibe coders". That is just not true. The process of creation remains the same. Break it down into smaller pieces. Work those pieces till a piece it done. Make sure it all comes together. People seriously using these agents are not writing "implement gta6. Make no mistakes".
Programming has a lot of nice features and accurately capturing them all in a hacker news comment isn’t that easy, so maybe you could have some consideration for that in the future. I also want effects to happen in the world. I also like the slow process of craft.
I’m anticipating not a very far away future when “implement gta 6” will not be that different from torrenting gta 3 in the 2000s. The current agent era feels very ephemeral. You’re choosing to interpret some version of AI based programming that you enjoy with what I think is the common interpretation of vibe coding. If you like your process then go ahead with it. I’m not stopping you.
You seem to want the AI supported gratification and the feeling of process and craft. So make your process and think of it however you like. There’s literally nothing I can or would do to get in your way.
Have you seen what most people do for jobs all day? They do repetitive brainless things, not always new and interesting work. For all of those people, doing this is at least something you do sitting down alone in an office and so they will be queueing to do this.
And LLMs are what make so many jobs feel repetitive and brainless. Jobs that were engaging, interesting, from which people were learning, developing expertise over time, are now prompting + babysitting agents
Lots of people job wasn’t. Now it is. A lot of people in software went into this field because they actually feel satisfied engaging with systems and their implementations.
Managing a fleet of agent is the exact opposite of what is engaging about software development.
The point is most people's jobs are like that. So it doesn't matter how interesting it was or that who did it before doesn't like it anymore because there's a queue of people lined up to do the brainless new version of the job.
It obviously does matter. The industry was built by people who care about the underlying details and want to engage with the hardware and software at a deep level. That has traditionally been the way people develop expertise - the exact thing you need to effectively use LLMs for anything else than boilerplate work. Ironically agentic stuff pull us away from from that
A hobby is something you enjoy the process of doing, not just the end result. Everybody likes a clean home, but cleaning is rarely a hobby. LLMs expedite achieving the end result. Take from that what you will.
I think part of it though is that LLMs (especially in novice hands) create code that is more difficult to review and has subtler bugs. Code review is (for many) the least fun part of programming.
Like imagine if your hobby was cooking, but you hated the cleanup. If you buy a robot that automates the fun part at the cost of making twice the mess, that is not a good trade off no matter how many extra cookies you end up with.
Or imagine a club where everyone cooks together and then cleans dishes together. Then one day some people show up at a club meeting with cooking robots that create 10x the amount of dirty dishes and then don’t help clean up at all. I bet the other members wouldn’t be happy, even if the robots mostly cooked decent food.
Then robots begin cleaning dishes too. And the other members aren't happy, because what's the point, we are here not for the dishes, but for the process.
I think that also hits what you see universally with every AI product trying to pass worse results off as a benefit: if you’re a business, the prospect of cutting your cost per cookie is really attractive as long as the decline in quality doesn’t cause people to stop buying and that’s going to get the focus of everyone who sees the financial reports more than they talk with customers. Anyone who’s going it because they care about cookies or take pride in your craft will have the opposite weighting.
We’re kind of at the fast food point: before LLMs, bring successful at all in software required some level of skill (witness all of the companies getting very expensive write-offs from off-shoring or big consultancies trying to save on payroll vs. staff), kind of like how the small cafe/bakery model needed to have at least some level of craft to have something people would buy before the option of making food at a factory and shipping it became an option.
It is, although in this case the cookie-making robot also makes 10% of its cookies with sand instead of flour. So it's still not a very good trade if you have any pride in your business or consequences for bad product.
> Code review is (for many) the least fun part of programming.
I'm lurking around the OpenBSD mailing list and it seems that code review is where collaboration starts. You send a patch, explain its purpose, people try it out and comment. If it's OK (various factors decide that), it's imported inside the codebase.
The main issue in corporate settings is that the author of the PR wants to browbeat you into accepting his/her changes. Either directly or by appealing to the manager. They don't want collaboration, they just see you as the blocker between them and the "Done" status on the ticket.
I used to work at a company that did in-person pair review. I wasn’t an engineer there but the engineers all legitimately enjoyed the process. They hired very capable people as a default, so it was watching two passionate and talented engineers debate.
Main course is this this and this, 1. Do this 2. Plug this 3. Run this.. step by step plan.
So you get a precise top down construction plan. Then it’s easy to split the parts where the coding needs to be manual (smoothing walls) vs where it can be whateverz (laying concrete).
LLMs reward seeing forest instead of trees.
This way one person can implement a whole new hobby project instead of only writing a single specialized part in an existing project.
Good luck making a grand strategy game in two months holidays without an LLM. Now you can.
People focused on the trees will however lag behind as implementation details matter less and less.
Yes manual touch is still needed in some parts of the construction process but architecting the whole thing is the most important in 2026.
ADHD wins over autism in the immediate future. It was balanced in the past for development to need both but now it’s a total hyperactive victory of general knowledge, master of none.
Specialists will be fed to the model weights, generalists will inherit the world.
> ADHD wins over autism in the immediate future. It was balanced in the past for development to need both but now it’s a total hyperactive victory of general knowledge, master of none.
That’s an interesting framing. I’m not sure I agree with your strong conclusion but that’s mostly because I think there’s a second axis about maintenance – you can make a grand strategy game in a couple months but it’s harder to make it balanced, fun, and to keep shipping updates which players like. A prototype or hobby experiment doesn’t have that problem nearly as much and I think that’s where we see such bimodal reactions where some people think it’s a huge productivity boost and others see modest or even negative results.
I think we’re on opposite sides but this is the first time I’ve seen someone in the wild share my suspicions that the LLM conflict is part of some vast ADHD vs autism hyperwar. At one point I was collecting posts I found where LLM boosters explicitly mentioned how it interacted with their ADHD.
The contest you described (people with ADHD vs. autism) would have an enormously negative effect. But there are reasons to think it's unlikely. FWIW at least 25% of ADHD adults manifest autistic traits. Hardly surprising, clinicians regard overlapping neurodevelopmental disorders as the rule and not the exception.
An online community works when participants learn and share knowledge gained. Contests, with self or others, disrupt learning and sharing, hence by definition constitute "bad behavior". If it exists, the hyperwar you allude to will harm communities more than all LLMs combined could do.
I'm pretty skeptical of people who are ADHD and claim to be getting boosted by LLMs.
It's hell for my ADHD, personally. It's an infinite rabbit hole generator and distraction machine. I have to use browser extensions to filter out stuff like Stack Exchange suggestions for interesting topics, too. I have had to block Wikipedia in the past to avoid rabbit holes there
I'm better now than I used to be but I am skeptical of people who claim to both be good at multi-tasking and also have ADHD. My experience with ADHD is that multi -tasking is not a strength, it's a sign of distraction
> ... multi-tasking is not a strength, it's a sign of distraction
Interesting point of view from one who's been there. Intuitively the comment makes a lot of sense.
In a way human multitasking resembles how single-threaded computer programs work. That is, we focus on one task at a time but tasks can be interleaved to give an illusion of doing several tasks at the same time. Managing several tasks requires remembering the next step to do when leaving one task to attend to another.
That suggests ADHD people lose track of where they left off when performing several tasks. Distraction plays a role, but I think problematic working memory is the major issue. Among ADHD researchers there's a saying, "people with ADHD know what to do but don't do what they know." ADHD task performance may improve with proper treatment.
Not diagnosed, on online “tests” I score just shy of the ADHD threshold, and I find myself in this. If only I properly remembered and agreed on what I wanted to do yesterday it would be fantastic. Even if I write it down and happen to remember to check the notes, it’s almost like I was a different person yesterday, todays priorities are all new
> Not diagnosed, on online “tests” I score just shy of the ADHD threshold, and I find myself in this.
Don't know the tests you refer to, but it's recognized that many adults have ADHD symptoms not quite meeting DSM-5 criteria for ADHD. It's called "subsyndromal" ADHD which causes as much disability and responds to treatment just as well as "full" ADHD cases. IOW it's absurd to regard someone with 6 symptoms as belonging in a different category vs. a person having "only" 5 symptoms. Anyway, ADHD is a condition that usually responds to treatment, no reason not to check into it.
Thanks, I know I could probably get a diagnosis and treatment for it. However, to be honest I kind of like it!
I think it keeps me on a different path wrt most people, so until it becomes a serious problem (financially) I’ll keep riding the waves of focus and distraction ;)
> That suggests ADHD people lose track of where they left off when performing several tasks
Not quite. At least not for me. My problem is hyperfocus. If I switch tasks I might go down a rabbit hole and forget to change back. It is similar to being distracted, except the distraction can still be legitimate work stuff. My focus is like a freight train, it's hard to change tracks
Which can seem like an attention problem when you're trying to change me from focused code mode to "jump on a call" mode
Yes, hyperfocus is indeed a frequent problem for people with ADHD. One version is getting overly involved in irrelevant details, due to interest or having difficulty pulling away. The latter takes much effort and a good measure of courage. Maybe "rabbit holes" are kind of like that.
For many with ADHD starting tasks is hard and resuming a task can be just as hard. Constant experience of "starting from scratch" is burdensome, and contributes to inertia and inefficiency. More than not, a tough problem to solve.
I've long thought disability associated with ADHD is generally underappreciated despite decades of study proving the point.
Self-observation isn't easy. I always considered it a sign of growth when people shared what they've seen.
For me it's a dopamine machine. I get my doses of lacking dopamine from completing things.
Agents let me complete things in parallel all the time.
They also let me build tools that manage all the attention-grabbing things in my life. Something I've been meaning to do "tomorrow" for a good decade.
I've already built custom notification tools for all the "I'll just quickly check hn/lobste.rs/tildes/whatever" so that they'll tell me if there's something actionable, closing down the rabbit hole for me. My email is pre-filtered by an agent that pings me if there's something urgent. My IRC client is custom-built for the specific way I use it etc etc.
No, it felt too invasive and creepy so I abandoned the idea. Or my lack of follow-through is a sign that I’m actually on the other side of the “war”. Reader decides! But suffice it to say that it was, to me, a significant trend especially around the “Claude Code moment” back in February/March is this year.
Autism should not be framed as being "opposted" to ADHD. The correct axis is "Autistic" (Neurosis) as opposed to Psychosis. Opposite poles but with a little bit of horseshoe theory dynamics
> This way one person can implement a whole new hobby project instead of only writing a single specialized part in an existing project.
I still do start from bottom up to refactor everything into workspace members and this way, parallelism is way easier
> LLMs reward seeing forest instead of trees.
Not just forest. LLMs reward seeing structures and designs, and it still requires a lot of creativity to design it well. I have to do a lot of architect works for that, like package layout, testing strategies (BDD vs TDD vs E2E), setting up access to platforms and digging out what strategies to apply...etc.
> People focused on the trees will however lag behind as implementation details matter less and less.
Trees like demonstrations and example code are even more important. This is because those are literally post-training materials for the LLMs
> Specialists will be fed to the model weights
I don't think so. New labor laws are certainly going to be introduced by the politicans to tackle this.
> generalists will inherit the world.
Now the problem is how far do you want to stretch the systemic thinking, and you also get burned out quicker.
Funny you should mention autism. I'm of the belief that if I really wanted to, my autism could help me become a very effective AI-assisted programmer, because autistic communication traits that actual humans find annoying (and which most people hence train out of themselves), like infodumping and using precise language, are things AI actually loves and produces better results with.
But I don't want to use AI to code, in large part because I iterate on ideas by working directly with code. Attempting to program through an LLM intermediary becomes like attempting to eat using soft silicone chopsticks.
Exactly. I have a backyard garden not because I need the food (it would be infinitely cheaper and easier to buy even high end farm-stand vegetables), but because I enjoy gardening. Buying a gardening robot (if such a thing existed) would completely defeat the point.
I use an LLM to make custom software for myself not because I enjoy coding but because I want software that I can't afford to hire a human to make for me. Buying a robot to do it (current agentic coding LLMs) is the only way it gets done at all.
These two things are not the same, and should not be viewed in the same light.
All that being said, I would 100% buy a weeding robot. Weeding sucks.
Sure, that's fair. I enjoy the process of figuring out ways software can solve some (usually quite trivial) problem for myself, but then I'm very happy to hand the act of making it to a robot.
It depends what you value. There's "zoning out going through the motions" for mindfulness and then there's "building stuff!" for interest, learning and fun.
And you can have any combination of the two. Like maybe doing the same planting routine but with different details for a different plant or experimenting if a new layout works better.
I don't care to do the labour of gardening. But if I had time, I'd spend it bikeshedding optimisations. IE what about grow lights? Is it worth having a greenhouse in my climate? etc etc. And all the little details about building stuff.
Aside: I too want a robot. One for lawn mowing (they exist but expensive) and one for spraying the little lawn weeds that pop up.
>> I have a backyard garden not because I need the food (it would be infinitely cheaper and easier to buy even high end farm-stand vegetables)
...cheaper how? My partner and I grow most of our veggies and, based on our calculations (including fixed costs that have gone into the garden, as well as e.g. water, fertilizer, etc.) we save a non-negligible amount of money every month.
One can piss away a lot of money gardening with little to show for it, especially if you buy bagged soil. Buy a raised planter, some soil, some fertilizer, some seedlings, a couple garden tools and you can easily be in for $150+ which is definitely more than you'll get out of a season worth of growth even if nothing fails.
On the other end of the spectrum, if you grow in fertile native soil from seed your costs can be close to zero
Yeah, in my experience, the people that got into software engineering because of the money are absolutely thrilled with how LLMs are changing the industry. They mostly just liked getting results and LLMs are great for that.
The people like me that sort of just stumbled into this career as a result of spending the evenings of our youth trying to get the computer to do interesting things don't seem to have as much appreciation now that the LLMs are taking over the fun part.
If I'm any example, I'm on the other side. I got into computers in the '70s: plotter art, BASIC, then Z-80s, forth, PDP11s, Amigas and so to now with GPUs and supercomputers and visualization. I'm recently retired and having enormous fun plowing through my lifetime backlog of "wouldn't that be a cool tool" (https://pcons.org) or a climate visualization (https://globe-viz.oberbrunner.com) or a travel packing list app (packzen) or a bunch of 3d printer tools... so on and so on. Totally co-engineered with Claude. What we're doing together, and getting out to the world as useful open source projects, is easily 10x what I could've done on my own, in the limited time available to me in this life.
> Yeah, in my experience, the people that got into software engineering because of the money are absolutely thrilled with how LLMs are changing the industry. They mostly just liked getting results and LLMs are great for that.
Those people are in for the biggest shock and largest disappointment: when a bright teenager can replicate, with an LLM, what they produced with an LLM the pay is going to fall off a cliff.
At least those who really are into development to the point of writing their own OS, Language, etc can still enjoy it as a hobby.
> Those people are in for the biggest shock and largest disappointment: when a bright teenager can replicate, with an LLM, what they produced with an LLM the pay is going to fall off a cliff.
The value of an engineer is not what task they work on, but what problems they can solve. To this teen, the code is a black box, the LLM has most of the value in their pairing, and another random teen will replace easily them.
LLMs were trained on the work of engineers, not prompting teens. The profession will continue to evolve, and the LLMs will improve along, thanks to professional, manual code still being written, not to vibe-coders.
I don't think that's quite fair. Computers have been my hobby since I was 4, I'm having a blast, this is so much fun. I've always had way more ideas I wanted to try than I've had time/energy to implement, especially true since having kids, now the implementation can almost keep up.
Computers were on a trajectory of becoming more consumption oriented, and this has shifted it strongly back towards creation. Normal people can make their own custom software again. This is awesome.
Really hate this characterization, although to be fair I love the ops/infra side of things more. I've used to set up so many parts of my homelab manually, but now I can just get it all done.
I've spent countless hours maintaining a Rpi server at the infancy of my interest in computing.
Truly a lot of it is maintenance and busywork tasks. I am having more fun than ever. Perhaps some people like those busywork sysadmin tasks and gain enjoyment of the homelab hobby just from that...honestly I can't say I really understand it.
That attitude isn’t universal to everyone who enjoys tinkering and making computers do interesting things. I just had lunch yesterday with one of the most tinkering people I know, who as a side hobby teaches kids to program using robotics etc. to make it interesting for them. Like me, he’s having a blast using LLMs.
LLMs are amazing tools for learning new skills and technology. I’ve recently moved into an ML adjacent role, am learning tons that I never knew before, and I’m delivering working systems more quickly.
One thing that’s been an eye-opener is how many details I miss in the code I write. LLMs can almost always suggest improvements, and usually, it’s quite obvious that they’re right. But this is making me a better developer, even though I already have decades of experience.
LLMs don’t force you to vibe code everything. You can use them as precisely or as vibe-ishly as you want. They can be a skill amplifier rather than a skill replacer, if you use them as such.
Not my experience at all. I taught myself assembly language when I was 11 just for fun, and I'm still programming as a hobby more than 40 years later, even though I also do it professionally.
And I love using LLMs. They take care of all the boring boilerplate, letting me focus on the parts I actually enjoy.
I enjoy optimizing code to the point of obsession, but I can't find economic (daily-job) justification to spend time on this. Lately I've had several extremely productive and enjoyable sessions that go like this:
- I run Claude on my desktop PC, explain the problem, and let it explore the code around the problem and build a test harness so it can run experiments
- I ask it to run experiments and try things; at that point I can leave my PC and continue over my phone in brief stolen moments here and there, over the course of many hours
- it comes back with results; I ask it to try new things based on those results
- after multiple rounds of this, I have explored a vast search space I could have never had time to do myself
- I ask it for an implementation of the most promising path; this is usually something simpl-ish (on the order of 50-100 LOC, not huge new systems)
- I return to the PC and do something between polish the implementation and redo it fully manually - partly because I enjoy it
This is something I can see myself do for fun in areas I care about, not for money and reasonable business objectives for my day job.
My hobby is making things. I used to make them with a screwdriver and a hammer and a handsaw. Then I bought a drill and got some screwdriving bits, an a pneumatic hammer and an arm saw. Now I can build things much much faster.
Similarly I've made lots of software projects. Over the last 15+ years I have over 250 github repos, the majority of them no one but me will be interested in but I had fun making stuff. LLMs let me make stuff faster just like the power tools. I'm still enjoying making stuff. I still direct, I still decide what to make, I still decide features, iterate on UX, etc, but at several X the speed of before. Loving it.
I think for those of us who got into software initially for the end result -- being the only way for the thing in your head to ever exist -- enjoying the engineering part was a nice, incidental surprise.
So while LLMs solve the technical implementation side, there's still upside since now we get to materialize the dang thing we were setting out to build in the first place.
Though there's a huge sense of loss since so much of our skills and identity and time and value were devoted to the engineering side.
I enjoy molding hyper specific software across my whole eco system of tech through prose with LLMs. Iterating over time to hone the right solutions based on very abstract reactions to the results I get. I would consider this a hobby because it costs me money and I get no value out of it outside of the act.
I also woodwork and I remember this discussion there between hand tools, power tools, and CNCs. It's all the same hobby just different entry points and doing it the least automated way doesn't make your work more or less than anyone else's.
> I would consider this a hobby because it costs me money and I get no value out of it outside of the act.
I LOLd very hard at that one, thanks. Very relatable.
A hobby of mine is also woodworking. I can make jigs, I can buy jigs. I can use hand tools, I can use power tools. I can use screws I can use joints. I can even buy made furniture and just assemble it. Though the latter is more of interior designing. But the point is it is a spectrum and everyone has their own niche. SDLC are not the same and there is different value to be derived. But the process of woodworking makes me handy. The experience allows me to elegantly and cost efficiently fix broken things. It enables me to do interesting things in interesting ways and there is more of a community around that. If I automated away all the woodworking and just bought the furniture... Well there isn't much of a community around that except flexing who has more cash. Idk, I'm trying to have empathy for both those who are for and against it.
> LLMs expedite achieving the end result. Take from that what you will.
It's an age-old debate in every hobby.
Some people are in it for the process. Others are in it for the end result, which is what they want to use or play with.
In woodworking there are some circles who will sneer at you for using certain power tools instead of hand tools. I do think it's more impressive when someone makes something entirely with hand tools. I don't have all the time in the world to work on a project, so I'm going to use the power tools when I can.
Some people care a lot about impressing others with how they made a thing. Some times that's appropriate, like if there's a competition for hand-made or human-written something. Other times, nobody cares. They're getting together to enjoy the end result, not the artisanal process used to make it.
> Some times that's appropriate, like if there's a competition for hand-made or human-written something.
Indeed, and this is the problem: unlike woodworking, where my power-assisted creation is independent of, or complements, your handcraft, this engine is directly competing with other engines. Twiss is being disingenuous when he says he has no interest in submitting it to competitions; it plays on a rated account on Lichess, for example.
Now that we can mass produce tables and chairs I would imagine there are far fewer people whose career is hand crafting tables and chairs. So machines literally did out compete them.
In fact handicraft people were out competed so hard many people forget that used to be a respectable career.
And to extend the analogy, before there were house cleaning machines the people who enjoyed cleaning were lumped in with the people who like having a clean house.
I will say that there are a lot of parts of programming I enjoy, but I used to spend an inordinate amount of time chasing down obscure errors or learning arcane systems just to get to the parts I personally enjoy: the feedback loop of making a thing, seeing the result, and improving it.
There is a reason stack overflow is dead. LLMs are a strictly better product.
I think this is why that I limit the use of LLM in my hobbyist project, but not in my work. I found LLM particularly useful of generating say VSCode configuration JSON files and it saved me a lot of headaches. It is also very good at finding the right tool for the job (e.g. which binary to use to write a floppy image).
What is the "end result"? You refer to cleaning here, but many people
use AI spam slop to bombard people with garbage text. That is not "cleaning
up", that is literally wasting the time of real people here. And that's just
one example of many more that can be given of AI slop not being so great.
What’s missing here is an understanding of the negative effect the rise of AI has had on programming communities overall: both the decrease in positive engagement as people have moved from discussing things and helping one another online to chatting with AI, and the increase in negative engagement. The amount of low-energy abandonware being “shared” into these communities makes it really hard to filter out the signal from the noise.
So at least in my experience, it’s not about gatekeeping as much as it’s about keeping the experience bearable for the existing hobbyist community.
To be fair, the decrease in discussion and helping one another online has also decreased some significant amount of negative engagement in the form of nonsensical arguments and frustrating ways of responding to questions (e.g. the biggest complaints with stackoverflow).
It was like this with cryptocurrency to a lesser extent too. Suddenly lots of people pretending to have technical background and explaining what a Merkle tree is.
And telling everyone if they don’t embrace cryptocurrency right away they will be left behind, how you are all limiting human flourishing by refusing to embrace the future, excusing every single negative side effect of the technology — because we have to move as fast as possible. And of course no words on how unsustainable the businesses are, how corrupt it is to have so much power and control into a few really shady individuals
Is it gatekeeping in this case? Would they not allow beginners to join and learn? Are they not fully sharing the knowledge to anyone who's interested to learn it?
To me, it amounts more to disapproving showing up to a handmade sweater community and showing off your machine made sweaters, or maybe a machine made sweater you added a handful of stitches too yourself. It's like not realizing that the community wasn't about cool looking sweaters, but about hand-made cool looking sweaters.
Having said that, there might be another community that's more about cool looking sweaters, and new machine techniques that deliver really nice looking sweaters would be of interest to them. I actually think it's what is happening a bit in programming, there are niche that remain more about the manual programming, and others that are about any process that can lead to better outcomes.
The term gatekeeping gets thrown around so much now that I feel like ive forgotten what it even means. What they are doing is having rules. All communities above a certain size need rules to maintain structure. Idk if that is or isnt gatekeeping, but it is definitely using rules to maintain their identity.
The entire AI-pilled side-software development has a single goal: shipping as fast and as much as possible. The problem is that it’s not better code or a better product—it's just more
I can understand why people who aren't interested in "shipping daily" and are more interested in figuring things out are kinda annoyed by all the LLM noise
Imagine, if you will, wanting a narrow application to assist you personally in some side hobby, that is data driven but would take weeks or months to set up. Now imagine it's done in hours (and not even purely focused hours) and does everything you need it to do in an instantly verifiable and rewarding way. Imagine doing this on the regular for all of your various use cases. And then imagine not trying to sell or market that tool, as you dont care about that, you only care about solving problems and empowering yourself.
The vast majority of developers are like this, and as such, are silent about it. They aren't making production applications. They're making bespoke solutions and iterating upon them over time.
Somehow purists think there's something wrong with the process of that. I say there's something wrong with them, perhaps on the spectrum of OCD. With a tinge of arrogance or narcissism thinking their applications are all of the purest order and construction, testaments to architectural beauty and edge case handling. With a tinge of fear and cope as so much of their identity is caught up in being a smart coder rather than a smart engineer.
If you are a software developer and don't see the value in and heavily utilize agentic tools, you are a poor engineer. Not every coder has the mindset of an engineer, but a coder should be able to recognize and understand a powerful tool when they see one instead of letting their emotions and irrationality take the wheel
Nothing says engineer mindset like calling people "something wrong with them, perhaps on the spectrum of OCD. With a tinge of arrogance or narcissism" and "emotions and irrationality"
I think you missed the entire point of this discussion, but you also have a very gross view of the world!
> Somehow purists think there's something wrong with the process of that. I say there's something wrong with them, perhaps on the spectrum of OCD.
Are you really saying it is a sign of serious mental illness because programmers want to enjoy their craft? You do realize our entire industry exists because of the passion of programmers years prior, right? I genuinely think you should be ashamed of what you wrote here.
I think you misunderstand. The entire industry exists because of the passion of engineers that program, not just programmers. And i specifically called out 'purists' for a reason, and yes I do believe OCD and perfectionism is a component. It would help if you read my comment critically rather than purely adversarily.
Well of course, if you enjoy writing software as a hobby then you value the experience of doing it, and you value the skill involved. You probably want a community of like minded individuals who all write code by hand, and where LLM use would be considered cheating.
That seems as natural as being against doping in sports or marking cards in poker.
None of this is anything against LLMs or AI assisted coding in general, even though the people involved probably have strong feelings about it given their preferences. It’s just the preferences of particular people and communities. Seems reasonable to me.
