I think the amount of turmoil around these deals is giving more weight to the possibility that we’re in a massive bubble thats quite divorced from any kind of fundamentals. Sooner or later the bubbles gonna burst.
Anthropic ARR went $1B -> $4B in the first half of this year. They're getting my $200 a month and it's easily the best money I spend. There's definitely something there.
"Sooner or later the bubble's gonna burst" and "There's definitely something there" aren't mutually exclusive - in fact they often go together.
It makes me perhaps a little sad to say that "I'm showing my age" by bringing up the .com boom/bust, but this feels exactly the same. The late 90s/early 00s were the dawn of the consumer Internet, and all of that tech vastly changed global society and brought you companies like Google and Amazon. It also brought you Pets.com, Webvan, and the bajillion other companies chronicled in "Fucked Company".
You mention Anthropic, which I think is in a good a position as any to be one of the winners. I'm much less convinced about tons of the others. Look at Cursor - they were a first moving leader, but I know tons of people (myself included) who have cancelled their subscription because there are now better options.
I probably mean it less as "I'm too old" and more of "Wow, time really flies".
To me, who started my career in the very late 90s, the .com boom doesn't really seem that long ago. But then I realize that there is about the same amount of time between now and the .com boom, and the .com boom and the Altair 8800, and I think "OK, that was a loooong time ago". It really is true what they say, that the perception of time speeds up the older you get, which is both a blessing and a curse.
Regarding AI, it's a bit fascinating to me to think that we really only had enough data to get generative AI to work in the very recent past, and nearly as soon as we had enough data, the tech appeared. In the past I would have probably guessed that the time between having enough data and the development of AI would have been a lot longer.
> It also brought you Pets.com, Webvan, and the bajillion other companies chronicled in "Fucked Company".
The irony with Webvan, they had the right idea about 15 years too early. Now we have InstaCart, DoorDash, etc. You really needed the mobile revolution circa 2010 for it to work.
Pets.com is essentially Chewy (successful pet focused online retailer)
So, neither of those ideas were really terrible in the same vain as say Juicera, or outright frauds like Theranos. Overvalued and ill-timed, sure
I genuinely don't understand what value Cursor itself brings. It's like a wrapper for some APIs, right? As far as I can tell there's like four actual AI firms in the world and everyone else is trying to whitelabel. It reminds me of the hosting industry in the early 2000s.
The value is the UX/DevX. Though, they are essentially just a fork of VS Code so it's hard to justify using instead of VS Code + Copilot or Continue which is almost the same UX now.
That's the problem with most "AI" products/companies that still isn't being answered. Why do people use your tool/service if you don't own the LLM which is most of the underlying "engine"? And further, how do you stay competitive when your LLM provider starts to scale RL with whatever prompting tricks you're doing, making your product obsolete?
Cursor isn't my preferred environment for development, but for me, it yields much better results than Copilot. Continue and Cline bugged out on me so badly, and so frequently, that I didn't find them worth using. YMMV
If you're interested, I recommend checking out Kiro from Amazon.
Every time I've tried Copilot or Cursor, it's happily gone off and written or rewritten code into a state it seemed very proud of, and which didn't even work, let alone solve the problem I put to it.
Meanwhile, Kiro:
1. Created a requirements document, with user stories and acceptance criteria, so that we could be on the same page about the goals
2. Once I signed off on that, it then created a design document, with code examples, error handling cases, and an architecture diagram, for me to review
3. After that looked good, it