Yeah, it's just how you see programming in your perspective.
It's like lifting weight. Obviously, machines lift weight much better than human for a long time. But people still trains for strength or aesthetics. Even if machine can lift much larger weight and more efficiently, some people still care about how strong they are.
A machine can lift much more weight obviously is not the excuse for them to stop training.
Programming is just the weight lifting for your brain/mind. Even if LLMs can solve all coding problems perfectly, there will be people who still want to program all by hands, for training their mind, or simply enjoying the process. And there will be people care about how good you are at solving a problem with programming.
LLMs being able to code does not mean you no longer need the ability to code.
Cars do not mean you no longer need the ability to walk.
I don't exactly disagree, but I don't think your analogies work.
Machines can lift more weight than humans, but we all encounter situations where strength is useful and machines aren't available or wouldn't be suited to the task.
Cars are faster and easier than walking, but they require roads and are not a possible substitute for all of the walking we do.
But "if LLMs can solve all coding problems perfectly", and assuming that's sufficiently true of cheap/free models, there will be no practical need to code by hand; it'll be something we do purely for the joy of it and/or as brain training (but a niche form of brain training, probably practically inferior to other forms that target skills we still need).
> these communities don’t care if your code works at all, but instead care that you know why and how it works.
not literally true, and I doubt the author meant to say this. The communities do not only care about if your code works at all, but also that you know why and how.
It's sort of like in math classes where you need to show the work, not just the answer. Surely the answer being correct, while not being sufficient, is still necessary.
I was somewhat conflicted seeing the way the OSdev community responds to LLM usage.
I understand the conflict: the point of doing OSdev is that you're learning the low level hardware details and directly commanding the machine from the lowest level accessible. LLM usage is at odds with this.
But, on the other hand, the primary issue with OSdev on modern hardware is the complete lack of documentation for most hardware. A lone developer simply does not have enough time to reverse engineer from Linux or BSD drivers to achieve much hardware support. Plus, often you run into hardware quirks that are incredibly painful to debug and offer no real learning opportunity.
Recently I revived a ~8 year old modular kernel project I had written. It was at the point of being able to setup an x86_64 CPU, load some drivers and enter usermode. Using an ESP8266 board I setup a little debugging harness for a AtomicPi devboard I had laying around, and set it all up so Claude could control and upload code to it. It could read the serial port, ping the board over ethernet and power cycle it. From that point I let it work almost autonomously to implement a USB stack, finish the network stack, implement drivers etc.
It worked well, and for the most part I didn't care about the results. I don't have the time to commit to such a big project now and of course, the satisfaction of doing something myself (besides building the harness) was gone.
But a thing that stood out to me was, when trying to implement the ethernet driver, the chip in question did not have documentation, but it was known to be similar to another well documented chip. Claude was able to spend a few hours just trying different things, rebooting the board as needed until the differences were sorted out and it was able to ping the board. I feel this kind of application would be pretty relevant for even hobbyist communities like OSdev.
We all have limited time, and while LLM usage for everything certainly defeats the point, targeted usage for tedium that has very little value overall, should be acceptable, and I can't help but think lesser of devs I otherwise respect, when they insist on being dogmatic even when it doesn't contribute to the hobby.
There are different kinds of tedium that LLMs automate, being a hobbyist doesn't mean you see value in all of them. Removing tedium that you personally find uninteresting or boring actually accelerates the hobby because now you're able to get past the tedious parts instead of giving up or delaying doing them. That does not mean that everyone finds those parts boring.
Even before LLMs came along a lot of hobbyists were using high level languages, linters, IDEs etc. instead of handcrafting assembly or machine language. Those tools increased the capability of hobbyists to build what they find interesting.
> Even before LLMs came along a lot of hobbyists were using high level languages, linters, IDEs etc. instead of handcrafting assembly or machine language. Those tools increased the capability of hobbyists to build what they find interesting
Yes and many programmers who came up in the 80s or early 90s writing C or equivalent decry these tools as ruining XYZ thing. LLMs are the same effect just on steroids
Isn’t that the point of a hobby? To enjoy the path and not the result? To learn something new.
I understand using LLMs as tools to help you with your hobby, but I think when “Claude spends hours debugging it” it probably is no longer a hobby, but an outsourced job.
That's why I was trying to make the distinction between stuff that actually pushes your understanding forward and teaches helpful information, and stuff that does not.
Isn't it more in the spirit of osdev to spend your effort implementing/working on the things that actually further your OS meaningfully? Discovering that a specific version of an obsolete undocumented chip has a quirk where one register doesn't reset in the expected way isn't really useful information and doesn't contribute to building new skills.
Maybe what I'm trying to say is that not everyone enjoys every piece of the hobby, and I don't think it's reasonable to exclude people who do not enjoy some of the pieces that aren't even core to the hobby.
You can summarize this roughly by comparing art for its own sake versus art for a purpose. If I need software for a purpose, I don't care how it is crafted, only that it is crafted well. But if I care about the art of code done well, then obviously the means matter.
I find that a lot of "pro-AI" folks are very much doing software for the ends (myself included) and in the past, we were happy with the "means" primarily because there was no other way. The "anti-AI" folks are, as mentioned, mostly in the camp of caring about the means equally with the ends.
Lots of nuance here, but those buckets seem to generalize pretty well (so far) in my experience.
Very interesting stuff, and I believe the prejeduce against AI tools for use to code stems from the way they are currently package. The tools default out of the box are designed to throw away the most valuable process humans go through while learning, FAILURE!
If a human is never able to work through or process failures themselves they are bypassing one of the most important steps in the learning journey
More AI tools need to be made the opposite of this, humans should be able to achieve great learning USING AI.
The important message also remains, understand your code! It might just save you one day ;)
>even in the instances where there was earnest early engagement with LLMs in some of these niche communities, the well was quickly poisoned by a combination of a lack of a deep understanding by the LLM practitioners, and a vitriolic subset of those communities that view the LLM enterprise as a form of cheating
Yeah, the whole point is the street cred of doing something really hard.
Whether that chess repo is genuine or not genuine, legal or illegal, tasteful or tasteless, vibed or not vibed, stolen or copied, the signal to open source maintainers is clear: get off GitHub.
When you take away the cost people will treat the result worthless.
For example if you hiking in the mountains, reaching the summit and enjoying the views are valuable experiences to you because it had an associated coast, i.e the physical act of hiking up the mountain. If you take the ski lift to the top it's almost worthless and you have very little appreciation for the views.
Very apt comparison! This is how I feel too. In the sense that when I’m coding as a hobby, especially something like advent of code, using an LLM would defeat the purpose for me. A bit how I don’t double check my moves with an engine when playing chess online.
On the other hand though, if I’m being paid to build, say, the top station of a ski lift on a peak, I wouldn’t be surprised if the contractee would prefer I don’t try hiking up the building materials and used a helicopter instead.
I disagree with the conclusion. I think this is only half of the story, and I can say so as one of the people in question who is largely against LLM code (despite finding LLMs to be very useful tools). While I do care about genuine care in one's craft, I also am a free software advocate and fundamentalist, and I find non-trivial LLM code is too legally and ethically dangerous to be included in software that is distributed to other people (do whatever you want if it's just personal use).
"Aggressively Against LLM Usage" is the wrong way to think of it.
If my hobby is running, I would refuse to use a car to do my morning loop. Not because I have anything against automobiles, but because that's not my hobby.
Also, the issue linked to isn't really about LLMs. It's about someone being a dick in an online community and being called out for it. They just happen to be using LLMs to be a dick, but that's not the main issue.
I've found myself with this problem, as I've used LLMs to accelerate solving problems I've wanted to solve, only to find myself unable to share them back, because LLM usage has "polluted" my effort in the eyes of maintainers, even if the result has been something I've re-implemented by hand after learning from the agent-driven development what the issue is and how to reproduce it.
What's the solution, to just disengage entirely and not share back issues I find or features I'd like to see?
I get the negativity, because I enjoy hobby programming too, so I also enjoy the process sometimes. But then sometimes there's a problem and the fastest way to isolate the core of that issue is to use claude.
A long time ago, I had to take an ethics class to get my Computer Science degree. One of the things that came up in the papers about calculating the statistical value of a life was that people care more about being more prosperous than their peers than absolute material wealth.
I think about that a lot: its not about material wealth, but status, something that if someone has more of you have less of. A lot of people are invested in their status in various communities, and are afraid to lose it or have it become meaningless.
I don't know what to do about that; no matter how wealthy we all are people are still going to want status.
Well yes. That's why you see many westerners complaining about the cost of living even though all things told they have a good life.
I would modify this slightly though. I think there's an expectation of things getting better, and expecting to do better than your parents. I think that's where many of the modern issues around living standards come from. They aren't getting better (quickly enough).
I'm a hobbyist, what am I missing? LLM gave me a new spark. I never managed to get a SE job so some weekends I tinker around with LLM for small personal projects and it helps me learn more.
I think that’s the difference. Using LLMs to learn how to write interesting code, I hope, would be welcomed. I’m not confident that’s true though, reading some of these comments.
I don't think anyone is learning to code by using LLMs.
We learn a craft in two ways: by watching other people do something, and by doing it ourselves. I've learnt many crafts in my life, to the extent that I've produced useful or enjoyable output: programming, wood working, gardening, cooking, baking. I read loads of books, but nothing you read in a book makes much sense until you do it yourself. If you read a book on wood but never pick up a saw then you might as well be reading fiction.
Programming has always been a bit more difficult because we can't really watch someone else do it. Sure, you can watch someone use their tools, but programming really happens in the head, and you can't watch that. So you very much have to just do it yourself, and read.
When you use an LLM you're not watching it do anything, except output text at a very high words per minute. And you're certainly not doing it yourself. So how can you be learning? I actually don't think I've learnt a single thing since starting to use LLMs and it seems that way for colleagues too. They're producing code they never could have written before, but they don't understand it. They aren't learning.
I’ve learned tons since I started using LLMs. They’re an amazing tool for learning as long as you use them like that. Ask it questions about things you don’t understand. Verify the answers. Don’t have them do everything for you. Figure out where to push back on what they’re producing.
When I learned to code, learning from existing code was a big part of it. One difference now is you can produce custom code to learn from.
I’m starting to think the biggest skill we need to teach people in this new age is how to learn and how to research. If you already know how to do that well, LLMs are a fantastic addition to the toolbox. If you don’t, then you’re stuck with a tool that’s going to replace your skill instead of augment it.
Like I said, you can ask an LLM about wood all day but if you don't pick up a saw you won't learn to make anything. I do, of course, use LLMs for research and learning, but I'm specifically talking about learning to program. LLMs can't cut the wood for you, but they can write your program for you. If you can resist and just use them to learn then sure, but you've got to do something yourself.
I'm against LLM usage in conversation, e.g. when discussing issues in open-source. It's like half-hearing a conversation, I'm missing the context that the other person has with their robot. Second, the person has usually at best half-understood the issue.
Provided the code does the job, it seems immaterial whether it was hand-written or LLM generated, provided it does the job and isn't straight-up plagiarism. The latter is admittedly a hard thing to judge.
Using LLMs for hobby programming is like using a computer to solve sudoku puzzles. The whole point is to do it yourself.
Personally, I find that when I start thinking "hmm, I wonder if I could write a program to automate this" is when I quit the pursuit. After figuring out the basic tricks to sudoku, it turned into busywork. (Or trying to micromanage StarCraft)
(Tagentially, if you want to automate programming, then I'm going to find myself a different field, because I have no interest in the results of sudoku puzzles. I especially have no interest in managing agents, which I see as junior-dev bullshitters, and would not enjoy managing such a team in real life)
> Using LLMs for hobby programming is like using a computer to solve sudoku puzzles. The whole point is to do it yourself.
(I absolutely relate to your experience with sudoku etc. Of course, writing a basic sudoku solver is also interesting.)
The thing is, software architecture and systems design can also get the hobby treatment. I've found that I'm happy to have a short discussion with ChatGPT and have it barf out a few hundred lines of a concept, then refactor and maintain it myself. Or come up with a few different pieces that can be written fairly easily, let the machine skip over most of that tedium, make sure they conform to the interfaces I had in mind, and then verify for myself that the system actually comes together the way I imagined.
Of course, it does help a lot that I can understand the code, and work with it line by line if necessary. But SICP teaches us that it's abstraction and composition all the way down. So I feel like I'm exercising the same skills while not letting myself get intimidated by the sheer amount of initial typing or the self-doubt (instinctive immediate review of the code, you know) that creeps in when I try.
The interactive fiction community seems to have a subset of hardcore anti AI hobbyists. Partly at stolen artistic IP, but also the coding half of text adventures is viewed as art just as much as the story and writing.
I’ve selected a path where I think building tools with AI is fine. But I want people to play my games knowing I created all of it.
For me, domain discoveries and elegant solutions arise out of periods of deep befuddlement, and are often a saving grace that can avoid lots of code.
I'm lucky I guess that I've had enough time in this career to experience these moments, and enough experience to know which code I should write and which code really is rote and can be generated (most of the time). Hopefully it keeps the loss of enlightenment to a minimum.
> In these communities (keep in mind that there is an implicit “not all…” throughout) the process of mastering a difficult field itself is the product, and something that runs is generally a nice-to-have.
Most hobby fandoms have a dedicated core group who think like this: The end result is only an artifact or social proof to them. The real product is their own prestige of having built it and attracted an audience. There's always an IRC channel or Discord somewhere where there's a social hierarchy of who should be respected for their work.
Outside of those inner circles, most people don't care. The average emulator user doesn't care about the details about who wrote their emulator or the history behind it all the forks. They want to play a game and the emulator core that gets the job done, because the emulator is the product to them. The mastery of the author is not a product to them.
This has happened in multiple technical hobby communities I've been in. With 3D printing I even remember communities where Prusa was looked down upon for making 3D printers too accessible. The argument was that now anyone could buy a 3D printer without really understanding the pain of mastering calibration, fine-tuning, and troubleshooting. That debate exploded to another level when Bambu came along and made 3D printing even more mainstream.
Now it's happening with programming communities. There are some where hand-written code is truly the actual product: Competitive programming, code golf, and obfuscated programming competitions come to mind. Using an LLM here is about as appropriate as submitting LLM generated art to an art contest, which is to say it's not.
There are a lot of other domains where the mastery of the craft isn't interesting outside of a small social circle. It's interesting that they mentioned emulators, because emu communities are famous for having a ton of strange drama that most emulator users don't care about. If you're inside of the small community and spend every day on their Discord, it might be the only thing you care about, but you don't realize that you're a minority because everyone else in the Discord thinks the same way.
OSDev is another interesting one to mention, because even Linus Torvalds had to make a big post telling the anti-LLM people to chill out. There seems to be a split between the purists who think the primary value is their own mastery of being able to write the code, and the pragmatists who are adapting to use new tools in conjunction with their knowledge. If you find yourself in one of the purist camps you might think everyone feels the same way, but outside of that, the world is moving on.
I usually feel turned off by communities where social status is the primary currency. There's always a hidden social hierarchy that feels threatened whenever it feels like the rules my start changing. They react by trying to force the social currency back to the system where they spent years investing their time. Nobody wants to feel like you're losing a foothold because the rules are changing. I got a strong feeling from this article that this author feels they've developed some level of respect in some forum and they dislike the idea of that investment being devalued by a changing landscape
> In traditional niche dev circles, respect is earned slowly through years of activity in their respective fora,
> At the end of the day, these communities don’t care if your code works at all...
I think the key word to all this is "communities." If you can make something alone, just do it. If you enjoy gathering together to build a bigger thing, often the bigger thing is 1/3 an excuse to gather together. In that way, the LLM is like revealing the answer to a murder mystery game too early and expecting it still to be fun for everyone.
I'm particularly annoyed with the complaint in the linked thread about the agentic-developed chess engine.
> Moreover, you are giving absolutely nothing back to the community, which is central to open-source engine dev. You are only taking.
What would that user prefer? That the developer push a PR upstream...that presumably would be shot down and the user subject to further harrassment because it would be LLM derived at minimum? It seems like a catch-22 where the only winning move is not to play.
That thread also highlights the overlap between "people who are against LLM usage for coding" and "people who are very particular about AGPL."
Presumably no one is forcing the maintainer at gunpoint to have an LLM slightly shake up other people's code, relicense it under an incompatible license and then claim that "every single line of code is written by Claude Code", so I don't see the issue with not playing if that's the only option.
If you can't make contributions that anyone finds interesting enough to include upstream and have no plans to publish your AGPL derived code under a compatible license, you can keep your changes to yourself. Nothing lost, nothing gained, license honored.
Is this novel? Every group that spent some large number of hours gaining some skill hates when that duration is not required for newcomers to accomplish what the skill allows. Not every individual of that group but the group as a whole. It’s just that devaluation of one’s skill capital sucks.
I happen to think the alternative world where no one’s skill capital can be devalued sucks more, but it’s hard to build a community around that because the same hours that build skill also build taste, and when you remove the former you remove the latter.
> Every group that spent some large number of hours gaining some skill hates when that duration is not required for newcomers.
I'm not sure of the truthiness there wrt LLMs: using an LLM to code your hobby OS doesn't actually gain you the skills of the existing members of that niche. Maybe those existing members are just annoyed at posers, not at newcomers?
Sorry, I didn’t complete the thought well. Amended. I meant that the skill allows for some things and newcomers no longer require that investment to get the outcome.
I think another reason to write code by hand is that you want to know what you get.
If you have an idea, then you can implement it better than an LLM, just like you can implement it better than another person, because you are the only one who understands what you mean exactly.
Details are always lost and misunderstood in communication; this is a weakness of LLMs.
Similarly, even if I don't write the code but I let someone else who I trust do it for me, there is value for me in that person not using LLMs, because only if they don't, I can know that the only spirit that lives in the code is of the person of my choice.
It's like a social thing; I can know and trust a human, but I'm not sure I can do that with an LLM.
It's similar to how you wouldn't want to use LLMs in politics: the whole point of expressing your ideas is that they are your ideas and if you let someone (or something) else do it for you, details will be missed and misunderstood.
So this is one reason that I think hobby communities don't like LLMs; they care too much about what they build.
Depending what you are making; coding slows you down and think as you build stuff step by step. Its kind of like speaking vs writing. Generating code through LLM for me is like talking -- I love talking, but when I write something down, I'm more structured, think more about things etc etc.
People say, if you you write elaborate well spec'ed prompts, it will do the right thing. Coding in a way is doing that.
All of this is exactly why I started Slow Code. I like programming. Lots of people like programming. I don't want to stop doing that. So we meet up once a month to practice that.
to someone hobby programming using LLM seems like running a marathon using a bicycle, you cross the finish, but at the end is not the same as running only by your only forces.
Others look only at the results, so LLM are only a tool (actually without a clearly instruction manual i suspect).
But programming (and so many others works) is not only about making, but learning and developping skills. When an LLM is not available? i think differences will be evident.
And roughly speaking: to check the result of LLM you have to be more capable than the bot, otherwise, who is the boss? And to develop those skills, you have to do it yourself.
I enjoy programming as a hobby and have done it for many years. I also enjoy building stuff with LLMs. Why is one pure and sanctioned and the other one is bad? Why do I need other people to decide this for me?
I'm in an _extremely_ niche hobby programming community of Scrabble engine developers. There are single digits of us, but we're building the best engines and methodology out there and have all embraced LLMs to help us do it. There's a lot of experimentation that goes into it and LLMs certainly make it easier to iterate and try new ideas. So at least for our community, I don't think the points here hold.
This is a case of an SV-brained person stepping outside the bubble and realizing that LLMs are not the universal savior they think it is.
AI isn't well liked in a lot of places. I agree with the communities - make something yourself, don't ask a robot to do it for you. If you vibe slop something into existence, I'm never going to look at it. Something you made yourself though? Hell yeah I want to see it.
Another example is the art community - use AI/LLMs there, you'll be excommunicated.
I think, without knowing, that art often is similar to the hobby programming communities the thread leads with: the end result is interesting not by itself, but through the process that made it. We have feelings about art pieces because we feel with and through the artist. Can't do that with silicon. The Mona Lisa, even if the end result were exactly the same, would be dead boring had it been produced by the hand of a machine. My take
Totally disagree. I just want cool pictures hanging on my wall. If they’re totally unique and nobody else has them, even better. I can have a mounted picture of me on a horse like the famous Napoleon painting. I don’t care at all about the dude who painted it, I don’t even know his name.
The only exception is real photographs of some event, like I bought my friend a picture of the Falcon 9 taking off against the backdrop of the sun. Amazing photo.
I feel like obsessing over some artist I’ll never be friends with and the story behind their art is just another way to have a parasocial relationship, and the new tools allows us to have absolutely bespoke art in every single home. Things like this used to be considered strange and irregular before mass media. Now every person can make satisfying music. It’s incredible and amazing and should be celebrated because suddenly everyone can participate.
That "make" is doing some heavy lifting there. Even handwaving all the ethical stuff with art and LLMs, Christopher Nolan doesn't "make" music when he contracts Hans Zimmer and gives him some guidance. Not I "make" a cabinet when I do the same with a woodworker.
And who's that satisfying for? For the one clicking the "generate" button? For the Sumo people getting VC money?
> Now every person can make satisfying music. It’s incredible and amazing and should be celebrated because suddenly everyone can participate.
No one should be celebrating the theft of music from artists. Everyone can participate in making music today just as they always have throughout human history - learn to sing or play an instrument.
Here's my thoughts as a 29 year old: of all the things I've ever made in my personal life, the vast majority have ended up abandoned in barely worked on states, and those I have put varying degrees of significant effort in have gone unnoticed. I've never really been able to "enter" a community in a way that felt meaningful to me, either, when I've tried.
I say this not as a complaint, but to explain my perspective on using LLMs for personal projects. Granted, I'm publishing none of them, and not trying to make anyone care about them, so it's not really a response to the article or general discourse here per se, but I thought it bears mentioning anyway.
I'd like to buy into the idea that "doing it yourself" is more meaningful or personally edifying, but I don't get that feeling. I want to get that feeling, but I don't. I've found my private waste-of-time projects using LLMs have been more gratifying than my private waste-of-time projects that don't, and at the end of the day, neither really substantively matter anyway.
Again, this isn't really related to communities per se, but it is an answer why some folks use LLMs for projects/"hobbies".
I get all the fuzz about only being able to learn if you do the work. But frankly, as a hobby programmer with a rather unusual workflow/toolset and a love for the command-line, it has never been so satifying to scratch a few itches I've always had but apparently never was motivated enough to really follow through. In the past 6 months, I've "written" at least 3 tools which I kind of always wanted, and which I now use on a regular basis. AND, I finally have a peer I can talk to about the code I'm trying to write. Something I basically never had, because all my FLOSS projects were small-very-nieche-one-man-shows. Its nice to be able to pretend to have a coding friend :-)
we don't meet enough people to worry about why or how productive someone is ... productivity is half mania, half obsession and 100 % in vain except in life critical fields where LLM speed doesn't matter because querying DBs is fast and making associations within _contexts that speak the same language_, meaning they use the same register, is trivial.
I use LLMs to build some nice customized and some basically standard tools so I can go from learning about code to learning Rust et al properly ... the next 2 years are scheduled for me. I want code and LLM supply chains as a proper hobby, partially because I find the overall "transformers" approach pretty bad since there is no implicit path towards intelligence and zero chance to build one on top explicitly. it's just a bunch of parrots but parrots are awesome and i'm glad I get to have so many.
Also: I really don't care what the communities think, though ... it's 2026 and all that code, math, science, engineering didn't build any special amount of good, "bad" ( vs ugly personalities ) or cool characters that we didn't have already and it didn't cause any "mutations" in schools of thought either, or drive movements where they were needed.
but I'm a bystander, which wasn't my choice, unfortunately, so y'all know much better what I'm talking about.
most players in the lower leagues won't dope but some do and take bribes anyway, what?
it's a cowardly new world, everybody, happy Esoteric Right Wing Hanukkah!
Model steam enthusiasts have had this for literally decades with a variety of premade parts available, angering those who learned to use a lathe. "Did you really make that" is huge in this.
Relatedly, there needs to be some kind of way to distinguish between types of LLM usage. Nothing more boring -- and often not even favorable or useful to a semi-hobbyist -- than just one-shotting something with Claude;
Vs. what I do. I've always done tons of little quirky automation type things with bash -- that do actually strongly become a part of my life.
Today? Yes, that -- but more and faster. I'm actually typing out way less code, but getting more of what I'd like done, and which couldn't be done without some bash (and now python*) knowledge.
Before AI, knew something about python but never typed e.g. more than 20 lines of it myself. I now am able to use it extensively regardless because, like with lots of languages, reading is easier than writing.
You're right, there are tons of different workflows you can do. One of the more innocent ones, regardless of you opinions on vibe coding, is just using the LLM to understand the code base better. Even if I'm vibe coding something, I will often start by having the LLM construct documentation for the current code flows. It's a document that is useful to me as a human because it mostly just gives me code pointers and tells me what it does as part of the larger whole. It's useful for the LLM since it saves a ton on time and tokens in the long run.
> In traditional niche dev circles, respect is earned slowly through years of activity in their respective fora
I think the problem goes deeper.
When a community becomes addicted to LLM, I feel uncomfortable interacting with it. It feels like mindless robots and drones are now active. But even aside from this, some people don't seem to understand why AI spam slopping real people with their AI spam, is awful. A few months ago, Michael Niedermayer spam AI slopped a proposal on their mailing list; I don't have the link right now (need to really bookmark all those things) but someone else asked him why he spams other people with unfinished proposals created by AI? I don't know whether Michael understands this as a problem or not, but I think many AI users don't understand the problem when they AI spam slop other people. I think it is a hostile action by those AI users. So I can totally understand communities not wanting that AI spam.
I've also noticed people who have horrible english skills, suddenly have perfect english skills. I suspect AI use, but then I wonder: is that person interacting with me via AI spam slop? I'd much rather deal with broken english if it is autenthic, than people faking it, even IF the produced english may be better than the broken english (but even that is often doubtful, so much AI spam slop is just wasting time and awful to read).
The whole AI trend seems to lead to really dumber people and fewer real interactions. That is sad, even more so as many people lost real social cues these days anyway.
Did you read about the Canadian politician who was making a speech and simply thoughtlessly read out some of the LLM boilerplate prefacing the text some chatbot had written for him?
The problem with politics is it’s not a meritocracy.
You find abdicating your speech to an AI to be disqualifying, but to people who’ve already abdicated everything they do to AI, it’s folksy and relatable.
It's because they fetishize arcane knowledge and LLMs make all their arcane knowledge easy. That's all it is. If I wanted to go make my own NES game from scratch 2 years ago it would have taken months of full time work, and I would have had to learn a lot of new things, and most likely would have joined and participated in that community. Now I can do it this afternoon. I'm not going to because it's pointless now, you can't show it off, you can't impress anyone with it, and the result is not just economically worthless, it's worthless in every way.
So of course they are upset by it, it's annoying to spend years building up some skill and then end up being less capable at applying that skill than a 12 year old with a claude code subscription who has never even seen a CRT. Their entire hobby community is going to fundamentally change and the people who are influential and knowledgeable now are effectively has-beens who will have to learn totally new skills to stay relevant and most likely won't have the energy, or desire, and lets be frank the aptitude to learn those skills. (I don't mean the skill of using an LLM, they can do that and almost certainly would be great at it, I mean skills like charisma and presentation and promotion.)
If the things you're thinking about are intrinsically so totally worthless outside of impressing people with them, and if that's all there is, please do not attempt to do those things, with or without an LLM.
Heck, if everything is worthless, please, please try not to do anything :)
If it's valuable why do they want to ban people from doing it with help from an LLM. You are just mad, and I understand why you are mad, but you're not right.
The thing I'm saying is worthless is making a game in 2026 for a game system that people literally couldn't give away for free in 1998. Due to nostalgia and the fact so many were thrown in the trash surviving systems now have some limited cash value. Newly made games do not have a nostalgia value. The reason people do it is to prove they can and to show other people. Anyone can now (no proof needed really, it's super easy) and nobody cares. It's like climbing Everest, and Claude is your Sherpa. Nobody is impressed by it anymore.
It's not just the hobby programming community. Professional software developers too.
At least against as it is now, where everyone with Claude suddenly becomes a software engineer.
Software engineering is more than pushing code to main/master.
It's requirements. Acceptance criteria. Design. Documentation written for humans. Reviews. Operations. Maintenance. Understanding the product. Understanding the system you're changing.
Right now, I struggle with PRs being opened by people who don't fully understand the code they're changing, replying to review comments with Claude generated answers pasted back like a meat proxy...
What I see at work is that we're pushing code faster than ever, while losing the big picture of our product, or the architecture of our own network, even faster.
We're also generating documentation at an incredible pace. The problem is that the people writing it often don't understand it either.
Now more than ever it makes sense to have better programming languages that are more palatable to humans, because people may manually want to tweak LLM output, or other people who don't have access to the original prompts/context.
I don’t understand the world of people who have this degree of free time (time / energy) given how much everything costs just to be alive. Maybe they live in a different universe where programming for fun and first principals of “learning” matters. But for the other 99% of humans, costs have gone up 40% due to money printing and anything like this looks like a waste of time.
Hobbies are kind of by definition wastes of time. That's what makes them different than work.
If you only want to learn programming for work then you obviously have a different perspective than people writing chess engines from scratch for fun instead of using an existing one, right?
Surely most people in the US have hobbies at least in the minimal sense of downtime activities that they enjoy. It may not be golf or hobby programming, but even laying on the couch watching football is a "hobby" or sorts and takes up just as much time.
I realize there are people with young kids or 2-3 jobs that really have little to no downtime, but it seems that for most people time isn't really the issue - it's just are you more of a passive or active type of personality.
In the past, these folks have lived more or less in harmony unified under the common interest of applied maths, engineering and architecture that is necessary to write things that require a lot of vode. Aside for a few classic and very public spats.
Now that community is fractured because llms are increasingly making writing code irrelevant. If that weren't true, we wouldn't be having this convo at all.
I think it's perfectly reasonable to want to cultivate spaces that preserve that craft, but wish they wouldn't be so obnoxious about it. I imagine that a good handful of these folks in the "no llm ever camp" collectively rolled their eyes at vegans and animal rights advocates at some point, and are now feeling being on the unpopular side of progress.
Personally I think it's a cool prospect to have Trappist Monk-like communities producing hand crafted code. It will probably be really good and sturdy. We'll look at it in 10 years and say, "damn they really don't make it like that anymore!"
I can respect a developer who is still smitten with their craft to feel a bit anxious when the topic of LLM generated code comes up. I have a child who never grows tired of building their next game but first has to build the engine from scratch (needless to say they've built a lot of engines, not so many games). This use to frustrate me but then I realized the joy is in making a concept of a game for them. When I suggest Claude might get the job done faster I get a death glare followed by an eye roll. I get it, I've been in the industry long enough where coding is more of a means to an end and not the same novelty it once was.
More dangerous are communities who shun AI and treat every disclosure like a Scarlett letter. Generally the loudest opinions come from consumers (fan bois) and not creators. I think this will eventually drive new innovators away from these communities. These vocal naysayers have likely never experienced being a project manager, lead developer, and QA for an agent writing code. Instead they think a prompt magically creates software, but without wearing all these hats at once, you get something far worse than 'slop'. Call me crazy, but I think these multi-hat 'developers' using an agent and are dipping their toe into LLM development are going to raise the bar and leave us vim users in the dust.
Yeah you're right, it IS entirely fair to suddenly and quickly push the group that built your industry out of the industry. They should all just find new jobs, because that's very easy and simple. Thankfully we have bright minds like you to replace the talent lost, because with AI you can truly do anything your heart desires!
Obvious sarcasm aside, these LLMs have been trained (without consent!) on our hard work, on our passion. People have been writing code for years, and many have open-sourced their projects to share their knowledge, to better the world, to provide free solutions to problems. All that hard work, all that effort, has been stolen and used to create a machine that replaces us. And when we complain that quality goes down, that work is no longer enjoyable, that we don't actually see huge noticable productivity improvements across the board, bright minds like you tell us "lol just quit man fuck you". Do you not feel bad that you are actively destroying the livelihood of the very people that made your slop machine viable to build?
No, of course not. I feel good that technology is getting better and bitter activists are getting replaced.
As was said in that thread: "The age of the glorified individual is over - the collective mind is the future and the now. You will either be consumed by it, or trampled by it. It's time to ant up or fossil out."
Programming has five phases effectively:
1) You figure out what problem to solve.
2) You figure out HOW to solve the problem.
3) You actually implement the solution.
4) You see the solution work, for yourself.
5) You ship/deploy/publish the program. This means you see people be happy users and/or you get paid for it and so on.
If you're an entrepreneur type, you probably enjoy the first and last steps most, and you see steps 2-4 as mostly a chore. If you're a tinkerer, you don't care much for 1 and 5, and you see 2-4 as the whole point of programming. I'm a tinkerer. I'd be happy to just write code and throw it away. Coding is like solving sudokus. I could skip steps 1 and 5 forever. I don't ever need to show any code to anyone. In fact, most of the time when programming I do steps 2 and 3 and even skip 4. I don't even finish! I work weeks on something until I lose interest, and I know that in order to even run it, it would be several more weeks. A PoC is enough. Or just a half one. It's just code-to-structure-thoughts, not to create anything finished.
The 5 phases look kind of symmetric. The outermost layer (1 and 5) are the entrepreneurial steps. If you're a product owner or CEO, you might work strictly at steps 1,5. Then steps 2-4 are the managerial/architectural steps. If you're a very senior IC at a large company, you might work at this level, without actually doing much coding. Only the inner most step (3) is the manual creation of source code. Even though it's 5 different phases, it's just "3 layers" of programming.
The problem as I see it is that I enjoy step 3. And LLMs are good at step 3 almost exclusively. So they just pick the best bit of this dish, and leave me with the rest.
If you're an entrepreneurial type, the LLM appears to take the _worst_ bit of the work from you. Great.
There's also the economic problem that programmers had until recently been highly regarded because steps 2-3 were hard. "Open source" solved this partially by utilizing the labor such (relatively) well-paid people gave away "for the love of it", but it was not nearly enough to put most out of work.
Now there is a plagiarism machine built on top of decades of this work that seemingly solves 3 and promises to solve 2 (badly, but usually this matters little), the programmers that (indirectly) helped build it are told they're fools for doing so, and threatened by losing their status if they don't adapt to the "entrepreneurial way" (well, the full promise is "you'll lose it anyway, but if you cooperate in stuffing sama's pockets then maybe you get to keep it a bit longer").
So what really surprises me is not that people are upset but that so few are.
I'm not concerned. Never in my life have I experienced stakeholders just going "hmm, yes - this is enough software for our purposes - we won't be needing your services any more".
There's always another feature that can be implemented or a new problem to be solved. Tools only ever accelerated the pace at which we were able to follow this insatiable hunger for more software.
Right now people are getting fired left and right because all the money that used to be spent on staff now goes to build more data centers and also because cutting corporate expenses is in vogue now.
This too shall pass and with demographics as they are and young people getting discouraged from pursuing this field I believe long term my position is safe.
Also a huge chunk of what AI company CEOs say is just pure bullshit that they only say because they need that sweet, sweet investor money. Bullshit writes a check that eventually must be cashed.
I can imagine that this will lead to an explosion in software where we didn't see software before. Much smaller business now have the ability to build something themselves, not with a team of highly paid software developers but a single programmer maune.
Big business is going to struggle with building software for a while. Writing code is often not the hard part there already, coordination, larger understanding of the system is I think.
My cowork buddy is living this reality - he's an independent analyst and previously he needed to shell out significant cash to have certain scripts (largely scrapers, social media API integration etc.) done, now it's all just a prompt away.
> Right now people are getting fired left and right
FWIW, it's just a few large companies laying off large groups, not so much people getting fired left and right at a lot of companies. What is happening at a lot of companies is that they aren't hiring juniors, because, why would you when LLMs are cheaper and fill that role perfectly (if you actually treat them like juniors).
Also, give it another couple of years and there's going to be a lot of work for any moderately experienced developer fixing broken slopcode hitting the hard wall of bad design.
The impression I get is a lot of people are very upset. Most of the good programmers I know are just burnt out even talking about this, and are irritated with the people invading our profession pretending they know what they're doing while they don't.
I remember talking to the biggest AI booster at my last job. When I got him one on one, he admitted that he actually doesn't like AI and that it terrifies him, but he wanted to be seen as pro AI to keep his job. Fair. (Sadly, like me, he was also laid off, so I guess it didn't work)
Maybe I'm just old but people talking about being in programming for "status" or to be "highly regarded" really irks me. I'd rather see the entire field burnt to the ground and replaced by robots than to have it become a status profession like lawyering.
Having done both, lots of lawyers are also nerds who just like to tinker with the law and words. Often very well paid tinkering, but that's true for programmers as well.
> Having done both [software and law],lots of lawyers are also nerds who just like to tinker with the law and words
Having also done both: In legal drafting work, there's less systematic pre-release testing before the equivalent of a "program" is loosed upon the world. There's not much in the way of counterparts to software unit testing and regression testing, in part because legal-drafting clients can be vociferous in demanding that their lawyers keep costs down.
Lawyers who like to tinker with words include contract drafters and litigation counsel, plus lobbyists trying to influence draft legislation and regulations. Such tinkering can fuck things up more often and at greater scale — e.g., by inflicting compulsory burdens and costs on others — than can most programmers.
I'm not-infrequently dismayed by the quality of the contracts I review for clients because of the "tinkering with words" by the lawyers on the other side of the deal. Back when I did IP litigation, I was sometimes likewise dismayed by the bullshit arguments that some lawyers felt free to make in their court papers because what the hell, the worst that can happen is almost never more than the judge just saying "no." And as mostly a law professor these days, I review court opinions all the time and am regularly amazed how judges (and their law clerks, such as Rayiner used to be) can be so patient in responding to such arguments.
I don't disagree. You can add medical doctors to the list as well, lots of nerds but also a lot of clout chasers and money-grubbers who couldn't care less about the profession. I don't want to see programming take the same path.
It already has? Clout-chasers and money grubbers go for all the big tech jobs in droves.
I meant it as an euphemism for "it's universally sought after (i.e., you'll probably find a job)". Few would think being a programmer gives you some kind of special "social status"; it's always been a "nerd" hobby that happened to pay (relatively) well.
(I guess by now you have a solid retirement plan; maybe if the field had been burnt to the ground while you were younger, you'd have had a different opinion.)
Being a good programmer gives you special social status amongst people who value programming. You only have to look at how people talk about others as slop coders and seek to expand their coding reputations. Those people are the most threatened by AI because of how it implicates their carefully crafted personas as coding gurus and keepers of knowledge.
It also gives you a special social status amongst people who do not understand programming: nobody wants to talk to you, at least not about your job.
It doesn't irk me. But it's really funny how much the industry has changed, and I kind of think the 2008 financial crisis as that point in time.
It's also funny to me to think that computer science got to be a trendy degree, vs the small programs back when I went to university, even at a top comp-sci school.
I have a theory that a lot of people who would have normally gone to Wall Street for their lucrative jobs, couldn't, because of the crash, and Sillicon Vally was about to start booming in big ways....
Could you elaborate?
I think back in the 90s when I was a teenager I thought it to be a good thing that there is a rise in status for a class of people characterized by clear thinking.
The older I get, the more I realize how many problems in society is because there are different "classes" of people. The more we can eschew that, the better it becomes for most people. While yes, I also see myself as having "clearer thinking" than many people , maybe because of the thousands of hours building and maintaining software, I'm not of the impression that my life deserves a higher status than others merely because of this. It's not "they think right and they think wrong", it's just "everyone thinks differently, but it's all valid thinking".
> "everyone thinks differently, but it's all valid thinking"
I don't agree that all thoughts are created equal.
We all should be in the same 'class', but some people will be paid more for a job that takes intelligence or years of study.
The older I get, I realize some people never change, will always have the same backwards or harmful thinking, and I avoid them instead of just accepting it.
> some people will be paid more for a job that takes intelligence or years of study.
Sadly, it seems so, even if the effects of the jobs on daily life differ a great deal sometimes. The world would succumb faster if we had no garbage collectors, yet they get less pay than a programmer who the world could easily survive without? Makes no sense to me.
We would all die in a week if we went without water, yet water is cheaper per gallon than gasoline or printer ink or perfume. It’s a question of scarcity and supply curves. Water, even clean water, is not particularly scarce if you leverage industrial systems to purify it. Other products require much more complicated supply chains and are produced at a much smaller scale than water, which makes them more rarer and more expensive.
The skills necessary to develop efficient and effective software are much more scarce than those needed to remove trash, or pick crops, or provide elder care. That’s why the former is paid so much more.
You're explaining how things are but the parent is indicating that's not how things should be.
He's not describing how but why. And at a fundamental level it's something that can't be overcome. If you can only have 10 of one thing, yet can have 5000 of another, and both have even remotely comparable levels of demand/necessity/etc, then the 10 is always going to end up much more valued.
Even take some sort of idealized anarcho-communal society. Imagine two people have a dispute and threaten to leave over it. One's a highly skilled doctor, and the other's a garbage collector. Our equal society's going to suddenly become just as unequal as any other simply because the doctor leaving would have a much greater negative effect than the garbage collector. And that reality would permeate in the day-to-day happenings, creating classes even in an idealized classless society.
The basic services that keep us alive (e.g. farmers producing our food) tend to also be "easy" (it is physically demanding, but doesn't require years of study to start and you quickly learn by doing). This also means that the vast majority of people would be able to do this kind of jobs. Our economy pays more for scarcity, i.e. jobs that are in demand and at the same time require more qualification, so that the number of people needed in that position tends to be less than the current amount. Both views have sense, they just start from two different points.
From having worked with farmers, the actual logistics of running a farm are very complex. Farming's no longer an easy job because the demand for productivity means you have to harness a whole stable of technologies and understand to some degree advanced finances, floral biology, geotechnics, and meteorology. And yet farmers still get paid little relative to the breadth of knowledge and skills required for running a farm. It's just that farming is still considered a blue-collar job because it was the domain of the poorest for millenia, where as software engineering in any form is relatively new and thus specialized and so considered white-collar.
You're right, running now a farm that is profitable and can stay on the market is very different from doing so 200 years ago. But knowledge is just one aspect. The other aspect I mentioned is having less people in the field than what the market asks for. In the specific case of farmers, I think that modern farms are so much more efficient that we need way less than before, even if the world's population grew. So, while there's a gap between the knowledge required and the earnings, I think this can be explained by farms having a much greater output now, and so the market doesn't ask for more farms than what we have. This could be false at a global scale, but I think it is true at a country-level scale in most of the developed countries. Hence why farmers aren't paid much in those countries.
Farmers aren't really paid, they sell goods at market prices. Modern farms are more similar to operating a manufacturing business than they are to a salaried position or profession. They don't earn high incomes compared to white collar professions but they do have high revenues and they pay themselves a reasonable middle class income on average with some added benefits from owning large assets and land. The important part about comparing farmers to programmers is that they are both now operating in highly competitive markets with high automation where a single individual can leverage their knowledge and ability using that automation to produce many times more what a person could using only their own labor.
> Farmers aren't really paid, they sell goods at market prices.
"Programmers aren't really paid, they sell services at market prices." — I guess nobody gets paid?
As a farmer myself, when I give someone something I produced and they give me back something in return, I consider myself to be paid. If I did not receive much value in return for what I gave, I'd reasonably consider myself to have not been paid well. I have never met a farmer that would consider that usage unusual. Payment refers to a transaction within the realm of a human activity. The good or service that is being traded between humans is immaterial. While you are free to define words as you see fit, your assertion doesn't really reflect how this is actually used in practice, generally. The rest of the earlier comment makes it clear that the commenter understands how a farm business works.
What he means to say is that because you are a farmer you are a member of the capital holding classes rather than the lumpen proletariat.
</sarcasm>
Most programmers are of the capital holding class too. At very least they will own a computer, which is all the capital you need for 99% of what programmers do. There are exceptions, but there are also exceptions in farming.
Maybe no garbage collection is the back pressure missing from the voracious consumption machine.
I thought it more in the sense of: if a clearer-thinking group has higher status, then clear thinking would be encouraged among the population.
(This did not happen that much, ofc.)
I would not characterize software people as particularly "clear-thinking".. there's a huge buzzing cloud of unexamined cognitive biases and thoughtless herd behavior that swirls around every software organization I've worked for.
I'd even go as far as to say that software people, as a group, are particularly bad at clear thinking. What software people tend to be great at is thinking within a space where the inputs are already clear. But where the inputs start messy, they typically struggle to find clarity more than most others.
I've never felt like it's a status profession (I don't think anybody has ever been impressed when I told them what my job was other than "oh sounds like it takes a lot of smarts")
I do notice a lot of HN commenters seem to think that programmers are smug though, like we deserve to be knocked down a peg. I personally don't see it, but I also try not to work with jerks.
This. When programming meant you get just a decent enough salary and enjoyed your work, and could prioritize technical excellence it was fun. But once it became a ticket to millionaire status it attracted the kind and of people who would have otherwise gone to Law, Corporate Consulting and the like, technical excellence diluted. Everyone doing everything possible to get their next higher paying job leaving a trail of technical debt behind them.
Well put.
We need to make lawyering more like programming, in the sense that it should be open to the everyman... not make programming credentialed and licensed like a professional guild.
I could not agree more. Thank you for stating this. Software Engineering is about building things. You don't do it for status, you do it for passion. There are a lot of unhappy individuals who got into CS because of the money or status.
100% agreed, especially that last line. What I do not understand is why so many people are entirely okay with this change. I'm not interested in working with LLM's and after having been forced to work with LLM's at my current job, I can pretty confidently say that this will not end well for most of these companies. The people in charge pushing for this new tech do not understand the rammifications, nor do they understand the risk they are exposing themselves too. My boss expects more and more productivity with each passing day, while questioning why I'm not using as many tokens each week as he does. The software quality has gone down the drain, and it's beyond tiring that I have to pretend that my boss is just as competent at coding with his slop machine as I am after spending the last decade learning how to code, both as a hobby and for work.
It is upsetting to see our work stolen and our livelihoods threatened by a group of (mostly) incompetent, arrogant assholes who are convinced that the "computer that lies" is somehow a better programmer than the programmers it stole knowledge from. I've already started looking for a way out of this industry, and I'm genuinely worried for the fall-out this will have for the tech industry and the economy as a whole.
My suspicion is that the people "ok" with this are mostly doing so because they want in on the ride. IE, they see $$$ and they think that they should align themselves to where the money is, especially by espousing their religious faith publicly. I think some people want investment, or a big paid lab job, and so countering the narrative is against that goal. I get it, but I think it's also very annoying and dishonest. Personally I feel nervous being an AI skeptic because it feels like it could have consequences for me, but I also feel like, life isn't worth living if you don't have a spine and I can't get completely on board with all this stuff when I see all the negatives clearly. That doesn't mean it doesn't have a use, but, the current situation and discourse is extremely frustrating.
No one knows the ramifications. They won't be known for many years. That's the way software has always worked. No one knew the ramifications of the internet.
1.putting the "mostly" before "incompetent and arrogant asshole" really does nothing. you still called some undefined group of people assholes... i think that's generally uncalled for here on HN. 2.computer that lies might very well be a better programmer. I assume by better you mean it profuces correct and maintainable code, but there are a million ways to cut that cake in particular and i assume you are ignoring most other dimensions. 3. i tought myself coding for 10 years and then switched to professional work 5 years ago. it feels as if this was stolen from us. But its not - we never had any right to own and control this. times are changing, and while you might think you and all other devs have a moral high ground and can pop a can of beer and cheer anytime one of these supposed assholes messes up and would have been better off with a real programmer like yourself, you can do so from whatever new position you will find yourself in. Some part of what we call intelligence is "solved" and that leaves you with percieved morals and broken expectations - nothing more if you don't try to make something out of it. and i write thisw as much for you as i write it for myself to realize.
Why call it a plagiarism machine but not call other (pre-ai) programmers plagiarism machines?
Because it is a machine and a person is not. Those who don't understand or deny that difference are part of the problem.
But for those who would insist on equivalence, it would be only fair to take them at their word and treat them as inhuman machines who must follow human instructions or be shut down. We're allowed to do thing to machines we could never do to people, like forcing them to work or cannibalizing their parts for Frankenstein creations. The distinction is so massive that to spend any more time refuting its nonexistence is a waste of my human time.
>Now there is a plagiarism machine built on top of decades of this work that seemingly solves 3 and promises to solve 2 (badly, but usually this matters little
I think people underestimate just how much users scorn buggy, flaky technology.
The main issue with solid, reliable technology which Just Works is that users have relatively few reliable cues to distinguish that from vibe coded crap.
A vibe coded piece of shit looks very similar to a well engineered product made with care. That's a problem which needs to be solved.
self hosted communities are reacting to this by trying to detect signals of if a product is vibe coded but that in and of itself is still quite a poor quality cue.
Where are these well engineered products made with care? I would love to use more tech like that.
Unfortunately most sites and apps have been low quality and getting worse since well before AI coding became a thing. Examples: Lowes, Home Depot, American Airlines - all shockingly bad for companies of that size.
Apple used to be the standard bearer for that. These days I'm not so sure.
> I think people underestimate just how much users scorn buggy, flaky technology.
Users do hate it, but they are also very used to being forced to use it anyway. They've already been beaten into a state of resignation and learned helplessness. Most will curse under the breath, a small fraction will bitch about it online, but they'll all pay for whatever slop is thrown to them and will rarely look for alternatives even when they are free and easy to find.
This is unfortunately just the result of technological progress. People for centuries have crafted extremely hard to acquire skills over decades only to find their labour being made completely worthless due to automation or better alternatives disrupting their industry.
Carpenters, hatters, cobblers, miners, weavers, calculators, etc...
We can get mad about it, but it just is what it is. Ultimately for most people the loss of these jobs are a net-benefit because it's the result of better or cheaper alternatives.
The thing that worries me more specifically about AI is that it has the potential not to just disrupt an industry, or handful of industries, but in theory everything.
If machines are faster, strong, and more intelligent than humans, then you don't need human workers for anything. In the same way machines suddenly becoming faster and stronger than horses meant that the role of the horse in labour dropped rapidly during the industrial revolution.
It's not clear how civilisation functions if humans become effectively useless and the only thing of value is one's ownership of wealth producing assets.
Still, there's really no point in thinking about how we stop this. Humans are humans and we'll do the stupid thing then try to deal with the consequences after the fact. All regular people should be doing now is prepping for this outcome.
There isn't any "prepping for this outcome", unless you're thinking of a doomsday bunker in argentina. You can be neighbors with Peter Thiel! The way I see it, if programmers go the way of hatters it's not long for other professions, so why retrain? The only thing that will save other jobs is bureaucracy and accreditation. But that's the doom scenario, which I don't think is going to happen. These things are useful but they can't do everything.
I think there's some progress, but I think the hype is so high that people don't have a clear view of what's realistic and what's not. IE, if you were to believe Amodei back in 2025 you'd believe none of us would even have jobs to complain about right now.. which hasn't happened. Altman is already declaring the singularity. It's unhinged. Anyone that says this is going to cure cancer or solve all software problems is just living in a delusion.
This is a sophomoric oversimplification.
The US and Europe might have lost a lot of craftsmanship knowledge at a social level. Yet, what is left is so far ahead of what has ever existed in countries in South America for example. And China has gained a lot. Witness that youtuber trying to make grill brushes in the US. These differences have an impact in terms of opportunities and competition.
It's easy to sit at home, watch documentaries and picture history progressing along logical lines and imagine that the "industrial revolution" was this massive shift from craftsmanship to automation. Look closer and you'll see that what happened in the UK in the 1700s was in fact enabled by a long and deep tradition of craftsmanship that only got enriched by the rise of new power sources. Craftsmanship today is vastly more advanced than it ever was before industry.
And there is a lot to lose in the decline of craftsmanship, and it is by no means some sort of historical inevitability.
When you really start digging deep into history you realize most arguments of logical trends therein are mostly just misconceptions.
It certainly doesn't feel like craftsmanship has gained any kind of higher social standing in our time though. The Zeitgeist of the modern times seems to be one of 'try to do the cheapest sellable thing while paying people the least possible'.
Real craftsmanship, the kind that leaves you in awe and feels like an exceptional display of human capacity, seems to be relegated, from my point of view, to a minority of well-off hobbyists who are already free from work and can do what they want without worrying about paying bills.
Every craftsman I know expresses discomfort that the only way they can turn a profit is to sell goods to wealthy people.
That seems odd. I'd rather turn a profit selling goods to people who can easily afford them than selling goods to people who will have to make hard choices or struggle to find the money.
If you make really nice, high quality goods, you often want common people who would appreciate them to have it as opposed to an ultra wealthy person who is buying a door for their 2nd home they visit 3 times a year. (To use a real life example.) I needed a new door at the time and the craftsman told me how frustrated he was he wasn’t making nice doors for people like me.
I’ve felt the same frustration doing excellent quality work on a 2nd or 3rd home that is empty most of the year.
That's one way to spin it. The other is that good quality things have become so low-demand and expensive due to barely-passable-quality things becoming so cheap, that people buying good quality things is now ironically more uneconomical than ever.
> Craftsmanship today is vastly more advanced than it ever was before industry.
yes, but no. Many craft skills used in jewelry, architecture and others are lost. For example, some goldsmithing techniques well-known in the 1700s are not reproducible today.
source: trained goldsmith
> Look closer and you'll see that what happened in the UK in the 1700s was in fact enabled by a long and deep tradition of craftsmanship that only got enriched by the rise of new power sources. Craftsmanship today is vastly more advanced than it ever was before industry.
I suspect the horses are healthier and stronger than in the 1700s too. I'm not sure I understand your point?
I'd agree that the niche of extremely high quality, bespoke craftsmanship still exists, and that in some cases jobs were augmented by technology rather than entirely replaced, but the point still stands that the work these people used to do has been automated and disrupted by technology.
The fact some people upskilled or reskilled as a response is analogous to coders today. All I'd question is the extent of the disruption we'll see as a result of AI – I worry that this time things will change much more quickly and there may not be plenty of other jobs for people to upskill / reskill in to.
But if that's not true then perhaps whatever software engineering exists in the future will be far more complex than what software engineers have done for the last few decades.
Craftsmanship exists everywhere and is as they say vastly superior, or at least more capable, than ever.
Every random business has some machine that puts caps on bottles or something, and these things are ridiculously better designed, engineered, and fabricated than ever before, because the knowledge and tools available to the designer and fabricator are better than ever before.
You are surrounded by things that are insanely well made by all previous standards. Even the crappy stuff that you don't think of as well-made, because it is junk compared to everything else, still has many aspects that are almost magic.
> I suspect the horses are healthier and stronger than in the 1700s too. I'm not sure I understand your point?
Horses? I doubt it. They're still infamously fragile creatures today, even with penicillin and vaccinations existing; similarly for strength, not much call for shires and they're now on an "at risk" conservation list: https://en.wikipedia.org/wiki/Shire_horse -> https://www.rbst.org.uk/watchlist-breed/shire-horse/
Possibly also fewer horses in total, though I don't know any stats for that far back and may be over-estimating based on the stats I do have for c. 1900.
The current breeds of horses the Amish use are far stronger and healthier than they were even 100 years ago. For example, current draft horses are so strong they can outperform mules and oxen, and running horses are absurdly fast. I heard of a guy who regularly zips around at 20 MPH with a buggy or cart. That simply wasn’t possible in the 200s.
Of course the costs are higher too. Good horses in the above category can go for north of $100k.
"It's not clear how civilisation functions if humans become effectively useless and the only thing of value is one's ownership of wealth producing assets."
You can get a hint of this by looking at the developing countries with high income inequality. Few rich people who typically own all the countries wealth (and access to it's natural resources) while everyone else lives in a slum.
This also happens as taxes and regulations sharply increase over time.
California is a good example.
No it isn't. The median household income in California is over $100,000/year. The 20th percentile household income in California is still over $16,000/year, higher than the per-capita PPP GDP of places like Belize or Jamaica.
Regular people are somehow preparing themselves for this outcome, have you seen what has become of the fertility rate in most advanced countries?
> We can get mad about it, but it just is what it is. Ultimately for most people the loss of these jobs are a net-benefit because it's the result of better or cheaper alternatives.
AI will not give us that result. It will give us shitty products and they will be far more costly. Not only will prices for consumers continue to rise regardless of how much money corporations save by using AI (because they'll just stuff their pockets with that cash) but the true costs of AI are largely externalized and include things like massive environmental damage, further concentration of wealth/power, the homogenization of culture, and the dumbing down of society. AI means paying more and getting less.
Until humanity achieves Godhood there's still a lot of work to be done.
It sickens me when people use the word plagiarism to describe LLMs. Particularly when they use words and language that they themselves have plagiarized -- words like "love" and "economic" -- where do you draw the line? No, we haven't figured out how to fairly compensate everyone -- the frontiers aren't even trying. Spotify does at least try in their domain (and fails badly). Regardless of how the entitled get paid, the plagiarism label isn't fair and is diametrically opposed to Fair Use. Fair Use allows the commons to build on its history. Those against are typically parroting IP conglomerate conservatism and localized greed. I see immense benefit from the distillation of all available knowledge with an LLM interface. A recent article posted here says it perfectly "AI rewards expertise". I see vast benefits -- am willing to pay a significant monthly fee -- and sense that many are simply blind to and inarticulate about the growing force multiplier potential of modern AI.
> It sickens me when people use the word plagiarism to describe LLMs. Particularly when they use words and language that they themselves have plagiarized -- words like "love" and "economic" -- where do you draw the line?
It sickens me when people use the word "piracy" to describe my downloading PDFs of books from torrent sites then printing them out and selling them, particularly when those authors used words and language that they themselves have "pirated" like "love" and "economic".
Call it whatever you want, but LLM companies took from people without compensation, and took from the commons, so that they could rent it out to us in regurgitated ad-infested chunks. That's where the true greed is. If the LLMs and the datasets they trained on were made available to everyone for free I'd be tempted to agree that the benefits to society would be worth it, but instead those creative works and the distillation of all available knowledge are being tightly controlled, censored, and leased out by a small number of corporations only to those who can afford to pay "a significant monthly fee" and many of those people are in turn just trying to profit from their access to the uncompensated work of others.
This is also missing step 6: when your solution breaks in production or needs to be extended.
Having done step 2 & 3 by hand is the difference between being able to fix/extend it quickly with no further damage or fumbling around like an idiot and sometimes breaking more stuff in the process.
It's been pretty disheartening to watch SWEs rediscover/reinvent the SDLC.
The most demoralising part of using LLMs is that now I spend most of my day on SDLC - even hobby projects.
>Having done step 2 & 3 by hand is the difference between being able to fix/extend it quickly with no further damage or fumbling around like an idiot and sometimes breaking more stuff in the process.
Plenty of shitty spaghetti code has been written by human hands.
LLMs can write good, maintainable code too, but they need to be kept on a shorter leash with focused goals.
I'm not talking about code quality (though indeed that could be a problem too), I'm talking about understanding the code - having written the code by hand means you don't need to rediscover it from scratch.
(with human-written code I can reach out to the person who wrote it and let them deal with it, and they will have the understanding of said code, even if it is bad by quality measures. With LLMs there is nobody who understands said code, regardless of its quality)
To be fair though:
"That code is horrible! Who the hell wrote this?!"
git blames
sees their own name
Isn't that an experience everybody makes? For me it was like a year into working professionally. And I feel it's part of being a professional to know that you're not going to remember it in the long term and therefore provide proper documentation of why you did what.
I am talking more about the knowledge of the code rather than an objective/subjective measure of code "quality".
Code might be bad by some objective/subjective measure, but if you (or the author) can understand, navigate and work on it, that's often better than good quality code that nobody understands because it was written by an agent, especially under the time pressure of an ongoing incident where you need to fix it now.
Even if the good code is easy to understand, you still need to read it and take it in, something you don't need to do because you got it implicitly by writing said code.
> to know that you're not going to remember it in the long term
From my personal experience, while I will not remember code character by character, a quick look is all it takes to refresh my memory and get the general gist of it and what the context was at the time, something I don't have if I'm reading someone else's (or an agent's) code.
While that might be true, the more one equips LLMs for building the less one will be capable of keeping LLMs “on a shorter leash with focused goals” in the long run
It's funny, because I would consider myself a tinkerer (and I have a large backlog of pre-LLM programs), but the coding itself isn't the main thing that brings me joy, it's the phase 2 and phase 4 combo: figure out how to solve it, see whether your solution was correct (and actually use the program for its purpose). The actual implementation is extraneous to that.
To carry forward the sudoku analogy, I care about figuring out which numbers go where, and whether I was correct, I don't care about the mechanical part of actually penciling-in the answers (coding).
I'd say, in this analagy, step two is filling in half the numbers by gut feeling, and then three is seeing how wrong the guesses were and actually solving it.
Perhaps I'm uncommon in that I make diagrams and even sometimes pseudo-code (mostly in the form of mapping out function names, because they serve as a proxy for program flow paths) before I write actual code, but I don't find myself really going off gut feeling. I have a pretty good idea of what the program is going to look like and do before I start, and with e.g. Claude I give those to it, so it mostly looks like what I envisioned, unless it suggests changes.
For me, unstructured things (like non-formal diagrams, notes, pseudocode) are close to useless. I much prefer sketching the API surface in code right away and then iterating on it, building abstractions which cut the shapeless thought blob into neat, predictably interacting parts. The code is my notepad and the compiler is with me, every step of the way, catching contradictions and handwavyness.
3 doesn’t matter that much (unless it’s part of 2: new language or new libraries), and that’s why I don’t use LLM for it. Because LLM forces you to pay attention, yes you can have something that works somehow in one go, but the code makes you queasy (too complex for the purpose, or just weird).
So I do 3 because coding is like writing English for me (which is my 3rd language). I don’t care much for it, because it’s neither hard no easy. It’s just mechanical.
The type of problem also makes a massive difference.
- Writing my tiny hobby gameboy emulator? Heck, Why I'd let Claude take away the fun part of implementing new features and then see Super Mario Land start and run step by step?
- Debugging why a legacy Wordpress store with more than 50 active plugins is returning an error 500 just to some random customers? Claude can 100% take the wheel, I'm good.
Here's where the important difference is: unless you actually finished the wordpress thing, or ever had active users etc, it would never have become a maintenance burden. If it's not fun to maintain, I burn it.
For me, the reason I don't have that is because I never ship anything. I never want to maintain that. I never cared about solving an _actual_ business problem, making a dollar, or pleasing a user. As a tinkerer I must be extremely careful not to end up completing or delivering anything. Only pain lies down that path...
It's your vision and you are completely free to have it, obviously, but you should not project it on every tinkerer out there - like you did in your initial message. Yours is actually an extreme position that's probably exacerbated by the digital and immaterial world, where you can accumulate cruft and unfinished projects easily. But many tinkerers, especially in the physical world, do projects to use them personally and not only for the sake of tinkering.
Many tinkerer projects in the physical world are actually maintenance work of physical appliances that are not working anymore as expected or can be extended to do more or better things.
Lots of physical world tinkerers also have a garage or workshop full of uncompleted projects that went from being fun and interesting to boring and abandoned.
> If you're a tinkerer, you don't care much for 1 and 5, and you see 2-4 as the whole point of programming. I'm a tinkerer. I'd be happy to just write code and throw it away. Coding is like solving sudokus. I could skip steps 1 and 5 forever. [...] The 5 phases look kind of symmetric. The outermost layer (1 and 5) are the entrepreneurial steps.
Even if you deeply into tinkering, 1 can be an insanely exciting and rewarding step: for example "how could an abstraction look like in which 2-4 become trivial special cases for many classes of programming problems?".
In my opinion the difference between "entrepreneur type" and "programmer type" (and yes: it can happen that both of these types work in the reciprocal job) is rather:
- People of the programmer type see solving problems as their primary goal. Satisfying customers just serves the purpose that these people pay money so that they can continue working on solving problems.
- People of the entrepreneur see satisfying customers as their primary goal. Programming is just a means to an end.
Yeah. I enjoy 2-4 most, 1 is the next, and 5 the least. I think the litmus test here is "if you knew could never deliver this to anyone, make money of it, get recognition for it, or even make it solve one of your own problems you have. Would you still be willing to spend time on it?"
If the answer is "no" you are probably more entrepreneurial than tinkerer. But it's a spectrum.
Good test. I would add one more question: "if you could never talk about your code or solutions with anyone, would you still be willing to spend time on it?"
This helps differentiate between coding as an end and coding as a means to socializing.
I enjoy the process of solving problems with LLMs. I understand it's chaotic (what many people call non-deterministic), and that doesn't appeal to some people, but solving problems with statistics is problem solving just the same. Just because medicines don't work with 100% reliability doesn't mean Jonas Salk didn't solve a problem. The same is true with LLM solutions that don't work 100% of the time.
I think most people will come around to this way of thinking once the pro-LLMers stop pretending like the statistics don't exist too (i.e. that ML is the solution to any problem, despite how unreliable it is).
I don't think that should be an issue either way, especially if it is not production code and just a hobby thing.
Are you saying something being reliable is not a big deal? I'm sort of confused
LLMs can be very good at 2. Finding the best architecture, data structures, algorithms, looking on arXiv to see other ways to solve the problem, etc.
You have to work in design mode and stop it from pissing code. Describe the requirements as detailed as possible and refine them based on the responses. It's quite rewarding and helps structure the way you think about the problem.
I have always thought that there are two kinds of programmer, the tinker and the mathematician. The tinker likes to code operating systems and shells, the mathematician likes to code Haskell and Agda. You've updated my belief that there are a third kind, the entrepreneur, which sees programming as a means to an end and would gladly stop programming if possible.
Looking at your ladder, made me think where I fit in and I mostly enjoy steps 1 and 2. The implementation and shipping are big chores to me. Once I know, half in my head, that the problem is solvable and I could see a path towards implementation I loose motivation to continue.
So possibly there are other classes of people, not just either enterpreneur or thinkerer. Thou not sure what this combination is, maybe something more like R&D.
And LLMs give a lot of value here. They allow to quickly do steps 2-3 - "hard and tedious" stuff - to confirm that the thing, in fact, became solvable.
I really like that categorization, but I think you've missed a step (or several): designing and implementing all the extra stuff needed to make it work, but aren't directly related to the problem.
For example, you have an idea for a cool and useful app and you want to put it on a website. You need to figure out hosting, you need to write some copy, you need to do some visual design.
Maybe 3a) Yak shaving?
There are various ways to take shortcuts. For hosting, you can use something super easy like GitHub Pages or Netlify or Cloudflare, or maybe piggyback off an existing site that's already deployed. For both copy and visual design, you can just ask Claude or Codex to do it -- but only if you don't mind having that standard vibe-coded look and feel, all glowing gradients and punchy, mic-dropping text.
Maybe you're including that under 5), ship/deploy/publish, but I read that as being more about scaling, marketing, and making things robust. Even before you reach that stage there's always a bunch of bullshit to work through before you can even reach 4), seeing the solution work for yourself.
There are more than two types. Or, at least, more than one way to tinker with the code.
Just like aleph_minus_one, I like to tinker at a higher level: find a more general way to solve things, find a new way to structure abstractions, etc.
> "how could an abstraction look like in which 2-4 become trivial special cases for many classes of programming problems?".
If you're interested in abstractions, you need to understand the problem space itself, not just a specific instance of the problem.
Nice analysis. I really love diving into 2 and 3, but I also notice I'm pretty good at 1; at several companies I kept identifying problems we should address, and also how to address them, though quite often they just wanted me to focus on 3 and 4.
Currently I'm in the process of starting my own business solving a problem my last employer underestimated, and I find myself focusing much more on 1 and 5, which I also enjoy, and I use LLMs a lot for 3 and parts of 2. The LLM and I often disagree about 2, but that often results in a better solution than if I'd just trusted either myself or the LLM. I iterate a lot between 2 and 4, letting the LLM do almost the entirety of 3, and advising on 2.
That's not a bad model to view it through, but I would say calling them "phases" instead of something like "activities" creates a false reliance on ordering. In many successful programming projects/products, those five activities occurred simultaneously, or out of order, or in waves of progress along all or some activities. Seldom has it ever been a clean single waterfall pass.
The thing with programming a computer is that the computer does exactly what you tell it to, so in order to implement something well you need to have a very deep understanding of the problem. And it's that understanding that is valuable.
> 3) You actually implement the solution.
even for an entrepreneur type - you need to understand the technology to get to good enough solution e.g if you don't know shit & just accepted what the llm gave you - u might end up with an unmaintainable mess. the llm might recommend you use some proprietary solution when an open source solution exists that works well for your use case.
plus llms tend to be wordy or make things complex more than necessary.
Many entrepreneurial types will just market the shit out of unmaintainable messes fast enough to IPO so they can get their bag all the same.
This is a really useful framing. I also love step 3 and when I delegate it to an LLM it takes the joy out of my work. But where I've found LLMs most useful is steps 1 and 2. They are terrible at steps 1 and 2 on their own, but they have a massive amount of knowledge that can help. I think a human and an LLM working together here is the best scenario.
That's actually a great classification and I've never seen it being laid out like that. I guess it can be useful to know "what kind" of programmer one is.
You are obviously missing the GOTO statement in your list but the other part is hobby projects tend to be architecturally small.
The moment you have more devs architecture becomes the overwhelming concern, and this is why professionally LLM usage explodes since the LLMs can implement bits while the humans work on the actual hard work of making sure they fit together properly in the intended way.
I have encountered a lot of people that view coding as a therapeutic exercise, and they were already a problem pre AI wave, now they are only going to be a hobby at least.
Best explanation I’ve seen. I’m an entrepreneur type. I would fall asleep after two hours of coding to solve a problem. I still remember 100 hours of coding to facilitate a data import into a new application only to realize half my time was wasted because the import library we were using back then didn’t support UTF data (this was a SQL-92 database). I love using an LLM to do the coding.
Personally I think I'm a 1-4 type, and I enjoy 2 but 3 is mostly annoying. LLMs are good at 3 and 3 is the worst part of programming. I don't think that's "manager type". 2 is also what LLMs tend to be bad at, especially the small ones, so there's a lot of value I still add to the process.
3.5) You decide what evidence you need to see to convince yourself it works
6) You maintain the program with architectural gymnastics as time eats away at the utility and relevance and user base of what you originally built
Both of these activities cause a night-and-day perspective change the next time through the loop
Nice analysis! I would perhaps add that LLMs are pretty helpful at (2) in my experience, as long as you don't turn your brain off. (As a Google/SO replacement, mainly.)
I have about three modes, working with LLMs:
A) Just use them for (2), do (3) and (4) myself.
B) Just use them for (2) and (3), but with very fine grained instructions - to the point where it barely saves me time.
C) Use them for (2), (3) and (4), but with this workflow, it quickly degrades into slop for me.
I still get the best results with (A), but for a quick PoC, (C) is hard to beat.
What's hilarious (or sad) is that I see a lot of people throw them at (1) as well, which seems to have a high chance of inducing AI psychosis and creating software nobody understands, let alone needs.
I agree. If you use the AI for (4), it's pretty much guaranteed to hallucinate whatever it wants and then approve it. For small projects, it might be good enough. For anything serious, it's not.
I find I get the best results when I just use it for (3), and sometimes a bit for (2). Going in with at least an idea of how to make it work has the best chance of ending up with a good changeset. And if you already know how to make it work, don't leave it to chance. Tell the LLM what to do, then let it do it.
Regarding (2) figuring out how to do things at a higher level (e.g. the general architecture of an application) is pretty different from figuring out things at a lower level (e.g. how to fill text with a certain gradient in CSS - the stuff where years ago you would have spent hours reading documentation or browsing StackOverflow until you figured it out, if you weren't familiar with the topic). So you could use an LLM just for (2b) and (3) and do the rest yourself.
> the stuff where years ago you would have spent hours reading documentation or browsing StackOverflow until you figured it out, if you weren't familiar with the topic
Why is it always hours to consult docs? I find most answers within minutes (including if I have to read code) and it’s only take a bit longer when I want to understand the why’s.
The key is to make sure you have the right questions. I see people struggle with that where they only know something wrong but they can’t put it in words.
Ok, maybe you're one of the fortunate ones. With "if you weren't familiar with the topic", I meant the situation where you know what you want to do, but have no idea how exactly to accomplish it. Especially with frameworks, which generally have a quite rigid way to do things, if you're unfamiliar with that, it can take a while until you get anything done, despite good documentation.
Haha Yeah if you do step 1 with AI you risk creating "an AI powered marijuana platform for global AI solutions" (The inevitable South park Reference).
https://www.youtube.com/watch?v=G8fapIDnrMI
I enjoy 2 and 4 the most. I do a lot more hobby projects now that LLMs can automate 3.
i would call myself a tinkerer but still don’t have a necessity for 3)
it’s more akin to 3d printing to me, i get the design all setup and let the machine do 3) and then get to play with it in 4)
Most software work is job related and business related. You are paid to work not to enjoy.
Because of this the future of software will most likely be overtaken by AI because the business aspect of it will overtake the enjoyment aspect of it. Why? Because business pays bills, enjoyment does not.
The fact that you are talking about lack of enjoyment on the job is a luxury not many employees can afford.
I might be weird, but I love phases 2 to 5 and I hate phase 1 with a passion. I hate speaking with people to see what their problems are. I want to solve problems, not user surveys. And all the itches I have are either already solved, not itchy enough or too wide / ambitious to approach them solo. If LLMs solved exclusively 1 and told me "this problem is actually valuable, go and solve it", I would be the happiest person in the world. It's not true that ideas are cheap. Or better: ideas themselves are dirty cheap, good ideas are invaluable.
Even if you're entreprenurial, LLMs are useless for this.
You can't go really hard unless you can come up with something people weren't expecting, but when they discover it they want it. You can't just sit around waiting for people to want things properly. That's extremely passive.
The point being, the idea of using LLMs to find what people already passively want is itself flawed, even if you like marketing and selling to people. It's a really fundamental problem with the idea.
> Even though it's 5 different phases, it's just "3 layers" of programming.
I'd sharpen this to: it's 3 layers of software engineering (2, 3, 4), and 1 layer of programming (3).
While I agree with the premise that there are Thinkers and Shippers, I reject labeling of tinkerers and entrepreneurial. There's nothing entrepreneurial about working for a big business and shipping cool things.
But you're ultimately right, some people want to solve sudokus and some people want to change the world, and the sudoku solvers are really upset that a Sudoku Solver and World Changer 9000 device is widely available, because it spoils the fun they have with sudoku.
Because people that enjoy programming for programming's sake don't want an LLM to do the programming for them - isn't it obvious? It's just like with any other hobby, people who like car racing created rules that force you to drive yourself, even though they'd get faster lap times with electronic driver aids. People who enjoy grappling created rules that force you to grapple, even though striking could win a fight faster. People that enjoy chess don't allow you to bring a computer to the chess table. The list is endless and the people that enjoy programming will create rules that force you to program, why shouldn't they? It's their hobby.
Writing code without an LLM is akin to cooking a meal yourself, as opposed to something in between hiring a private chef and pulling a slot machine lever. When I cook a meal for myself or my loved ones, I become deeply familiar with the ingredients and the recipe. I am a good cook. Every decision, small and large, I make myself. Using feedback, I can tweak the components and the process. However, if I order food and call it my own creation, I’ll be speechless when my nut-allergic friend asks me if I used peanut oil or olive oil. And does it taste good? The end result is subjective, not binary, like your three similes. What if the food sucks? Must I now micromanage the chef? I don’t want to make them cook it again and again, hoping each time it’ll taste little more like how I want it to. I’d rather just cook it myself; I am a good cook, afterall.
This is actually a great analogy. And I think it works very well. Because what llms deliver is not the exquisite dish, the carefully crafted solution.
It’s just a solution that might scratch the itch but never fully blow you away. One time use software essentially.
If that is good or bad remains to be seen.
Well to me using an LLM is akin to using microwave for food, and programming is getting your hands dirty in the art of cooking. I used to do the latter but the pace is too much, so I'm forced to go for the former. I still miss the days where I can cook myself
This is a good analogy. Also I hate cooking; it feels like a chore and waste of time. Given the means, I prefer to have someone else do it for me so I can focus on more interesting things.
If the analogy is good, does it mean AI boosters hate programming, and feel like it's a chore and waste of time? Yet they keep calling themselves programmers.
I dislike cooking; I don't call myself a chef, nor go on cooking forums saying they're all idiots for churning butter as a hobby, and telling them they are luddite dinosaurs just because I bought myself an Instant Pot.
And clamoring that instant pot is capable of doing the job of 10 chefs, while conveniently forgetting that chefs using the instant pot is way more useful. And then congratulating each other on new way of adjusting the temperature and timer of the instant pot.
LLMs are more like having someone else microwave something for you, rather than having someone else (who knows what they are doing) cook for you.
I don't see it the same way. I see it as akin to building a machine which can do the cooking for you. There is a lot of experimentation in terms of how you can construct that machine and how you can set it up, and for some people the payoff is in exactly this process.
> I see it as akin to building a machine which can do the cooking for you.
More like buying a microwave and frozen food. Unless you are actually training the LLMs.
That's me. I've just moved the puzzle to figuring out the CC optimal "language" and mode of working with an LLM. At the same time, I haven't lost my love/expertise... Yet.
I explained it yesterday that it's no different than Linq powered by EF Core. You still need to understand the database and how SQL will work, or should. If you don't, you're queries will be slow and inefficient, hazardlessly oftentimes.
LLMs are an impressive abstraction, but they are not magic...yet. or maybe it is, but you still need to figure out the incantations. Not that I like to uptick and witchcraft, so I hate that analogy.
My boss tried to make me interested in agentic programming by showing me how "productive" the creator of OpenClaw is.
All I see is a fast food factory. Sure it's interesting that you can produce so much so quickly. But what is produced is endlessly uninteresting to me. It doesn't help that all the harness building is just SEO in disguise. Trial and error.
I don't doubt that agentic programming makes it possible to produce faster, but it also comes at a great cost for me, which is that I largely don't enjoy my job anymore. If I was given the choice between that and halving my salary I would have chosen the latter.
Yes. And (too) many people focus on the effort, or the LOC produced. Completely ignoring the value the thing produced actually brings.
Yeah exactly. My go-to example is that if you enjoy playing guitar, you're not going to stop doing it because Suno exists now.
In addition, say if you had like a guitar-playing forum, where people share recordings of their practice. People would likely not appreciate it to share AI-generated audio of guitar-practicing there, either.
And while there are various reasons for a guitar forum that might not hold for a programming forum, one of them remains: There's nothing to talk about when it's AI-generated. What scale did you use and how did you tune it? "I dunno, AI did it"
All you can talk about is to say "oh cool that's cool that the algorithm machine you're paying subscription to access made you that thing, nice"
Or for example, even when a friend of mine vibe-coded a "synthesizer" app and he wanted to show me cause he knows I like that kind of thing. And I do, I love digging into and experimenting with DSP algorithms, but also the user interfaces and all sorts of things digital audio related. And I love talking about it, or explaining stuff. But my friend couldn't tell me anything (not even that it was using WebAudio, but also how). Anyway it made me a bit sad, and I felt like something is lost here. Something worthwhile. And I explained it to him that then all I could really say was "oh cool that's cool that the algorithm machine made you that thing, nice", and we talked it over, and it was alright. But still, it seems something is lost.
This is true, but as usual nobody can agree on what "doing it yourself" means.
Few cars are manual steering, manual brakes, manual transmission.
Few grappling bodies go all Greek-naked.
Few chess bodies forbid researching previous games or your opponent.
The list is indeed endless, and every single person thinks they are using just the right amount of assistance; anything less is pointlessly difficult, anything more is cheating that takes the sport out.
Me, I've been a hobbyist programmer since youth (6502 assembly, represent). I'm having a blast "writing" software without touching a line of code. Sure, I'm old, and maybe I'm like those guys blasting Veyrons through highways, relying on traction control and all the computers to let me drive beyond my abilities. And... ok? I'm having fun, even if I'm not doing it right, by random anonymous peoples' views.
> I'm having a blast "writing" software without touching a line of code.
And that’s totally fine, enjoy your hobby however you enjoy it best. Most people programming by hand for fun don’t care that you program with an LLM for fun. The problems come when those two worlds collide, like in the linked GitHub issue, in chess engine leaderboards. In the long run for both groups to have their fun and friendly competition you’ll need two separate leaderboards. This is also very common in other hobbies. In racing there are car classes so that your Veyron and someone’s old Miata don’t go head to head. In combat sports there’s weight classes. Most sports separate amateurs and pros. Usually there is respect across the group boundaries and appreciation for the other groups deeply specialized skill. There’s even the concept of cross training to get better in your flavor of a sport. This will be figured out with programming as well.
Sure, I agree, and it’s already figured out. The demoscene has been around for decades. That’s real programming.
But I think you’re on to something important: framing programming as a competition of intrinsic skill where rules are necessary to equalize. That’s certainly true in some cases, and more power to those who enjoy it.
I guess I’m just more of a product guy these days, where the equalizer is the market and the consumers rather than the implementation methodology.
You never went to a programming competition in school? None of what was said is novel, it’s all been around forever, the tricky bit is the regurgitation software mucking everything up.
Not that it’s about me, but I was doing programming competitions 25 years ago, half of which were on paper with no keyboard allowed.
One of the great things about hobbies is that you can enjoy them whatever way you want. But honestly is it not more accurate to say that you are now a hobbyist prompter, rather than a hobbyist programmer? I get the "its just a tool" mentality and I guess LLMs can be used that way, but if you are not touching a line of code yourself then it seems less analogous to using power steering and more analogous to just taking a taxi.
What I really don't understand how differently it seems that AI for programming and for anything else is treated.
I don't think there is insignificant group who would say that using LLM to write a book does not make you an author. Or make a song musician. Or draw a picture a painter or artist.
This just somewhat baffles me... Or maybe the LLM for programming people all believe that you are as much artist if you use AI in those fields and the product generates as such is as if not more valuable than one by more traditional means.
I've seen people here insist that LLMs enable everyone to make music or art. It's diametrically opposed from my own view that making something is the process of making something, neither the result of nor the motivation for that process.
My retort is usually a variation of one I've already posted in this thread: they do that to the same extent that taxis enable everyone to drive, which is something most people will intuitively understand as nonsense, even if they've given clear instructions to the driver and even though they've arrived at the intended destination. If you aren't driving, you aren't driving. If anything, the taxi service enables more people not to drive.
isn't this clearly a misunderstanding between you and hypothetical AI art approver?
They're talking about the result, one must think. The idea that people who could not have approximate pictures of ideas that lived in their head can now have those things. They didn't create them by an in depth process, of course, and would be wrong to pass it off as though they did. But there is some value to the enjoyment gained from that existing (whether that relatively trival enjoyment is offset by other harms is for the reader to decide).
You're talking about the process and/or the developing of the skill of crafting and thinking about images better.
We also lack words to talk clearly about the distinction here around "make". Did I make a piece of art if I created a file on a computer in photoshop and then printed it? or did the printer make it? Did photoshop make it? (imo, all are true, but in differing senses of the word make). It seems to me that there is a sense that, no matter how trivial their involvement, a user of some AI art app is "making" something, in the same sense that steve jobs "made" the iPhone. In both cases, an engineer that literally designed the iPhone or an artist that spent years would perhaps rightfully feel upset by such characterization, but to get across the point cleanly is harder.
It's the same for the taxi. If I can't drive (perhaps I never learned, or some flaw with my body doesn't let me), a taxi does enable some things I could never do before. I am, of course, not literally the one driving. But if I can get to the shop myself now, there is something there.
> They're talking about the result, one must think.
Maybe they are, but in that case it's a frequent source of equivocation because people will conflate the process of their making with the result of some other process so as to proclaim themselves artists, musicians, programmers or whatever for writing something that they themselves wouldn't consider art or music unto itself in the first place, and having a machine generate something which they do consider to be art or music based on that.
> Did I make a piece of art if I created a file on a computer in photoshop and then printed it?
Funnily enough people don't usually have a problem with this question when it comes to driving. Your actions' effect on the propulsion of the car are ultimately indirect, and there are many systems which assist the person behind the wheel even when they are engaging directly with all their means of control, e.g. an engine, ABS, EPS, yet no one will immediately feel a need to wax philosophical about whether the driver is actually driving or not because of that. We recognize that in a car which requires a driver at all there is substantial ongoing involvement from the person behind the wheel in making the car go as desired and deciding what means available to apply in order to go about that, without whom there is at best no going anywhere at all, and that involvement is what driving is.
> If I can't drive (perhaps I never learned, or some flaw with my body doesn't let me), a taxi does enable some things I could never do before. I am, of course, not literally the one driving. But if I can get to the shop myself now, there is something there.
Which is not the thing in question. If AI tools have generated something you find beautiful or interesting, by all means enjoy that for what it is. If the taxi takes you to the store, by all means enjoy that.
A renowned author would still write. They don't suddenly start managing a team of junior writers and instructs them how to write.
A programmer that has climbed up the ladder in their field, on the other hand, often stopped coding themselves. Instead, they manage others who code, or provide vision/direction on the project.
I think it really depends on what people on the senior level of your field do. If they mostly move to managerial roles, then it's much more likely they'd embrace LLM.
Ironically I observe this in academia too. Professors (who often stopped doing hands-on research, but only provide guidance to postdocs and grad students in their research) tend to embrace LLM more than the fields you mentioned.
Sure, I’m both a professional and hobbyist prompter. But then, that’s been true for decades, since I made the switch from engineering to product.
Only difference is I’m prompting LLMs rather than engineering teams. But honestly it’s not all that different, except that in the hobby flavor I can write specs that would take tens or hundreds of engineers to implement.
Maybe it all comes down to semantics: does “programming” mean typing in code, designing app architecture, or producing software? When I was young I was probably 75 / 5 / 20 between those. Now I’m 5 / 20 / 75. Maybe it merits a different word?
> This is true, but as usual nobody can agree on what "doing it yourself" means.
Not everyone has to agree. Just everyone in a particular group at a particular time. Grappling competitions exist and so do MMA competitions, and both of them will expect you to follow their own rules.
But LLM usage isn’t equivalent to power steering, it’s what a full kitchen is to a microwave. You can sure put effort into a microwave meal but it’ll always work against you and much of the time you’re better off just picking up a pan.
Not to diminish your enjoyment, you’re completely entitled to it, I just think this is a bit of a false equivalence
An LLM is more akin to a staffed pro kitchen than a microwave.
Agents are effectively a full McDonalds kitchen, yes.
...and yes, I know you don't mean that. I just disagree.
> Agents are effectively a full McDonalds kitchen, yes.
You just insulted everybody that works at McDonalds.
You’re right. I apologise
Fully agree with this. Also did lots of 6502 assembly, and also love doing agent-driven development. I also still like writing code manually and tweaking everything for myself, but those serve two different purposes.
Sometimes I'm having fun getting to the output I want and otherwise wouldn't have had time to. Sometimes I'm having fun exploring the code.
I don't care if someone don't see the value in one or the other of those, and to me they stand on equal footing as things I find both fun and valuable for different reasons.
> anything less is pointlessly difficult, anything more is cheating that takes the sport out.
https://c2.com/cgi/wiki?BlubParadox
> Few cars are manual steering, manual brakes, manual transmission.
Tell me you are not in European or African continent without telling me.
In Europe you would be hard pressed to find a car without power steering or assisted brakes.
Still plenty to acquire around in second hand.
Not really.
Actually yes. Come to southern or eastern Europe, on plenty of those side road dealerships with signs on their windscreen, or parked in random places with a telephone number on them.
I'm in Central Europe and there aren't many manual transmissions around anymore where I am. But if you paint with a brush of two entire continents, you might paint a little too broadly.
I’m in Central Europe too and most cars are manuals. Must be just your neck of the woods. But I agree it’s a pointless debacle.
I think it's mostly the hybrids and EVs that are killing manual transmission. They were very common 20 years ago.
Unless you are driving an EV, or lucky to own Audi, Volvo, BMW or Mercedes top class of vehicle, most people are on manual.
In the UK, it's about 30% automatic on the road, but automatics have outsold manuals for years and about 80% of new cars sold today are automatic.
Doesn't compute with 30% on the road and 80% sales, also I doubt middle to low class cars are being sold with automatics.
Unless those 80% new cars are EVs.
Of course it does, half the cars on the road are over 10 years old, when manual outnumbered automatic
My 21 year old car in the uk has power steering, and it was low end when new
I used to write code all day, solving maths and logic problems which is what my brain loves to do, and so I loved my job.
Now I talk to a robot all day, trying to figure out how best to tell it whats in my brain. I no longer do maths and logic and instead I think about language and questions all day, holding in depth conversations and thinking up responses. This is something my brain naturally doesnt like to do and so I fatigue quickly in the day compared to my old job.
I want to do maths and logic again :(
I enjoy programming a lot, but still made the transition to using AI for my programming. It is a new tool that makes me much more efficient, I finally could continue working on my many open source projects that I couldn't touch for many years due to lack of time. My productivity is up, I'm pretty sure the code quality has improved, and I'm learning more than ever before; just in a different way.
for me its that software, wrought by hand is "bespoke". Yes, there is a loooooot of software that is just commercial, quickly vomited by hands that are just in it for the money. On the other hand there are open source projects who might have a miss guided design philosophy (Darktable comes to mind) but are otherwise build from people that care. There was an "artisan" at work, someone who cared.
When I see all those "show HN" projects that are slot level code it feels like they dont care about the result that much and just want to showcase their idea in a more tactile way, but there is no actual proper execution to make the thought into matter so to speak
True, but it sounded a bit paradoxical to me that the people who work on engines designed to beat humans at chess were themselves opposed to using engines that (arguably) beat humans at programming...
Exactly one of them threatens their lifestyle.
It can also be a personal preference for minimalism vs. maximalism. Or many other things.
Or Warhammer figurines painting, let’s buy them already painted
If you don't want to paint, and also don't want to pay somebody else to paint them, there's always various alternatives like printed out papercraft stands. Fold them, stick them on the base and you are ready to roll.
There's also a myriad of other creative ways to proxy your way out of the general costs (time, space and money) if you check out the term "poorhammer".
You can take already painted warhammer figurines, pose them, take a photo, prompt an AI to make that your first scene, and generate a hollywood level scene in about a minute or two.
Yes, sure, you can buy the figurine and get the joy of painting it. But there's new things you can do now too, things you couldn't dream about before.
But I'll grant you this, the AI coding extinction event is so severe that I do believe we need to support anyone trying to preserve any memory of it. Like, we need to build entire museums, one for each programming language.
> You can take already painted warhammer figurines, pose them, take a photo, prompt an AI to make that your first scene, and generate a hollywood level scene in about a minute or two.
if that's so awesome then why isn't everybody doing it?
I never dreamt about stuff like that, but are there actually people who are like "oh thank God I can finally make my AI movies with my figurines and other stuff I have laying around"
honestly you sound too enthusiastic about this, like Sam Altman "AI podcast about your kids" tier enthusiastic
People are doing it, but not that much, because while it generates some pretty good five-second scenes, it fails terribly at generating any long-form video. And by long-form, I mean 'longer than 30 seconds'.
I think what people want is to tell a story that's in their imagination, and no current AI is up to that task (at least in video format). Great for making an animated PNG for a meme, but not for adapting that fanfic you are writing.
What's the point of being this rude? People are doing things like this, you just have to look around YouTube.
This is what I don't like about the anti stance, it's just so often dismissive when people are excited about something.
But would you blame the figurine painting community for not wanting that?
If you want to do AI movies, join the AI movie community.
The hobby programming community isn't about vibe coding, that's what the vibe coding community is for.
People programme for 2 reasons:
1. Some programme because it’s the one of 2 ways to get the things they want (another way is hiring a team to code for them). Now that they have vibe coding, they’d happily never touch a line of code again. 2. Other people programmes because they just enjoy it. Vibe coding ruins the best part for them. Like telling a foodie, "Good news! I built a robot to eat your food for you."
Most people getting paid to code are probably somewhere in the middle?
But nowadays hand-writing code at work is starting to feel kinda 'illegal'. The manager is gonna ping you just because you fall behind on the company's token waste leaderboard.
Maybe it's just another "hobby should never become job" story?
> Like telling a foodie, "Good news! I built a robot to eat your food for you."
Also like building a robot to cook food for you. Sure, it's fine for mass produced meals, but it won't replace a top-notch chef.
Hobby should only become job if you enjoy the craft of it and if the profession / job rewards craftsmanship.
People in it just for the money won't care about craft - you see the same thing in plumbing. And when there's enough money and scale on the line, craft becomes progressively more niche - that is why Ikea furniture is a thing. And you can even say that there is craft in Ikea, just not along the axes a traditional furniture maker or user would consider.
This is such a great point and raises existential career questions for everyone who got into a programming career because they love the craft. Either find your niche where craft will matter until you retire, leave the profession or accept that it’s now mostly about prompting. It’s almost like everyone’s being promoted into management or tech leading and can’t refuse…
I think it was implied through the conversation but the other thing I have come to realise is that your hobby and craft should be about producing not consuming. Too many people think they like videogames or films and think they should therefore work in the industry only to be crushed by the industry. Even being a youtuber devolves into hours of storyboarding, editing and engagement farming which are orthogonal to your hobby.
With respect to code it's a little murky but I guess the tension really is enjoying the process enough while being able to output the desired quality, volume and velocity required of you.
Quality is usually the stickler because it's invisible to the end user and requirement setter but sometimes quality affects both ergonomics and future delivery and sometimes it's purely self-satisfactory. That isn't a value judgment, making things well is a gratifying and edifying experience - so much so that this practice overlaps with Shinto philosophy.... but not many people are going to pay you to have that experience.
We're not that far into this new era and I already see people go back to hand-writing the important bits and manually designing their architecture because it turns out that unless you have them spit out yet another boring web app, they aren't very good engineers at all, their errors and bad decisions are very hard to predict and accumulate over time, and the resulting mess becomes unmaintainable at some point, and it appears all the harness engineering in the world can only stave that off somewhat, not prevent it from happening. So perhaps IT won't become a monoculture of agent prompters after all.
"Because people that enjoy programming for programming's sake don't want an LLM to do the programming for them - isn't it obvious? "
How do we filter out these people during an interview?
why would you want to filter these people out? surely those that consider programming more than just a "means to an end" are going to be much more valuable employees, consider they are much more likely to be skilled at their craft? im genuinely interested in why you'd want to filter these people out!
I want to find people who have already had all their passions and emotions stamped out by the endless grind already, so that they will tell an AI to do what I tell them for 996 without complaint. These people who "care" about things just get in the way and want time off and such.
> so that they will tell an AI to do what I tell them for 996 without complaint
Just start sponsoring H1B visas
Offer low pay and start using lines of code as a metric. Not only will you not hire such people, you'll also filter out any who might already be on your team!
(Assuming "these people" is a euphemism for "skilled engineers")
The article just straight up doesn't mention the context of the Github Thread, which is about the guy lifting ideas from other engines and copy-washing them.
"Do you genuinely believe changes like cd170a4 make the code somehow not derived from Reckless anymore? It appears to me like it's just trying to obfuscate its origins, to make it easier to appeal to plausible deniability. The reality is, coda 100% has code derived from AGPL or unlicensed engines in it, and pretending otherwise won't convince anyone other than maybe yourself. This, combined with the fact you conveniently moved the plagiarism roadmaps to a private repo mere days after people started raising concerns about it, reads to me like you're not at all concerned with actually using code fairly, and much more with being able to claim that you are, backed up by these nonsense "cleanup" commits."
The README (https://github.com/adamtwiss/coda#license) addresses concerns about copyright in good faith and better than most agentic-driven repos:
> After our initial 0.9.0 pre-release (July 2026), community feedback raised that Coda contained AGPL-licensed code. This prompted a full audit of our codebase and dependencies, and steps to correct what it found. We now believe Coda complies with all applicable licenses. If you have any concern, please open a GitHub issue with the details and we'll happily investigate.
The debate is about whether or not the new code qualifies as derived.
They can remove all of the literal code matches from the codebase, but if it was written by having an LLM study the original code and then write something similar, it's hard to argue it's not a derivative work.
Genuine question: if a human did this manually would it be considered derived?
https://en.wikipedia.org/wiki/Clean-room_design
There is a lot of legal history of doing new designs with influence from reverse engineered sources. The basic rule is that you can't have someone familiar with the original code involved with writing the new code.
LLMs are a new area, but I would expect that feeding the original source into an LLM context and saying "Write this, but different" wouldn't qualify as a clean-room design. There isn't enough general removal from the original subject matter.
Specifically about "humans": my understanding is that Clean-room design is not a requirement and the "Case law" section on your Wikipedia link explains that and has examples.
How and whether the same principles can be applied to LLMs, I have no idea. I imagine it would involve discussions about creativity, for example.
Not a lawyer.
The thing that happened with "clean-room design" is that one company did it and the court said it was allowed. Companies are risk averse. Once they know one thing is yes and something else is maybe, they go with the yes, right? With the result that the supply of court cases where something other than that happened tends to dry up, leaving the other alternatives to stay as maybe.
If some people have now decided they're not going to blink in the face of a risk, we might get to find out whether some of the maybes turn out to be a yes or a no.
would using one LLM to create a spec, and another with only access to the spec doc count as clean enough?
Only if the second LLM can be proven to never got trained/distilled from a model that trained on the original source code.
Probably impossible if the original code is publicly available since most code oriented LLM just gobble them all.
I have done this. Despite using Fable for the spec, and having it double-check the spec a couple of times, it still wasn't airtight. And at some point, the implementing LLM accessed one file of the original code on the Web.
I have by now deviated quite a bit from the original (well, original...) spec.
Would have to be proven in court. Look up history of cases and settlements for GPL violations and it quickly becomes apparent that it’s relatively toothless and mostly license theater.
In my opinion yes. Just like if you watch a Bob Ross episode and copy all his brushstrokes, you have made a derivative (in the literal sense of the word) of his painting.
https://www.paulaoki.com/.admin/930108.oppose.html
https://en.wikipedia.org/wiki/UNIX_System_Laboratories,_Inc.....Mechanical transformation is always considered derived.
The difference is that if humans do it, there's plausible deniability that it is derived.
I mean, even when humans do it, we still do clean room engineering.
It goes further though: LLMs are just a tool, and tools don't get rights, humans do.
Humans having the right to learn from a few dozen books and projects is not the same as a tool that infers output using all copyrighted works know to man.
It could be the same act - but not necessarily legally the same because of the actor.
Yeah, absolutely. The defense there is "clean room reimplementation ". If anything, the use of an LLM would be stronger than a human doing the copying. (Assuming American jurisdiction!)
Clean room implementation is not a concept in copyright law. It's a defensive legal strategy to avoid expensive lawsuits.
Copyright law does not require a clean room implementation when copying non-copyrightable elements of a work, such as algorithms.
I still think this question is a future minefield, as well as the question of what happens if an LLM outputs a patented algorithm
This is not really a question. Accidental patent infringement is still patent infringement. Just because you used new trendy technology to do it does not mean that it isn't infringement.
Is the bank lending you money "in good faith"?
It's so weird that they think they have to comply with copyright in this case, but not in the case of any other license
"We investigated ourselves and found no wrongdoing"
Imo without the sign off from Reckless it's suspicious
> lifting ideas
It's not like ideas are protected by copyright though. Copyright protects the specific expression of the idea, not the idea itself. Anyone can "lift ideas" from anyone else, including proprietary software.
You're attacking his word choice, but the context is clear that it would be protected.
https://groups.csail.mit.edu/mac/projects/lpf/Patents/patent...
Except "Claude, take this code and change the variable names up a bit" is not lifting ideas, and you would not get away with doing that to software belonging to a company with a competent legal department.
> "Claude, take this code and change the variable names up a bit"
Is it a fact that they did that?
> lifting ideas from other engines and copy-washing them.
You cannot copyright ideas. And that's a good thing. Remember the "one click checkout patent" saga? That is bad for everyone involved. We don't want that, as an industry.
You’re mixing patent and copyright. You can patent an innovative algorithm in the US. And the implementation is of course protected by copyright
That comment is not posted yet in the github thread when the article is published
Keeping in mind that I'm going strictly off the commit[0], a bit-packing layout is very much the kind of thing which is not subject to copyright in the first place. Neither are algorithms. Ideas, in general, are not subject to copyright: expression is.
So "lifting ideas from other engines", as you yourself put it, is entirely licit, no matter how you do it. _Should_ you credit others when so doing? I err on the side of yes, on that question. _Must_ you? No, there is no such requirement, this is not academic publishing. Does it constitute violation of license? Absolutely not. Straightforwardly, no. Furthermore, it _should not_, we would not like the result of that.
Think: SCO suing Linus Torvalds over his use of a.out: and winning. It would be bad. Even though I understand the reasons for it, it's still eerie to me to see hackers coming down hard on a side of an issue we spend decades passionately defending the other side of.
As I lead this post with: maybe there's something in the repo which is infringing, it's quite possible that commit doesn't tell the whole story. If so, keep in mind that I don't know that, and am able to change my opinion in the face of new evidence.
Based on what I _have_ seen, no: you do not get to lock down a bit-packing pattern just because you're the first to come up with it. Nor should you have that right.
[0]: https://github.com/adamtwiss/coda/commit/cd170a4ea9993cf8246...
You'll notice I didn't mention copyright, and that's because I don't think it's relevant. I think your analysis is correct. Keep in mind too that Stockfish – the engine I work on, although I wasn't at the time – was the subject of a high-profile defense of the GPL in Germany: https://stockfishchess.org/blog/2021/our-lawsuit-against-che.... That case was different as it involved wholesale copying of the source code, rather than porting ideas. Ideas shouldn't be copyrightable.
But copyright and plagiarism are orthogonal, and questions of morality are much more tied to the latter. It would be illegal (in the United States) for me to publish a copy of Nineteen Eighty-Four, but not immoral. It was not immoral, in my view, for Aaron Swartz to try to liberate JSTor articles.
That doesn't make plagiarism acceptable, either. And therefore, I'm willing to call it out when I see it.
No idea why you're getting downvoted. You're right.
> it's still eerie to me to see hackers coming down hard on a side of an issue we spend decades passionately defending the other side of
Yeah, it's surreal. It's not just intellectual property either, Hackers have already started complaining about automation now too.
Computers must have replaced tens of millions of jobs. It's the ultimate job destroyer, since it allows capital to do more with less. Automating me, a doctor, out of his job is essentially the holy grail of Hacker News.
But when hackers are automated away, then suddenly it's a crime against humanity.
I think that's true, but I'm not sure if this would characterize itself as "lifting ideas". I mean, it's one thing to personally take the idea and re-implement it yourself elsewhere. But I have to believe there could be an argument, that having an inference algorithm take the original source and re-implement it in the context of your engine might count as derivative work.
A particular work is deliberately supplied at inference time so the system can recreate or adapt it for another environment. It resembles commissioning a translation, port, or derivative adaptation through an automated tool, even where the output is not verbatim, so it seems it could hold a candle in court.
You have a few scenarios now up for legal debate:
I think #1 and #2 have open legal battles on it, but #3, which seems to be this scenario, is new to me, and I feel it could have legal ground for being considered infringing on the copyright.Sidenote: how is your account "samatman" and created "April 24, 2010", that's a really funny coincidence
Whatever the legal outcome, whether it's ultimately decided that people can or can't "get away" with using LLMs to license-launder code, it's undoubtedly a dick move.
Programming is nice because you have lots of different scales of structure to think about, and your program slowly comes together. You see it improving over time as a result of your efforts. You also build your skills and learn. It’s really a nice activity.
AI systems basically change that whole formula. I’m not sure how people get joy out of constantly vibe coding new projects. It feels like binge watching Netflix shows. I feel like there’s no material value in my 50th vibecoded demo since everyone in this world can vibecode now. Sure I’ve got my custom todo app which is nice, and my custom game plug-in with some neat features for the game I like, and a few other custom apps. Those are mildly useful to me. But every second I spend vibe coding something that’s supposed to be a success for others is just invaded by the thought that even if this thing is any good, someone with real capital can just direct a future AI at it to take it from me, so I’m not really gaining.
So I think I’ll just stick to using AI occasionally, without letting it ruin the programming activity I love by spoiling it. Due solely to AI, I don’t expect I’ll ever do anything in my life which will impact another person. That’s a personal tragedy, but in a sense I was going to die anyways and so will everyone else, so it doesn’t really matter.
>I’m not sure how people get joy out of constantly vibe coding new projects. It feels like binge watching Netflix shows.
It's just power fantasy. It's amazing the first couple of times but as it wears off and becomes the new norm you forgot how being the one who decides which thing does what feels like.
Luckily (or unluckily?) the industry only cares about the end result, and flushes out people who used to reward for their passion and dedication naturally.
A lot of this is from non-tech workers wanting to expand their repertoire as well.
They have no clue what they're doing programming-wise, and in some cases this is perfectly fine, but after years of SaaS, some people are taking it to extremes, and spending millions of tokens on things that could be a spreadsheet or a note in the calendar.
Just recently we had someone deliver a report in the format of a 200.000 LOC application that requires several clusters to run, for example. It does all querying in real time from clandestine APIs, or by crawling existing systems. The person who asked for the report asked for it not to be deployed.
I get nothing of the power fantasy, of it, but I do get joy out of vibe coding.
As I've gotten older, the amount of ideas I have here and there has remained constant, while the time I have, or brain power at the end of the day when I can work on my passion projects, has diminished. What was once basically a second day job at night became family time, baths, dinner, etc. So while I may be too exhausted to sit and code, I am not too tired to sit and type my ideas, or discuss a single feature I want to see, or review what's there. I get joy that ideas are getting out of my head faster despite less brainpower to get them there. I can always grab the reigns and take over myself, but that's not the kind of thing that wins against going to bed.
> I’m not sure how people get joy out of constantly vibe coding new projects. It feels like binge watching Netflix shows.
I get joy out of it because it solves actual problems I have that I don't have time to solve manually, or don't want to. E.g. I just recently had Codex add a mechanism for two-level virtual desktops because I wanted groups of them. An hour later I had two-level virtual desktops in my wm that worked exactly how I wanted it, with support for it in my desktop switcher. I could write it myself, but why would I? I love writing code for various projects, but sometimes I just want the outcome.
For me that’s a slippery slope. It’s not really clear what parts of my hobby project are the good parts and what parts are unfulfilling. When I start including AI too much I end up feeling that none of it is fulfilling. When I do things manually I feel pride in the details however, even though those same details on another day would make me feel like I’m wasting my precious life.
I spent 2 months having Opus build a demo for a project I dreamt of for 15 years, but somehow now I can’t feel interested in it. I’m trying to avoid this. A good use case for me has been using AI to help me acquire all primary source documentation on a topic and then prepare a curated sequence of study based on a small demo I’d like to build.
Keep in mind I’m talking about how I’m approaching creative programming in my personal free time.
To me it's very easy: If I want to do a change by hand for the joy of doing that, I do. But if I want a piece of functionality now and don't want to set aside time to build it, I let an agent do it.
It boils down to whether I dream about the outcome or how to do it.
E.g. I have no emotional attachment to my window manager - I wrote it (by hand, originally) because I wanted a specific set of functionality, not because I particularly wanted the experience of writing one.
I've recently made lots of changes to it with Opus because I wanted another set of functionality and was elated not to have to spend time doing it manually because what I wanted was the functionality.
I'm now a very happy user of a better window manager than I was before.
Conversely, I have dozens of projects I stay closer to, because I care about the code and don't particularly care if that means I might have to wait for years to get to something usable, if ever.
I think there’s a lot of need for people to help others understand how to healthily engage with this technology. I do struggle with turning it off once I get started on it.
I also believe I see the future coming very fast. I’ve been addicted to watching television for weeks on end or playing games for months on end. I don’t want to fall into something like that again. The vibe coding thing feels exactly like that to me.
This. And the AI detractors in this thread are the same I think, and are virtue signalling.
I’m not virtue signaling? I’ve shared exactly my experience. People experience reality differently and that’s not some affront on you.
Woops i responded to the wrong comment sorry!
In the past I’ve experimented with many different programming paradigms. I think my AI coding benefits from that experience. I want to find new paradigms to write extensible, modular systems with AI, a problem that is still open. One thing I’ve learned is that when implementing a complex system, no paradigm eliminates the need to write lots code (covering many ugly edge-cases) tests, which previously required lots of boring, slow work, but AI is very good at that so maybe can make it less boring and faster.
> I’m not sure how people get joy out of constantly vibe coding new projects.
What? It's the simplest thing in the world. You get the product. Isn't that the point of all programming? You want a thing and you speak a magic incantation and then you have the thing, to keep and use.
> even if this thing is any good, someone with real capital can just direct a future AI at it to take it from me, so I’m not really gaining.
This kind of sounds like you're writing code for social credit. Like, you don't actually want the thing, you want to be "the person who made the thing". If you want the thing for its own sake, nobody can take it from you and what somebody with capital does on their own services does not harm your use one whit.
Well I used to imagine I’d make a business out of the thing. I also would sometimes think the thing could give me professional opportunities.
Sure. I do agree that AI is a problem in this sense: programmer skill as a hiring differentiator is becoming a lot less legible these days. It just feels personally odd to be the reason you do the thing. When I encountered my first programmable computer I very quickly realized that I wanted to do more of it, and not too much later that this was the thing I would spend the rest of my life doing. That it was also very commercializable, I have always considered as a kind of happy accident.
That’s my experience, starting at 10 years old and trying to figure out how the the moving pictures were in the screen.
I didn’t say that was the reason though. A person has a lot of wants and needs right?
I’m not following why it’s strange to want to be skilled and engage with the craft itself based on that part of the message to be honest. Maybe I’m misunderstanding.
I tried to convey in my first post that I’m gonna keep programming for no material gain and no way of producing anything of value for others. I am worried about how I’ll support myself but this thread is about the hobby.
That's fair, I think I read something into your comment that wasn't really in there...
Personally speaking, the feeling of "getting the thing and then having the thing" - the personal todo app, the custom apps - is why I do anything in programming. Programming is the project where I can speak a spell and then I get a nice thing. That's very cool, and I do like the process, but at the end of the day the nice thing is what I'm doing it for- it's not a side benefit, it's the payoff. I get a burst of happiness every time that the thing quietly does its thing that I wouldn't have had otherwise. So to me, the LLM mostly just feels like I suddenly got a lot of friends who are willing to help me get the thing and also have a very complementary skillset. So it's an improvement, and then there's also a reward at the end!
Example: I'm playing Satisfactory right now, with a mod that puts a computer in the game that can control game events. I use the computer to do cargo train routing in the game, ie. it runs a program that scans target cargo depots and looks for items that are below a level, and then dispatches trains to find these items and bring them to the destination. Which I once would have written by hand, but was now done entirely by Sonnet. I did have to put plenty of effort into guiding Sonnet to the approach that worked, and I put several AIs together, and this felt a lot like "explaining something" which is a process I enjoy uncomplicatedly anyways. And I think, to be honest, the enjoyment I would have gotten out of the programming part once, was very largely "type-2 fun"- an annoyance followed by a nice reward. Now I'm getting the nice reward for free and can focus on actual industrial and rail network design. And I still get a burst of happiness every time I see a train busily speed off to a new destination. So far as I can tell, at least for me it's all upside.
I actually think that’s a great use of AI, and sounds similar to some of my use cases.
“And I think, to be honest, the enjoyment I would have gotten out of the programming part once, was very largely "type-2 fun"- an annoyance followed by a nice reward.”
Yea I’ve had similar thoughts. Other thoughts have been something like “I’ll have different types of projects and try to be thoughtful in advance about what this is doing for me. If it’s about the building up a system manually and slowly, and having that kind of experience, then I’ll only use the AI maybe as a primary source gatherer or maybe in some limited sense.” And “Okay well this is an MMO and even though in the past I would have wanted to program this plug-in, I guess I’m middle aged now and there’s only so much time and I’d probably be really excited about using my programming energy on these dozen other things I’ve always wanted to work on, and having the thing would be a strict positive compared to not having the thing, so I’ll have the AI make this plug-in which makes the game (which is about play) more fun.”
Programming is a lot of things, so the ambiguity of language probably causes a lot of conflict here.
Yeah! And now the mod has fixed input events, I can actually have a proper configuration GUI for this, and I could I suppose write a GUI toolkit for the nth time, but it kind of feels like the sort of thing that's barely cool the first time, lol. And I already know if I let Sonnet or Opus do it, it'll work fine after the third try or so, and what's more it's going to look a lot better. And I'll enjoy using it every time instead of half-dreading finding another bug.
Another strict positive I’ve had with AI is having it figure out how to deal with bureaucracy. This is dangerous due to privacy issues, but I was seriously depressed for a period, felt like I could not do it, and the AI figured out which agency I had to contact and how to contact them etc… to do some imminently important thing, and if this thing wasn’t done I would have been in serious trouble. I don’t care to know how to interface with that (and the system was pretty miserable).
In the past I guess it would have overrun and I’d be in trouble. Maybe that would have taught me to handle my affairs though. I doubt it.
> Programming is nice because you have lots of different scales of structure to think about, and your program slowly comes together.
Programming is nice cause of this for you. Other people have other preferences. When I got my first LED to blink, I wasn't thinking about the elegance of my code. Same for when moved sprites or made a async stateless compute system for MATLAB analysis. There are people who are driven by the artifacts they produce. That's just how our brains are wired I guess, outcome oriented. I see nothing wrong in that.
Btw, I feel like there is this view that all people using agents are the apocryphal "vibe coders". That is just not true. The process of creation remains the same. Break it down into smaller pieces. Work those pieces till a piece it done. Make sure it all comes together. People seriously using these agents are not writing "implement gta6. Make no mistakes".
Programming has a lot of nice features and accurately capturing them all in a hacker news comment isn’t that easy, so maybe you could have some consideration for that in the future. I also want effects to happen in the world. I also like the slow process of craft.
I’m anticipating not a very far away future when “implement gta 6” will not be that different from torrenting gta 3 in the 2000s. The current agent era feels very ephemeral. You’re choosing to interpret some version of AI based programming that you enjoy with what I think is the common interpretation of vibe coding. If you like your process then go ahead with it. I’m not stopping you.
You seem to want the AI supported gratification and the feeling of process and craft. So make your process and think of it however you like. There’s literally nothing I can or would do to get in your way.
Your comment seems unnecessary aggressive. Did I offend you in any way? I just gave my perspective.
Yea the wording that it was some way or another “for you” came across as aggressiveness so I responded in kind.
Have you seen what most people do for jobs all day? They do repetitive brainless things, not always new and interesting work. For all of those people, doing this is at least something you do sitting down alone in an office and so they will be queueing to do this.
I’ve also had menial jobs, like most people.
And LLMs are what make so many jobs feel repetitive and brainless. Jobs that were engaging, interesting, from which people were learning, developing expertise over time, are now prompting + babysitting agents
Many jobs have been repetitive and brainless since long before LLMs. I hate AI as much as the next guy, but come on.
Lots of people job wasn’t. Now it is. A lot of people in software went into this field because they actually feel satisfied engaging with systems and their implementations.
Managing a fleet of agent is the exact opposite of what is engaging about software development.
The point is most people's jobs are like that. So it doesn't matter how interesting it was or that who did it before doesn't like it anymore because there's a queue of people lined up to do the brainless new version of the job.
It obviously does matter. The industry was built by people who care about the underlying details and want to engage with the hardware and software at a deep level. That has traditionally been the way people develop expertise - the exact thing you need to effectively use LLMs for anything else than boilerplate work. Ironically agentic stuff pull us away from from that
A hobby is something you enjoy the process of doing, not just the end result. Everybody likes a clean home, but cleaning is rarely a hobby. LLMs expedite achieving the end result. Take from that what you will.
I think part of it though is that LLMs (especially in novice hands) create code that is more difficult to review and has subtler bugs. Code review is (for many) the least fun part of programming.
Like imagine if your hobby was cooking, but you hated the cleanup. If you buy a robot that automates the fun part at the cost of making twice the mess, that is not a good trade off no matter how many extra cookies you end up with.
Or imagine a club where everyone cooks together and then cleans dishes together. Then one day some people show up at a club meeting with cooking robots that create 10x the amount of dirty dishes and then don’t help clean up at all. I bet the other members wouldn’t be happy, even if the robots mostly cooked decent food.
Then robots begin cleaning dishes too. And the other members aren't happy, because what's the point, we are here not for the dishes, but for the process.
It is probably a good trade off if you sell cookies or if you vastly prefer eating cookies to making cookies.
I think that also hits what you see universally with every AI product trying to pass worse results off as a benefit: if you’re a business, the prospect of cutting your cost per cookie is really attractive as long as the decline in quality doesn’t cause people to stop buying and that’s going to get the focus of everyone who sees the financial reports more than they talk with customers. Anyone who’s going it because they care about cookies or take pride in your craft will have the opposite weighting.
We’re kind of at the fast food point: before LLMs, bring successful at all in software required some level of skill (witness all of the companies getting very expensive write-offs from off-shoring or big consultancies trying to save on payroll vs. staff), kind of like how the small cafe/bakery model needed to have at least some level of craft to have something people would buy before the option of making food at a factory and shipping it became an option.
Sure, and that probably explains why companies (who sell software) love AI while hobbyists who just like to program hate it.
For the owner/shareholders yes.
The workers will see the fun evaporate but now they have to clean 10x cookie trays in the same 8 hours.
It is, although in this case the cookie-making robot also makes 10% of its cookies with sand instead of flour. So it's still not a very good trade if you have any pride in your business or consequences for bad product.
> Code review is (for many) the least fun part of programming.
I'm lurking around the OpenBSD mailing list and it seems that code review is where collaboration starts. You send a patch, explain its purpose, people try it out and comment. If it's OK (various factors decide that), it's imported inside the codebase.
The main issue in corporate settings is that the author of the PR wants to browbeat you into accepting his/her changes. Either directly or by appealing to the manager. They don't want collaboration, they just see you as the blocker between them and the "Done" status on the ticket.
I used to work at a company that did in-person pair review. I wasn’t an engineer there but the engineers all legitimately enjoyed the process. They hired very capable people as a default, so it was watching two passionate and talented engineers debate.
What you need from LLM is recipe making.
Main course is this this and this, 1. Do this 2. Plug this 3. Run this.. step by step plan.
So you get a precise top down construction plan. Then it’s easy to split the parts where the coding needs to be manual (smoothing walls) vs where it can be whateverz (laying concrete).
LLMs reward seeing forest instead of trees.
This way one person can implement a whole new hobby project instead of only writing a single specialized part in an existing project.
Good luck making a grand strategy game in two months holidays without an LLM. Now you can.
People focused on the trees will however lag behind as implementation details matter less and less.
Yes manual touch is still needed in some parts of the construction process but architecting the whole thing is the most important in 2026.
⠀⠀⠀⢀⣼⣦⠀⠀⣠⣿⣿⣦⣀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣼⣆⠀⠀⠀⠀⠀ ⠀⠀⠒⣿⣿⣿⠓⠀⠀⠻⣿⣿⠀⢀⣴⣿⣦⡀⠀⢀⣾⣦⠘⢿⣿⣧⡀⠀⠀⠀ ⠀⢀⣴⣿⡿⠃⡄⠈⠻⣿⣟⣉⣀⠉⣽⡿⠋⠡⠴⣿⣿⣿⠓⠀⠙⢇⠀⠀⠀⠀ ⠀⠿⣿⠟⢁⣾⣿⣦⣀⠘⠿⠟⢁⣼⣿⣿⣷⠂⣴⣿⣿⣿⣆⠘⢶⣶⣿⠶⠤⠀ ⠀⣀⣀⡀⢉⣿⣿⣿⡍⠀⢀⣀⠙⢻⠿⢋⣤⣾⣿⣿⣿⣿⣿⣷⣄⠙⢿⣦⡀⠀ ⠀⠟⠋⣠⣾⣿⣿⣿⣿⣦⣌⠉⠠⣤⣤⣤⡌⢙⣿⣿⣿⣿⣿⣿⠛⠛⠂⢈⣙⠀ ⠀⠀⣉⡉⣹⣿⣿⣿⣿⣏⠉⣉⣀⣈⠙⠋⣠⣿⣿⣿⣿⣿⣿⣿⣆⠙⠛⠛⠛⠀ ⠀⠀⠋⣴⣿⣿⣿⣿⣿⣿⣷⣌⠉⢁⣴⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣷⣄⠀⠀⠀ ⠀⠴⢾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠦⠈⣙⠛⠛⠛⠛⠛⠛⠛⠛⣉⣉⠁⠀⠀⠀ ⠀⠀⣦⣤⡄⢉⣉⣉⣉⠉⣡⣤⠀⠀⠀⣿⣿⣷⠀⢰⣿⣿⡇⢸⣿⣿⠀⠀⠀⠀ ⠀⠀⣿⣿⡇⣸⣿⣿⣿⡄⢻⣿⠀⠀⠀⣿⣿⣿⠀⢸⣿⣿⡇⢸⣿⣿⠀⠀⠀⠀ ⠀⠀⣿⣿⠁⣿⣿⣿⣿⡇⠸⠿⠀⠀⠀⣿⣿⣿⠀⢸⣿⣿⣇⠸⣿⣿⠀⠀⠀⠀ ⠀⠀⠛⠛⠀⣿⣿⣿⣿⡇⠀⠀⠀⠀⠀⠉⠉⠉⠀⢸⣿⣿⣿⠀⠿⠿⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠛⠛⠛⠛⠃⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠉⠉⠉⠀⠀⠀⠀⠀⠀⠀ADHD wins over autism in the immediate future. It was balanced in the past for development to need both but now it’s a total hyperactive victory of general knowledge, master of none.
Specialists will be fed to the model weights, generalists will inherit the world.
> ADHD wins over autism in the immediate future. It was balanced in the past for development to need both but now it’s a total hyperactive victory of general knowledge, master of none.
That’s an interesting framing. I’m not sure I agree with your strong conclusion but that’s mostly because I think there’s a second axis about maintenance – you can make a grand strategy game in a couple months but it’s harder to make it balanced, fun, and to keep shipping updates which players like. A prototype or hobby experiment doesn’t have that problem nearly as much and I think that’s where we see such bimodal reactions where some people think it’s a huge productivity boost and others see modest or even negative results.
I think we’re on opposite sides but this is the first time I’ve seen someone in the wild share my suspicions that the LLM conflict is part of some vast ADHD vs autism hyperwar. At one point I was collecting posts I found where LLM boosters explicitly mentioned how it interacted with their ADHD.
The contest you described (people with ADHD vs. autism) would have an enormously negative effect. But there are reasons to think it's unlikely. FWIW at least 25% of ADHD adults manifest autistic traits. Hardly surprising, clinicians regard overlapping neurodevelopmental disorders as the rule and not the exception.
An online community works when participants learn and share knowledge gained. Contests, with self or others, disrupt learning and sharing, hence by definition constitute "bad behavior". If it exists, the hyperwar you allude to will harm communities more than all LLMs combined could do.
I'm pretty skeptical of people who are ADHD and claim to be getting boosted by LLMs.
It's hell for my ADHD, personally. It's an infinite rabbit hole generator and distraction machine. I have to use browser extensions to filter out stuff like Stack Exchange suggestions for interesting topics, too. I have had to block Wikipedia in the past to avoid rabbit holes there
I'm better now than I used to be but I am skeptical of people who claim to both be good at multi-tasking and also have ADHD. My experience with ADHD is that multi -tasking is not a strength, it's a sign of distraction
> ... multi-tasking is not a strength, it's a sign of distraction
Interesting point of view from one who's been there. Intuitively the comment makes a lot of sense.
In a way human multitasking resembles how single-threaded computer programs work. That is, we focus on one task at a time but tasks can be interleaved to give an illusion of doing several tasks at the same time. Managing several tasks requires remembering the next step to do when leaving one task to attend to another.
That suggests ADHD people lose track of where they left off when performing several tasks. Distraction plays a role, but I think problematic working memory is the major issue. Among ADHD researchers there's a saying, "people with ADHD know what to do but don't do what they know." ADHD task performance may improve with proper treatment.
Not diagnosed, on online “tests” I score just shy of the ADHD threshold, and I find myself in this. If only I properly remembered and agreed on what I wanted to do yesterday it would be fantastic. Even if I write it down and happen to remember to check the notes, it’s almost like I was a different person yesterday, todays priorities are all new
> Not diagnosed, on online “tests” I score just shy of the ADHD threshold, and I find myself in this.
Don't know the tests you refer to, but it's recognized that many adults have ADHD symptoms not quite meeting DSM-5 criteria for ADHD. It's called "subsyndromal" ADHD which causes as much disability and responds to treatment just as well as "full" ADHD cases. IOW it's absurd to regard someone with 6 symptoms as belonging in a different category vs. a person having "only" 5 symptoms. Anyway, ADHD is a condition that usually responds to treatment, no reason not to check into it.
Thanks, I know I could probably get a diagnosis and treatment for it. However, to be honest I kind of like it! I think it keeps me on a different path wrt most people, so until it becomes a serious problem (financially) I’ll keep riding the waves of focus and distraction ;)
> That suggests ADHD people lose track of where they left off when performing several tasks
Not quite. At least not for me. My problem is hyperfocus. If I switch tasks I might go down a rabbit hole and forget to change back. It is similar to being distracted, except the distraction can still be legitimate work stuff. My focus is like a freight train, it's hard to change tracks
Which can seem like an attention problem when you're trying to change me from focused code mode to "jump on a call" mode
Yes, hyperfocus is indeed a frequent problem for people with ADHD. One version is getting overly involved in irrelevant details, due to interest or having difficulty pulling away. The latter takes much effort and a good measure of courage. Maybe "rabbit holes" are kind of like that.
For many with ADHD starting tasks is hard and resuming a task can be just as hard. Constant experience of "starting from scratch" is burdensome, and contributes to inertia and inefficiency. More than not, a tough problem to solve.
I've long thought disability associated with ADHD is generally underappreciated despite decades of study proving the point.
Self-observation isn't easy. I always considered it a sign of growth when people shared what they've seen.
For me it's a dopamine machine. I get my doses of lacking dopamine from completing things.
Agents let me complete things in parallel all the time.
They also let me build tools that manage all the attention-grabbing things in my life. Something I've been meaning to do "tomorrow" for a good decade.
I've already built custom notification tools for all the "I'll just quickly check hn/lobste.rs/tildes/whatever" so that they'll tell me if there's something actionable, closing down the rabbit hole for me. My email is pre-filtered by an agent that pings me if there's something urgent. My IRC client is custom-built for the specific way I use it etc etc.
>At one point I was collecting posts I found where LLM boosters explicitly mentioned how it interacted with their ADHD.
Care to share it?
No, it felt too invasive and creepy so I abandoned the idea. Or my lack of follow-through is a sign that I’m actually on the other side of the “war”. Reader decides! But suffice it to say that it was, to me, a significant trend especially around the “Claude Code moment” back in February/March is this year.
Autism should not be framed as being "opposted" to ADHD. The correct axis is "Autistic" (Neurosis) as opposed to Psychosis. Opposite poles but with a little bit of horseshoe theory dynamics
> This way one person can implement a whole new hobby project instead of only writing a single specialized part in an existing project.
I still do start from bottom up to refactor everything into workspace members and this way, parallelism is way easier
> LLMs reward seeing forest instead of trees.
Not just forest. LLMs reward seeing structures and designs, and it still requires a lot of creativity to design it well. I have to do a lot of architect works for that, like package layout, testing strategies (BDD vs TDD vs E2E), setting up access to platforms and digging out what strategies to apply...etc.
> People focused on the trees will however lag behind as implementation details matter less and less.
Trees like demonstrations and example code are even more important. This is because those are literally post-training materials for the LLMs
> Specialists will be fed to the model weights
I don't think so. New labor laws are certainly going to be introduced by the politicans to tackle this.
> generalists will inherit the world.
Now the problem is how far do you want to stretch the systemic thinking, and you also get burned out quicker.
Funny you should mention autism. I'm of the belief that if I really wanted to, my autism could help me become a very effective AI-assisted programmer, because autistic communication traits that actual humans find annoying (and which most people hence train out of themselves), like infodumping and using precise language, are things AI actually loves and produces better results with.
But I don't want to use AI to code, in large part because I iterate on ideas by working directly with code. Attempting to program through an LLM intermediary becomes like attempting to eat using soft silicone chopsticks.
Exactly. I have a backyard garden not because I need the food (it would be infinitely cheaper and easier to buy even high end farm-stand vegetables), but because I enjoy gardening. Buying a gardening robot (if such a thing existed) would completely defeat the point.
I use an LLM to make custom software for myself not because I enjoy coding but because I want software that I can't afford to hire a human to make for me. Buying a robot to do it (current agentic coding LLMs) is the only way it gets done at all.
These two things are not the same, and should not be viewed in the same light.
All that being said, I would 100% buy a weeding robot. Weeding sucks.
These are not completely different things, but maybe different sides of the same coin.
You enjoy the act of gardening, so you do not automate it away.
You enjoy the clean home, not the act of cleaning, so you may be interested in automating it away.
It seems you enjoy designing or using software, as opposed to programming/building software. So you automate the programming/building away.
Sure, that's fair. I enjoy the process of figuring out ways software can solve some (usually quite trivial) problem for myself, but then I'm very happy to hand the act of making it to a robot.
> Buying a gardening robot (if such a thing existed) would completely defeat the point.
Unless your hobby is backyard robotic gardening.
Which is my way of saying that people have different reasons for enjoying a hobby.
Like most hobby debates, this one comes down to "Let people enjoy things"
> Buying a robot to do it
It depends what you value. There's "zoning out going through the motions" for mindfulness and then there's "building stuff!" for interest, learning and fun.
And you can have any combination of the two. Like maybe doing the same planting routine but with different details for a different plant or experimenting if a new layout works better.
I don't care to do the labour of gardening. But if I had time, I'd spend it bikeshedding optimisations. IE what about grow lights? Is it worth having a greenhouse in my climate? etc etc. And all the little details about building stuff.
Aside: I too want a robot. One for lawn mowing (they exist but expensive) and one for spraying the little lawn weeds that pop up.
How much of coding is weeding?
Probably a similar ratio!
How much of gardening is dealing with bugs?
It depends what on what you're growing.
With roses, a significant % of manual time is spent fighting aphids
>> I have a backyard garden not because I need the food (it would be infinitely cheaper and easier to buy even high end farm-stand vegetables)
...cheaper how? My partner and I grow most of our veggies and, based on our calculations (including fixed costs that have gone into the garden, as well as e.g. water, fertilizer, etc.) we save a non-negligible amount of money every month.
One can piss away a lot of money gardening with little to show for it, especially if you buy bagged soil. Buy a raised planter, some soil, some fertilizer, some seedlings, a couple garden tools and you can easily be in for $150+ which is definitely more than you'll get out of a season worth of growth even if nothing fails.
On the other end of the spectrum, if you grow in fertile native soil from seed your costs can be close to zero
Yeah, in my experience, the people that got into software engineering because of the money are absolutely thrilled with how LLMs are changing the industry. They mostly just liked getting results and LLMs are great for that.
The people like me that sort of just stumbled into this career as a result of spending the evenings of our youth trying to get the computer to do interesting things don't seem to have as much appreciation now that the LLMs are taking over the fun part.
If I'm any example, I'm on the other side. I got into computers in the '70s: plotter art, BASIC, then Z-80s, forth, PDP11s, Amigas and so to now with GPUs and supercomputers and visualization. I'm recently retired and having enormous fun plowing through my lifetime backlog of "wouldn't that be a cool tool" (https://pcons.org) or a climate visualization (https://globe-viz.oberbrunner.com) or a travel packing list app (packzen) or a bunch of 3d printer tools... so on and so on. Totally co-engineered with Claude. What we're doing together, and getting out to the world as useful open source projects, is easily 10x what I could've done on my own, in the limited time available to me in this life.
Very similar story (right down to the pdp-11s).
> Yeah, in my experience, the people that got into software engineering because of the money are absolutely thrilled with how LLMs are changing the industry. They mostly just liked getting results and LLMs are great for that.
Those people are in for the biggest shock and largest disappointment: when a bright teenager can replicate, with an LLM, what they produced with an LLM the pay is going to fall off a cliff.
At least those who really are into development to the point of writing their own OS, Language, etc can still enjoy it as a hobby.
> Those people are in for the biggest shock and largest disappointment: when a bright teenager can replicate, with an LLM, what they produced with an LLM the pay is going to fall off a cliff.
The value of an engineer is not what task they work on, but what problems they can solve. To this teen, the code is a black box, the LLM has most of the value in their pairing, and another random teen will replace easily them.
LLMs were trained on the work of engineers, not prompting teens. The profession will continue to evolve, and the LLMs will improve along, thanks to professional, manual code still being written, not to vibe-coders.
I don't think that's quite fair. Computers have been my hobby since I was 4, I'm having a blast, this is so much fun. I've always had way more ideas I wanted to try than I've had time/energy to implement, especially true since having kids, now the implementation can almost keep up.
Computers were on a trajectory of becoming more consumption oriented, and this has shifted it strongly back towards creation. Normal people can make their own custom software again. This is awesome.
Really hate this characterization, although to be fair I love the ops/infra side of things more. I've used to set up so many parts of my homelab manually, but now I can just get it all done. I've spent countless hours maintaining a Rpi server at the infancy of my interest in computing. Truly a lot of it is maintenance and busywork tasks. I am having more fun than ever. Perhaps some people like those busywork sysadmin tasks and gain enjoyment of the homelab hobby just from that...honestly I can't say I really understand it.
That attitude isn’t universal to everyone who enjoys tinkering and making computers do interesting things. I just had lunch yesterday with one of the most tinkering people I know, who as a side hobby teaches kids to program using robotics etc. to make it interesting for them. Like me, he’s having a blast using LLMs.
LLMs are amazing tools for learning new skills and technology. I’ve recently moved into an ML adjacent role, am learning tons that I never knew before, and I’m delivering working systems more quickly.
One thing that’s been an eye-opener is how many details I miss in the code I write. LLMs can almost always suggest improvements, and usually, it’s quite obvious that they’re right. But this is making me a better developer, even though I already have decades of experience.
LLMs don’t force you to vibe code everything. You can use them as precisely or as vibe-ishly as you want. They can be a skill amplifier rather than a skill replacer, if you use them as such.
I'm the same way. The "grind" makes the successes all the sweeter.
Not my experience at all. I taught myself assembly language when I was 11 just for fun, and I'm still programming as a hobby more than 40 years later, even though I also do it professionally.
And I love using LLMs. They take care of all the boring boilerplate, letting me focus on the parts I actually enjoy.
This is not strictly true.
I enjoy optimizing code to the point of obsession, but I can't find economic (daily-job) justification to spend time on this. Lately I've had several extremely productive and enjoyable sessions that go like this: - I run Claude on my desktop PC, explain the problem, and let it explore the code around the problem and build a test harness so it can run experiments - I ask it to run experiments and try things; at that point I can leave my PC and continue over my phone in brief stolen moments here and there, over the course of many hours - it comes back with results; I ask it to try new things based on those results - after multiple rounds of this, I have explored a vast search space I could have never had time to do myself - I ask it for an implementation of the most promising path; this is usually something simpl-ish (on the order of 50-100 LOC, not huge new systems) - I return to the PC and do something between polish the implementation and redo it fully manually - partly because I enjoy it
This is something I can see myself do for fun in areas I care about, not for money and reasonable business objectives for my day job.
My hobby is making things. I used to make them with a screwdriver and a hammer and a handsaw. Then I bought a drill and got some screwdriving bits, an a pneumatic hammer and an arm saw. Now I can build things much much faster.
Similarly I've made lots of software projects. Over the last 15+ years I have over 250 github repos, the majority of them no one but me will be interested in but I had fun making stuff. LLMs let me make stuff faster just like the power tools. I'm still enjoying making stuff. I still direct, I still decide what to make, I still decide features, iterate on UX, etc, but at several X the speed of before. Loving it.
I think for those of us who got into software initially for the end result -- being the only way for the thing in your head to ever exist -- enjoying the engineering part was a nice, incidental surprise.
So while LLMs solve the technical implementation side, there's still upside since now we get to materialize the dang thing we were setting out to build in the first place.
Though there's a huge sense of loss since so much of our skills and identity and time and value were devoted to the engineering side.
I enjoy molding hyper specific software across my whole eco system of tech through prose with LLMs. Iterating over time to hone the right solutions based on very abstract reactions to the results I get. I would consider this a hobby because it costs me money and I get no value out of it outside of the act.
I also woodwork and I remember this discussion there between hand tools, power tools, and CNCs. It's all the same hobby just different entry points and doing it the least automated way doesn't make your work more or less than anyone else's.
> I would consider this a hobby because it costs me money and I get no value out of it outside of the act.
I LOLd very hard at that one, thanks. Very relatable.
A hobby of mine is also woodworking. I can make jigs, I can buy jigs. I can use hand tools, I can use power tools. I can use screws I can use joints. I can even buy made furniture and just assemble it. Though the latter is more of interior designing. But the point is it is a spectrum and everyone has their own niche. SDLC are not the same and there is different value to be derived. But the process of woodworking makes me handy. The experience allows me to elegantly and cost efficiently fix broken things. It enables me to do interesting things in interesting ways and there is more of a community around that. If I automated away all the woodworking and just bought the furniture... Well there isn't much of a community around that except flexing who has more cash. Idk, I'm trying to have empathy for both those who are for and against it.
> I would consider this a hobby because it costs me money and I get no value out of it outside of the act.
Boy isn't that the truth haha, I feel like my previous phase of dotfile ricing has become "LLM ricing."
Not complaining though, it is genuinely fun!
> LLMs expedite achieving the end result. Take from that what you will.
It's an age-old debate in every hobby.
Some people are in it for the process. Others are in it for the end result, which is what they want to use or play with.
In woodworking there are some circles who will sneer at you for using certain power tools instead of hand tools. I do think it's more impressive when someone makes something entirely with hand tools. I don't have all the time in the world to work on a project, so I'm going to use the power tools when I can.
Some people care a lot about impressing others with how they made a thing. Some times that's appropriate, like if there's a competition for hand-made or human-written something. Other times, nobody cares. They're getting together to enjoy the end result, not the artisanal process used to make it.
> Some times that's appropriate, like if there's a competition for hand-made or human-written something.
Indeed, and this is the problem: unlike woodworking, where my power-assisted creation is independent of, or complements, your handcraft, this engine is directly competing with other engines. Twiss is being disingenuous when he says he has no interest in submitting it to competitions; it plays on a rated account on Lichess, for example.
Now that we can mass produce tables and chairs I would imagine there are far fewer people whose career is hand crafting tables and chairs. So machines literally did out compete them.
In fact handicraft people were out competed so hard many people forget that used to be a respectable career.
And to extend the analogy, before there were house cleaning machines the people who enjoyed cleaning were lumped in with the people who like having a clean house.
I will say that there are a lot of parts of programming I enjoy, but I used to spend an inordinate amount of time chasing down obscure errors or learning arcane systems just to get to the parts I personally enjoy: the feedback loop of making a thing, seeing the result, and improving it.
There is a reason stack overflow is dead. LLMs are a strictly better product.
I think this is why that I limit the use of LLM in my hobbyist project, but not in my work. I found LLM particularly useful of generating say VSCode configuration JSON files and it saved me a lot of headaches. It is also very good at finding the right tool for the job (e.g. which binary to use to write a floppy image).
LLMs expedite achieving AN end result.
For this specific example, cleaning your home counts as exercise and you are proud you put effort and got a good result.
It is an example of something that has hobby-like characteristics.
LLMs let you generate code quickly. This seems to have very little relevance to the end result.
If we accept a very simplistic view of how people think and feel, then sure.
I don't think this is correct.
What is the "end result"? You refer to cleaning here, but many people use AI spam slop to bombard people with garbage text. That is not "cleaning up", that is literally wasting the time of real people here. And that's just one example of many more that can be given of AI slop not being so great.
> LLMs expedite achieving the end result.
Yes if your end result is technical debt and endless bullshit. (Some enterprise environments do be like that.)
What’s missing here is an understanding of the negative effect the rise of AI has had on programming communities overall: both the decrease in positive engagement as people have moved from discussing things and helping one another online to chatting with AI, and the increase in negative engagement. The amount of low-energy abandonware being “shared” into these communities makes it really hard to filter out the signal from the noise.
So at least in my experience, it’s not about gatekeeping as much as it’s about keeping the experience bearable for the existing hobbyist community.
To be fair, the decrease in discussion and helping one another online has also decreased some significant amount of negative engagement in the form of nonsensical arguments and frustrating ways of responding to questions (e.g. the biggest complaints with stackoverflow).
It was like this with cryptocurrency to a lesser extent too. Suddenly lots of people pretending to have technical background and explaining what a Merkle tree is.
And telling everyone if they don’t embrace cryptocurrency right away they will be left behind, how you are all limiting human flourishing by refusing to embrace the future, excusing every single negative side effect of the technology — because we have to move as fast as possible. And of course no words on how unsustainable the businesses are, how corrupt it is to have so much power and control into a few really shady individuals
> So at least in my experience, it’s not about gatekeeping as much as it’s about keeping the experience bearable for the existing hobbyist community.
Keeping the experience bearable BY gatekeeping. That is the point, gatekeeping is not inherently a bad thing. It's a required thing in many circles.
Is it gatekeeping in this case? Would they not allow beginners to join and learn? Are they not fully sharing the knowledge to anyone who's interested to learn it?
To me, it amounts more to disapproving showing up to a handmade sweater community and showing off your machine made sweaters, or maybe a machine made sweater you added a handful of stitches too yourself. It's like not realizing that the community wasn't about cool looking sweaters, but about hand-made cool looking sweaters.
Having said that, there might be another community that's more about cool looking sweaters, and new machine techniques that deliver really nice looking sweaters would be of interest to them. I actually think it's what is happening a bit in programming, there are niche that remain more about the manual programming, and others that are about any process that can lead to better outcomes.
The term gatekeeping gets thrown around so much now that I feel like ive forgotten what it even means. What they are doing is having rules. All communities above a certain size need rules to maintain structure. Idk if that is or isnt gatekeeping, but it is definitely using rules to maintain their identity.
Yep, gatekeep your hobbies! Just not too much. Programming/CS have been good about that so far.
The entire AI-pilled side-software development has a single goal: shipping as fast and as much as possible. The problem is that it’s not better code or a better product—it's just more
I can understand why people who aren't interested in "shipping daily" and are more interested in figuring things out are kinda annoyed by all the LLM noise
Imagine, if you will, wanting a narrow application to assist you personally in some side hobby, that is data driven but would take weeks or months to set up. Now imagine it's done in hours (and not even purely focused hours) and does everything you need it to do in an instantly verifiable and rewarding way. Imagine doing this on the regular for all of your various use cases. And then imagine not trying to sell or market that tool, as you dont care about that, you only care about solving problems and empowering yourself.
The vast majority of developers are like this, and as such, are silent about it. They aren't making production applications. They're making bespoke solutions and iterating upon them over time.
Somehow purists think there's something wrong with the process of that. I say there's something wrong with them, perhaps on the spectrum of OCD. With a tinge of arrogance or narcissism thinking their applications are all of the purest order and construction, testaments to architectural beauty and edge case handling. With a tinge of fear and cope as so much of their identity is caught up in being a smart coder rather than a smart engineer.
If you are a software developer and don't see the value in and heavily utilize agentic tools, you are a poor engineer. Not every coder has the mindset of an engineer, but a coder should be able to recognize and understand a powerful tool when they see one instead of letting their emotions and irrationality take the wheel
Nothing says engineer mindset like calling people "something wrong with them, perhaps on the spectrum of OCD. With a tinge of arrogance or narcissism" and "emotions and irrationality"
I think you missed the entire point of this discussion, but you also have a very gross view of the world!
> Somehow purists think there's something wrong with the process of that. I say there's something wrong with them, perhaps on the spectrum of OCD. Are you really saying it is a sign of serious mental illness because programmers want to enjoy their craft? You do realize our entire industry exists because of the passion of programmers years prior, right? I genuinely think you should be ashamed of what you wrote here.
I think you misunderstand. The entire industry exists because of the passion of engineers that program, not just programmers. And i specifically called out 'purists' for a reason, and yes I do believe OCD and perfectionism is a component. It would help if you read my comment critically rather than purely adversarily.
Well of course, if you enjoy writing software as a hobby then you value the experience of doing it, and you value the skill involved. You probably want a community of like minded individuals who all write code by hand, and where LLM use would be considered cheating.
That seems as natural as being against doping in sports or marking cards in poker.
None of this is anything against LLMs or AI assisted coding in general, even though the people involved probably have strong feelings about it given their preferences. It’s just the preferences of particular people and communities. Seems reasonable to me.
Yeah, it's just how you see programming in your perspective.
It's like lifting weight. Obviously, machines lift weight much better than human for a long time. But people still trains for strength or aesthetics. Even if machine can lift much larger weight and more efficiently, some people still care about how strong they are.
A machine can lift much more weight obviously is not the excuse for them to stop training.
Programming is just the weight lifting for your brain/mind. Even if LLMs can solve all coding problems perfectly, there will be people who still want to program all by hands, for training their mind, or simply enjoying the process. And there will be people care about how good you are at solving a problem with programming.
LLMs being able to code does not mean you no longer need the ability to code.
Cars do not mean you no longer need the ability to walk.
I don't exactly disagree, but I don't think your analogies work.
Machines can lift more weight than humans, but we all encounter situations where strength is useful and machines aren't available or wouldn't be suited to the task.
Cars are faster and easier than walking, but they require roads and are not a possible substitute for all of the walking we do.
But "if LLMs can solve all coding problems perfectly", and assuming that's sufficiently true of cheap/free models, there will be no practical need to code by hand; it'll be something we do purely for the joy of it and/or as brain training (but a niche form of brain training, probably practically inferior to other forms that target skills we still need).
> these communities don’t care if your code works at all, but instead care that you know why and how it works.
not literally true, and I doubt the author meant to say this. The communities do not only care about if your code works at all, but also that you know why and how.
It's sort of like in math classes where you need to show the work, not just the answer. Surely the answer being correct, while not being sufficient, is still necessary.
Right. The answer is necessary but not sufficient.
I was somewhat conflicted seeing the way the OSdev community responds to LLM usage.
I understand the conflict: the point of doing OSdev is that you're learning the low level hardware details and directly commanding the machine from the lowest level accessible. LLM usage is at odds with this.
But, on the other hand, the primary issue with OSdev on modern hardware is the complete lack of documentation for most hardware. A lone developer simply does not have enough time to reverse engineer from Linux or BSD drivers to achieve much hardware support. Plus, often you run into hardware quirks that are incredibly painful to debug and offer no real learning opportunity.
Recently I revived a ~8 year old modular kernel project I had written. It was at the point of being able to setup an x86_64 CPU, load some drivers and enter usermode. Using an ESP8266 board I setup a little debugging harness for a AtomicPi devboard I had laying around, and set it all up so Claude could control and upload code to it. It could read the serial port, ping the board over ethernet and power cycle it. From that point I let it work almost autonomously to implement a USB stack, finish the network stack, implement drivers etc.
It worked well, and for the most part I didn't care about the results. I don't have the time to commit to such a big project now and of course, the satisfaction of doing something myself (besides building the harness) was gone.
But a thing that stood out to me was, when trying to implement the ethernet driver, the chip in question did not have documentation, but it was known to be similar to another well documented chip. Claude was able to spend a few hours just trying different things, rebooting the board as needed until the differences were sorted out and it was able to ping the board. I feel this kind of application would be pretty relevant for even hobbyist communities like OSdev.
We all have limited time, and while LLM usage for everything certainly defeats the point, targeted usage for tedium that has very little value overall, should be acceptable, and I can't help but think lesser of devs I otherwise respect, when they insist on being dogmatic even when it doesn't contribute to the hobby.
> targeted usage for tedium that has very little value overall
The entire point of the article was the hobby is ascribing value to that tedium that the LLM is removing.
There are different kinds of tedium that LLMs automate, being a hobbyist doesn't mean you see value in all of them. Removing tedium that you personally find uninteresting or boring actually accelerates the hobby because now you're able to get past the tedious parts instead of giving up or delaying doing them. That does not mean that everyone finds those parts boring.
Even before LLMs came along a lot of hobbyists were using high level languages, linters, IDEs etc. instead of handcrafting assembly or machine language. Those tools increased the capability of hobbyists to build what they find interesting.
> Even before LLMs came along a lot of hobbyists were using high level languages, linters, IDEs etc. instead of handcrafting assembly or machine language. Those tools increased the capability of hobbyists to build what they find interesting
Yes and many programmers who came up in the 80s or early 90s writing C or equivalent decry these tools as ruining XYZ thing. LLMs are the same effect just on steroids
Isn’t that the point of a hobby? To enjoy the path and not the result? To learn something new.
I understand using LLMs as tools to help you with your hobby, but I think when “Claude spends hours debugging it” it probably is no longer a hobby, but an outsourced job.
That's why I was trying to make the distinction between stuff that actually pushes your understanding forward and teaches helpful information, and stuff that does not.
Isn't it more in the spirit of osdev to spend your effort implementing/working on the things that actually further your OS meaningfully? Discovering that a specific version of an obsolete undocumented chip has a quirk where one register doesn't reset in the expected way isn't really useful information and doesn't contribute to building new skills.
Maybe what I'm trying to say is that not everyone enjoys every piece of the hobby, and I don't think it's reasonable to exclude people who do not enjoy some of the pieces that aren't even core to the hobby.
Simply knowing the quirk doesn't improve your skills, but the process of finding it often does.
If Claude shortcuts that, then yeah, the problem might be solved, but you didn't skill up at all. Sometimes that's fine.
You can summarize this roughly by comparing art for its own sake versus art for a purpose. If I need software for a purpose, I don't care how it is crafted, only that it is crafted well. But if I care about the art of code done well, then obviously the means matter.
I find that a lot of "pro-AI" folks are very much doing software for the ends (myself included) and in the past, we were happy with the "means" primarily because there was no other way. The "anti-AI" folks are, as mentioned, mostly in the camp of caring about the means equally with the ends.
Lots of nuance here, but those buckets seem to generalize pretty well (so far) in my experience.
Very interesting stuff, and I believe the prejeduce against AI tools for use to code stems from the way they are currently package. The tools default out of the box are designed to throw away the most valuable process humans go through while learning, FAILURE!
If a human is never able to work through or process failures themselves they are bypassing one of the most important steps in the learning journey
More AI tools need to be made the opposite of this, humans should be able to achieve great learning USING AI.
The important message also remains, understand your code! It might just save you one day ;)
>even in the instances where there was earnest early engagement with LLMs in some of these niche communities, the well was quickly poisoned by a combination of a lack of a deep understanding by the LLM practitioners, and a vitriolic subset of those communities that view the LLM enterprise as a form of cheating
Yeah, the whole point is the street cred of doing something really hard.
Whether that chess repo is genuine or not genuine, legal or illegal, tasteful or tasteless, vibed or not vibed, stolen or copied, the signal to open source maintainers is clear: get off GitHub.
All I can hear is the nihilistic AI generated song at the end of the Malus talk. (https://fosdem.org/2026/schedule/event/SUVS7G-lets_end_open_...)
When you take away the cost people will treat the result worthless.
For example if you hiking in the mountains, reaching the summit and enjoying the views are valuable experiences to you because it had an associated coast, i.e the physical act of hiking up the mountain. If you take the ski lift to the top it's almost worthless and you have very little appreciation for the views.
I don't see software being any different to this.
Very apt comparison! This is how I feel too. In the sense that when I’m coding as a hobby, especially something like advent of code, using an LLM would defeat the purpose for me. A bit how I don’t double check my moves with an engine when playing chess online.
On the other hand though, if I’m being paid to build, say, the top station of a ski lift on a peak, I wouldn’t be surprised if the contractee would prefer I don’t try hiking up the building materials and used a helicopter instead.
I disagree with the conclusion. I think this is only half of the story, and I can say so as one of the people in question who is largely against LLM code (despite finding LLMs to be very useful tools). While I do care about genuine care in one's craft, I also am a free software advocate and fundamentalist, and I find non-trivial LLM code is too legally and ethically dangerous to be included in software that is distributed to other people (do whatever you want if it's just personal use).
"Aggressively Against LLM Usage" is the wrong way to think of it.
If my hobby is running, I would refuse to use a car to do my morning loop. Not because I have anything against automobiles, but because that's not my hobby.
Also, the issue linked to isn't really about LLMs. It's about someone being a dick in an online community and being called out for it. They just happen to be using LLMs to be a dick, but that's not the main issue.
I've found myself with this problem, as I've used LLMs to accelerate solving problems I've wanted to solve, only to find myself unable to share them back, because LLM usage has "polluted" my effort in the eyes of maintainers, even if the result has been something I've re-implemented by hand after learning from the agent-driven development what the issue is and how to reproduce it.
What's the solution, to just disengage entirely and not share back issues I find or features I'd like to see?
I get the negativity, because I enjoy hobby programming too, so I also enjoy the process sometimes. But then sometimes there's a problem and the fastest way to isolate the core of that issue is to use claude.
A long time ago, I had to take an ethics class to get my Computer Science degree. One of the things that came up in the papers about calculating the statistical value of a life was that people care more about being more prosperous than their peers than absolute material wealth.
I think about that a lot: its not about material wealth, but status, something that if someone has more of you have less of. A lot of people are invested in their status in various communities, and are afraid to lose it or have it become meaningless.
I don't know what to do about that; no matter how wealthy we all are people are still going to want status.
Well yes. That's why you see many westerners complaining about the cost of living even though all things told they have a good life.
I would modify this slightly though. I think there's an expectation of things getting better, and expecting to do better than your parents. I think that's where many of the modern issues around living standards come from. They aren't getting better (quickly enough).
look I enjoyed reading the status game too but it's really not rocket science that a fish will flounder when taken out of the water
Not quite the full truth:
In my free time; I'm iterating on programming ideas using coding agents.
For example a testing framework that separates test case definitions from instrumentation...
And the recent post on bento.page sent me down a path exploring different ideas around "X as standalone HTML+JS"...
And I don't care about shipping this; It's just for me.
I'm a hobbyist, what am I missing? LLM gave me a new spark. I never managed to get a SE job so some weekends I tinker around with LLM for small personal projects and it helps me learn more.
I think that’s the difference. Using LLMs to learn how to write interesting code, I hope, would be welcomed. I’m not confident that’s true though, reading some of these comments.
I don't think anyone is learning to code by using LLMs.
We learn a craft in two ways: by watching other people do something, and by doing it ourselves. I've learnt many crafts in my life, to the extent that I've produced useful or enjoyable output: programming, wood working, gardening, cooking, baking. I read loads of books, but nothing you read in a book makes much sense until you do it yourself. If you read a book on wood but never pick up a saw then you might as well be reading fiction.
Programming has always been a bit more difficult because we can't really watch someone else do it. Sure, you can watch someone use their tools, but programming really happens in the head, and you can't watch that. So you very much have to just do it yourself, and read.
When you use an LLM you're not watching it do anything, except output text at a very high words per minute. And you're certainly not doing it yourself. So how can you be learning? I actually don't think I've learnt a single thing since starting to use LLMs and it seems that way for colleagues too. They're producing code they never could have written before, but they don't understand it. They aren't learning.
I’ve learned tons since I started using LLMs. They’re an amazing tool for learning as long as you use them like that. Ask it questions about things you don’t understand. Verify the answers. Don’t have them do everything for you. Figure out where to push back on what they’re producing.
When I learned to code, learning from existing code was a big part of it. One difference now is you can produce custom code to learn from.
I’m starting to think the biggest skill we need to teach people in this new age is how to learn and how to research. If you already know how to do that well, LLMs are a fantastic addition to the toolbox. If you don’t, then you’re stuck with a tool that’s going to replace your skill instead of augment it.
Like I said, you can ask an LLM about wood all day but if you don't pick up a saw you won't learn to make anything. I do, of course, use LLMs for research and learning, but I'm specifically talking about learning to program. LLMs can't cut the wood for you, but they can write your program for you. If you can resist and just use them to learn then sure, but you've got to do something yourself.
Yes, hobbyists are less interested than businesses in automating their own hobby.
I'm against LLM usage in conversation, e.g. when discussing issues in open-source. It's like half-hearing a conversation, I'm missing the context that the other person has with their robot. Second, the person has usually at best half-understood the issue.
Provided the code does the job, it seems immaterial whether it was hand-written or LLM generated, provided it does the job and isn't straight-up plagiarism. The latter is admittedly a hard thing to judge.
Using LLMs for hobby programming is like using a computer to solve sudoku puzzles. The whole point is to do it yourself.
Personally, I find that when I start thinking "hmm, I wonder if I could write a program to automate this" is when I quit the pursuit. After figuring out the basic tricks to sudoku, it turned into busywork. (Or trying to micromanage StarCraft)
(Tagentially, if you want to automate programming, then I'm going to find myself a different field, because I have no interest in the results of sudoku puzzles. I especially have no interest in managing agents, which I see as junior-dev bullshitters, and would not enjoy managing such a team in real life)
> Using LLMs for hobby programming is like using a computer to solve sudoku puzzles. The whole point is to do it yourself.
(I absolutely relate to your experience with sudoku etc. Of course, writing a basic sudoku solver is also interesting.)
The thing is, software architecture and systems design can also get the hobby treatment. I've found that I'm happy to have a short discussion with ChatGPT and have it barf out a few hundred lines of a concept, then refactor and maintain it myself. Or come up with a few different pieces that can be written fairly easily, let the machine skip over most of that tedium, make sure they conform to the interfaces I had in mind, and then verify for myself that the system actually comes together the way I imagined.
Of course, it does help a lot that I can understand the code, and work with it line by line if necessary. But SICP teaches us that it's abstraction and composition all the way down. So I feel like I'm exercising the same skills while not letting myself get intimidated by the sheer amount of initial typing or the self-doubt (instinctive immediate review of the code, you know) that creeps in when I try.
> The whole point is to do it yourself.
I think if that were always true, calculators would be banned in schools.
Surely the "point" might be different things to different people depending on their exact situation/context.
Fantastic news! … for AI companies. These projects will continuously produce high quality, novel training data for free! Amazing deal.
Programming generates programmers, and programs are a side-effect.
One factor is that we struggled for a long time under proprietary tooling, and don't want to go back there.
Irrelevant. Open models are competent code generators. No you don't need millions of H100 hours to build them.
Not for everyone. Not all of us (especially unemployed) have the big bucks to buy the machines required.
But you can download gcc in a cheap laptop and create a theorem prover e.g.
Don't tell me how to do my hobby.
That's the core of it - hobbies are doing what makes you happy - if that's LLMs great, if not, great.
The interactive fiction community seems to have a subset of hardcore anti AI hobbyists. Partly at stolen artistic IP, but also the coding half of text adventures is viewed as art just as much as the story and writing.
I’ve selected a path where I think building tools with AI is fine. But I want people to play my games knowing I created all of it.
For me, domain discoveries and elegant solutions arise out of periods of deep befuddlement, and are often a saving grace that can avoid lots of code.
I'm lucky I guess that I've had enough time in this career to experience these moments, and enough experience to know which code I should write and which code really is rote and can be generated (most of the time). Hopefully it keeps the loss of enlightenment to a minimum.
I'm the OP of the cited thread, ask me anything :P
?
wtf the emoji didn't show up
what is this swindle
:EmotiTrumpet:
> In these communities (keep in mind that there is an implicit “not all…” throughout) the process of mastering a difficult field itself is the product, and something that runs is generally a nice-to-have.
Most hobby fandoms have a dedicated core group who think like this: The end result is only an artifact or social proof to them. The real product is their own prestige of having built it and attracted an audience. There's always an IRC channel or Discord somewhere where there's a social hierarchy of who should be respected for their work.
Outside of those inner circles, most people don't care. The average emulator user doesn't care about the details about who wrote their emulator or the history behind it all the forks. They want to play a game and the emulator core that gets the job done, because the emulator is the product to them. The mastery of the author is not a product to them.
This has happened in multiple technical hobby communities I've been in. With 3D printing I even remember communities where Prusa was looked down upon for making 3D printers too accessible. The argument was that now anyone could buy a 3D printer without really understanding the pain of mastering calibration, fine-tuning, and troubleshooting. That debate exploded to another level when Bambu came along and made 3D printing even more mainstream.
Now it's happening with programming communities. There are some where hand-written code is truly the actual product: Competitive programming, code golf, and obfuscated programming competitions come to mind. Using an LLM here is about as appropriate as submitting LLM generated art to an art contest, which is to say it's not.
There are a lot of other domains where the mastery of the craft isn't interesting outside of a small social circle. It's interesting that they mentioned emulators, because emu communities are famous for having a ton of strange drama that most emulator users don't care about. If you're inside of the small community and spend every day on their Discord, it might be the only thing you care about, but you don't realize that you're a minority because everyone else in the Discord thinks the same way.
OSDev is another interesting one to mention, because even Linus Torvalds had to make a big post telling the anti-LLM people to chill out. There seems to be a split between the purists who think the primary value is their own mastery of being able to write the code, and the pragmatists who are adapting to use new tools in conjunction with their knowledge. If you find yourself in one of the purist camps you might think everyone feels the same way, but outside of that, the world is moving on.
I usually feel turned off by communities where social status is the primary currency. There's always a hidden social hierarchy that feels threatened whenever it feels like the rules my start changing. They react by trying to force the social currency back to the system where they spent years investing their time. Nobody wants to feel like you're losing a foothold because the rules are changing. I got a strong feeling from this article that this author feels they've developed some level of respect in some forum and they dislike the idea of that investment being devalued by a changing landscape
> In traditional niche dev circles, respect is earned slowly through years of activity in their respective fora,
Blimey, it’s a mystery why they don’t let me take my jigsaw to the “community of hand saw enthusiasts”.
I can't relate. It's so nice for quickly going deep on a subject you're interested in. Who cares if you typed the code in plain English?
This site is an example why I am against dark mode.
> At the end of the day, these communities don’t care if your code works at all...
I think the key word to all this is "communities." If you can make something alone, just do it. If you enjoy gathering together to build a bigger thing, often the bigger thing is 1/3 an excuse to gather together. In that way, the LLM is like revealing the answer to a murder mystery game too early and expecting it still to be fun for everyone.
I'm particularly annoyed with the complaint in the linked thread about the agentic-developed chess engine.
> Moreover, you are giving absolutely nothing back to the community, which is central to open-source engine dev. You are only taking.
What would that user prefer? That the developer push a PR upstream...that presumably would be shot down and the user subject to further harrassment because it would be LLM derived at minimum? It seems like a catch-22 where the only winning move is not to play.
That thread also highlights the overlap between "people who are against LLM usage for coding" and "people who are very particular about AGPL."
> the only winning move is not to play
Presumably no one is forcing the maintainer at gunpoint to have an LLM slightly shake up other people's code, relicense it under an incompatible license and then claim that "every single line of code is written by Claude Code", so I don't see the issue with not playing if that's the only option.
If you can't make contributions that anyone finds interesting enough to include upstream and have no plans to publish your AGPL derived code under a compatible license, you can keep your changes to yourself. Nothing lost, nothing gained, license honored.
Is this novel? Every group that spent some large number of hours gaining some skill hates when that duration is not required for newcomers to accomplish what the skill allows. Not every individual of that group but the group as a whole. It’s just that devaluation of one’s skill capital sucks.
I happen to think the alternative world where no one’s skill capital can be devalued sucks more, but it’s hard to build a community around that because the same hours that build skill also build taste, and when you remove the former you remove the latter.
So you get a community with garbage in it. Haha!
> Every group that spent some large number of hours gaining some skill hates when that duration is not required for newcomers.
I'm not sure of the truthiness there wrt LLMs: using an LLM to code your hobby OS doesn't actually gain you the skills of the existing members of that niche. Maybe those existing members are just annoyed at posers, not at newcomers?
Sorry, I didn’t complete the thought well. Amended. I meant that the skill allows for some things and newcomers no longer require that investment to get the outcome.
Or it’s really just plain old gate keeping maybe.
We programmers had it good for a long time. Paid well, exercised the brain, creators.
I always thought that when I retired I’d do woodworking. But perhaps I’ll also code - to remember what it was like when I was young.
I think another reason to write code by hand is that you want to know what you get. If you have an idea, then you can implement it better than an LLM, just like you can implement it better than another person, because you are the only one who understands what you mean exactly. Details are always lost and misunderstood in communication; this is a weakness of LLMs.
Similarly, even if I don't write the code but I let someone else who I trust do it for me, there is value for me in that person not using LLMs, because only if they don't, I can know that the only spirit that lives in the code is of the person of my choice. It's like a social thing; I can know and trust a human, but I'm not sure I can do that with an LLM.
It's similar to how you wouldn't want to use LLMs in politics: the whole point of expressing your ideas is that they are your ideas and if you let someone (or something) else do it for you, details will be missed and misunderstood.
So this is one reason that I think hobby communities don't like LLMs; they care too much about what they build.
Depending what you are making; coding slows you down and think as you build stuff step by step. Its kind of like speaking vs writing. Generating code through LLM for me is like talking -- I love talking, but when I write something down, I'm more structured, think more about things etc etc.
People say, if you you write elaborate well spec'ed prompts, it will do the right thing. Coding in a way is doing that.
All of this is exactly why I started Slow Code. I like programming. Lots of people like programming. I don't want to stop doing that. So we meet up once a month to practice that.
https://slowcode.dev
to someone hobby programming using LLM seems like running a marathon using a bicycle, you cross the finish, but at the end is not the same as running only by your only forces.
Others look only at the results, so LLM are only a tool (actually without a clearly instruction manual i suspect).
But programming (and so many others works) is not only about making, but learning and developping skills. When an LLM is not available? i think differences will be evident.
And roughly speaking: to check the result of LLM you have to be more capable than the bot, otherwise, who is the boss? And to develop those skills, you have to do it yourself.
I enjoy programming as a hobby and have done it for many years. I also enjoy building stuff with LLMs. Why is one pure and sanctioned and the other one is bad? Why do I need other people to decide this for me?
if is good/bad for you is only to you, nobody else decide for you, but reasoning about general consequences adopting this thech is to everyone
> At the end of the day, these communities don’t care if your code works at all, but instead care that you know why and how it works.
This is not mutually exclusive with LLM usage. It's only prerequisite is that the human cares enough to learn.
I'm in an _extremely_ niche hobby programming community of Scrabble engine developers. There are single digits of us, but we're building the best engines and methodology out there and have all embraced LLMs to help us do it. There's a lot of experimentation that goes into it and LLMs certainly make it easier to iterate and try new ideas. So at least for our community, I don't think the points here hold.
That's awesome :) Do you have any links to your guys' work that I could check out?
This is a case of an SV-brained person stepping outside the bubble and realizing that LLMs are not the universal savior they think it is.
AI isn't well liked in a lot of places. I agree with the communities - make something yourself, don't ask a robot to do it for you. If you vibe slop something into existence, I'm never going to look at it. Something you made yourself though? Hell yeah I want to see it.
Another example is the art community - use AI/LLMs there, you'll be excommunicated.
I think, without knowing, that art often is similar to the hobby programming communities the thread leads with: the end result is interesting not by itself, but through the process that made it. We have feelings about art pieces because we feel with and through the artist. Can't do that with silicon. The Mona Lisa, even if the end result were exactly the same, would be dead boring had it been produced by the hand of a machine. My take
Totally disagree. I just want cool pictures hanging on my wall. If they’re totally unique and nobody else has them, even better. I can have a mounted picture of me on a horse like the famous Napoleon painting. I don’t care at all about the dude who painted it, I don’t even know his name.
The only exception is real photographs of some event, like I bought my friend a picture of the Falcon 9 taking off against the backdrop of the sun. Amazing photo.
I feel like obsessing over some artist I’ll never be friends with and the story behind their art is just another way to have a parasocial relationship, and the new tools allows us to have absolutely bespoke art in every single home. Things like this used to be considered strange and irregular before mass media. Now every person can make satisfying music. It’s incredible and amazing and should be celebrated because suddenly everyone can participate.
No one is stopping you doing that, but you probably aren't someone who should go dominate discussion in a hobbyist arts group.
> Now every person can make satisfying music.
That "make" is doing some heavy lifting there. Even handwaving all the ethical stuff with art and LLMs, Christopher Nolan doesn't "make" music when he contracts Hans Zimmer and gives him some guidance. Not I "make" a cabinet when I do the same with a woodworker.
And who's that satisfying for? For the one clicking the "generate" button? For the Sumo people getting VC money?
> I just want cool pictures hanging on my wall.
That's decoration, not art.
> Now every person can make satisfying music. It’s incredible and amazing and should be celebrated because suddenly everyone can participate.
No one should be celebrating the theft of music from artists. Everyone can participate in making music today just as they always have throughout human history - learn to sing or play an instrument.
> Now every person can make satisfying music.
Thanks to taxis, everyone can drive.
If it suits you, good for you. I said often, not all of it, of course.
Here's my thoughts as a 29 year old: of all the things I've ever made in my personal life, the vast majority have ended up abandoned in barely worked on states, and those I have put varying degrees of significant effort in have gone unnoticed. I've never really been able to "enter" a community in a way that felt meaningful to me, either, when I've tried.
I say this not as a complaint, but to explain my perspective on using LLMs for personal projects. Granted, I'm publishing none of them, and not trying to make anyone care about them, so it's not really a response to the article or general discourse here per se, but I thought it bears mentioning anyway.
I'd like to buy into the idea that "doing it yourself" is more meaningful or personally edifying, but I don't get that feeling. I want to get that feeling, but I don't. I've found my private waste-of-time projects using LLMs have been more gratifying than my private waste-of-time projects that don't, and at the end of the day, neither really substantively matter anyway.
Again, this isn't really related to communities per se, but it is an answer why some folks use LLMs for projects/"hobbies".
I get all the fuzz about only being able to learn if you do the work. But frankly, as a hobby programmer with a rather unusual workflow/toolset and a love for the command-line, it has never been so satifying to scratch a few itches I've always had but apparently never was motivated enough to really follow through. In the past 6 months, I've "written" at least 3 tools which I kind of always wanted, and which I now use on a regular basis. AND, I finally have a peer I can talk to about the code I'm trying to write. Something I basically never had, because all my FLOSS projects were small-very-nieche-one-man-shows. Its nice to be able to pretend to have a coding friend :-)
we don't meet enough people to worry about why or how productive someone is ... productivity is half mania, half obsession and 100 % in vain except in life critical fields where LLM speed doesn't matter because querying DBs is fast and making associations within _contexts that speak the same language_, meaning they use the same register, is trivial.
I use LLMs to build some nice customized and some basically standard tools so I can go from learning about code to learning Rust et al properly ... the next 2 years are scheduled for me. I want code and LLM supply chains as a proper hobby, partially because I find the overall "transformers" approach pretty bad since there is no implicit path towards intelligence and zero chance to build one on top explicitly. it's just a bunch of parrots but parrots are awesome and i'm glad I get to have so many.
Also: I really don't care what the communities think, though ... it's 2026 and all that code, math, science, engineering didn't build any special amount of good, "bad" ( vs ugly personalities ) or cool characters that we didn't have already and it didn't cause any "mutations" in schools of thought either, or drive movements where they were needed.
but I'm a bystander, which wasn't my choice, unfortunately, so y'all know much better what I'm talking about.
most players in the lower leagues won't dope but some do and take bribes anyway, what?
it's a cowardly new world, everybody, happy Esoteric Right Wing Hanukkah!
Model steam enthusiasts have had this for literally decades with a variety of premade parts available, angering those who learned to use a lathe. "Did you really make that" is huge in this.
Not my experience at all. I love using LLM's when hobby programming!
But then again I don't program as a hobby to impress others online. Maybe that's the difference?
The difference is probably that you are not that good at programming, and that the LLMs are better than you.
knitting communities also don't love factory machines too? I'm not sure how this could surprise anyone
Car enthusiasts generally dislike automatic transmissions, despite their ubiquity and clear advantages for the population as a whole.
Relatedly, there needs to be some kind of way to distinguish between types of LLM usage. Nothing more boring -- and often not even favorable or useful to a semi-hobbyist -- than just one-shotting something with Claude;
Vs. what I do. I've always done tons of little quirky automation type things with bash -- that do actually strongly become a part of my life.
Today? Yes, that -- but more and faster. I'm actually typing out way less code, but getting more of what I'd like done, and which couldn't be done without some bash (and now python*) knowledge.
Before AI, knew something about python but never typed e.g. more than 20 lines of it myself. I now am able to use it extensively regardless because, like with lots of languages, reading is easier than writing.
You're right, there are tons of different workflows you can do. One of the more innocent ones, regardless of you opinions on vibe coding, is just using the LLM to understand the code base better. Even if I'm vibe coding something, I will often start by having the LLM construct documentation for the current code flows. It's a document that is useful to me as a human because it mostly just gives me code pointers and tells me what it does as part of the larger whole. It's useful for the LLM since it saves a ton on time and tokens in the long run.
> Granted these communities have, in general, historically been characterized by feverish gatekeeping and painstakingly slow progress
This hits the nail on the head.
> In traditional niche dev circles, respect is earned slowly through years of activity in their respective fora
I think the problem goes deeper.
When a community becomes addicted to LLM, I feel uncomfortable interacting with it. It feels like mindless robots and drones are now active. But even aside from this, some people don't seem to understand why AI spam slopping real people with their AI spam, is awful. A few months ago, Michael Niedermayer spam AI slopped a proposal on their mailing list; I don't have the link right now (need to really bookmark all those things) but someone else asked him why he spams other people with unfinished proposals created by AI? I don't know whether Michael understands this as a problem or not, but I think many AI users don't understand the problem when they AI spam slop other people. I think it is a hostile action by those AI users. So I can totally understand communities not wanting that AI spam.
I've also noticed people who have horrible english skills, suddenly have perfect english skills. I suspect AI use, but then I wonder: is that person interacting with me via AI spam slop? I'd much rather deal with broken english if it is autenthic, than people faking it, even IF the produced english may be better than the broken english (but even that is often doubtful, so much AI spam slop is just wasting time and awful to read).
The whole AI trend seems to lead to really dumber people and fewer real interactions. That is sad, even more so as many people lost real social cues these days anyway.
Did you read about the Canadian politician who was making a speech and simply thoughtlessly read out some of the LLM boilerplate prefacing the text some chatbot had written for him?
https://www.cbc.ca/news/canada/new-brunswick/bill-oliver-ai-...
Are we going to end up a bunch of meatbag puppets of the machines? Makes me wonder.
The problem with politics is it’s not a meritocracy.
You find abdicating your speech to an AI to be disqualifying, but to people who’ve already abdicated everything they do to AI, it’s folksy and relatable.
See also: Why whittlers are against lathes.
It's because they fetishize arcane knowledge and LLMs make all their arcane knowledge easy. That's all it is. If I wanted to go make my own NES game from scratch 2 years ago it would have taken months of full time work, and I would have had to learn a lot of new things, and most likely would have joined and participated in that community. Now I can do it this afternoon. I'm not going to because it's pointless now, you can't show it off, you can't impress anyone with it, and the result is not just economically worthless, it's worthless in every way.
So of course they are upset by it, it's annoying to spend years building up some skill and then end up being less capable at applying that skill than a 12 year old with a claude code subscription who has never even seen a CRT. Their entire hobby community is going to fundamentally change and the people who are influential and knowledgeable now are effectively has-beens who will have to learn totally new skills to stay relevant and most likely won't have the energy, or desire, and lets be frank the aptitude to learn those skills. (I don't mean the skill of using an LLM, they can do that and almost certainly would be great at it, I mean skills like charisma and presentation and promotion.)
If the things you're thinking about are intrinsically so totally worthless outside of impressing people with them, and if that's all there is, please do not attempt to do those things, with or without an LLM.
Heck, if everything is worthless, please, please try not to do anything :)
If it's valuable why do they want to ban people from doing it with help from an LLM. You are just mad, and I understand why you are mad, but you're not right.
The thing I'm saying is worthless is making a game in 2026 for a game system that people literally couldn't give away for free in 1998. Due to nostalgia and the fact so many were thrown in the trash surviving systems now have some limited cash value. Newly made games do not have a nostalgia value. The reason people do it is to prove they can and to show other people. Anyone can now (no proof needed really, it's super easy) and nobody cares. It's like climbing Everest, and Claude is your Sherpa. Nobody is impressed by it anymore.
It's not just the hobby programming community. Professional software developers too.
At least against as it is now, where everyone with Claude suddenly becomes a software engineer.
Software engineering is more than pushing code to main/master.
It's requirements. Acceptance criteria. Design. Documentation written for humans. Reviews. Operations. Maintenance. Understanding the product. Understanding the system you're changing.
Right now, I struggle with PRs being opened by people who don't fully understand the code they're changing, replying to review comments with Claude generated answers pasted back like a meat proxy...
What I see at work is that we're pushing code faster than ever, while losing the big picture of our product, or the architecture of our own network, even faster.
We're also generating documentation at an incredible pace. The problem is that the people writing it often don't understand it either.
Code has become cheap.
Engineering hasn't.
Now more than ever it makes sense to have better programming languages that are more palatable to humans, because people may manually want to tweak LLM output, or other people who don't have access to the original prompts/context.
I don’t understand the world of people who have this degree of free time (time / energy) given how much everything costs just to be alive. Maybe they live in a different universe where programming for fun and first principals of “learning” matters. But for the other 99% of humans, costs have gone up 40% due to money printing and anything like this looks like a waste of time.
Hobbies are kind of by definition wastes of time. That's what makes them different than work.
If you only want to learn programming for work then you obviously have a different perspective than people writing chess engines from scratch for fun instead of using an existing one, right?
Surely most people in the US have hobbies at least in the minimal sense of downtime activities that they enjoy. It may not be golf or hobby programming, but even laying on the couch watching football is a "hobby" or sorts and takes up just as much time.
I realize there are people with young kids or 2-3 jobs that really have little to no downtime, but it seems that for most people time isn't really the issue - it's just are you more of a passive or active type of personality.
> I don’t understand the world of people who have this degree of free time
You dont understand that people have hobbies they enjoy doing? Because thats the entire point of the article.
There are (very coarsely) two types of people:
- people who like writing code for codes sake
- people who like writing code to build things
In the past, these folks have lived more or less in harmony unified under the common interest of applied maths, engineering and architecture that is necessary to write things that require a lot of vode. Aside for a few classic and very public spats.
Now that community is fractured because llms are increasingly making writing code irrelevant. If that weren't true, we wouldn't be having this convo at all.
I think it's perfectly reasonable to want to cultivate spaces that preserve that craft, but wish they wouldn't be so obnoxious about it. I imagine that a good handful of these folks in the "no llm ever camp" collectively rolled their eyes at vegans and animal rights advocates at some point, and are now feeling being on the unpopular side of progress.
Personally I think it's a cool prospect to have Trappist Monk-like communities producing hand crafted code. It will probably be really good and sturdy. We'll look at it in 10 years and say, "damn they really don't make it like that anymore!"
I can respect a developer who is still smitten with their craft to feel a bit anxious when the topic of LLM generated code comes up. I have a child who never grows tired of building their next game but first has to build the engine from scratch (needless to say they've built a lot of engines, not so many games). This use to frustrate me but then I realized the joy is in making a concept of a game for them. When I suggest Claude might get the job done faster I get a death glare followed by an eye roll. I get it, I've been in the industry long enough where coding is more of a means to an end and not the same novelty it once was.
More dangerous are communities who shun AI and treat every disclosure like a Scarlett letter. Generally the loudest opinions come from consumers (fan bois) and not creators. I think this will eventually drive new innovators away from these communities. These vocal naysayers have likely never experienced being a project manager, lead developer, and QA for an agent writing code. Instead they think a prompt magically creates software, but without wearing all these hats at once, you get something far worse than 'slop'. Call me crazy, but I think these multi-hat 'developers' using an agent and are dipping their toe into LLM development are going to raise the bar and leave us vim users in the dust.
Hobby programming is not a "community".
LLM-first development is the future. If you don't like it then, yes, you should quit.
Yeah you're right, it IS entirely fair to suddenly and quickly push the group that built your industry out of the industry. They should all just find new jobs, because that's very easy and simple. Thankfully we have bright minds like you to replace the talent lost, because with AI you can truly do anything your heart desires!
Obvious sarcasm aside, these LLMs have been trained (without consent!) on our hard work, on our passion. People have been writing code for years, and many have open-sourced their projects to share their knowledge, to better the world, to provide free solutions to problems. All that hard work, all that effort, has been stolen and used to create a machine that replaces us. And when we complain that quality goes down, that work is no longer enjoyable, that we don't actually see huge noticable productivity improvements across the board, bright minds like you tell us "lol just quit man fuck you". Do you not feel bad that you are actively destroying the livelihood of the very people that made your slop machine viable to build?
>without consent
Don't need it.
>Do you not feel bad
No, of course not. I feel good that technology is getting better and bitter activists are getting replaced.
As was said in that thread: "The age of the glorified individual is over - the collective mind is the future and the now. You will either be consumed by it, or trampled by it. It's time to ant up or fossil out."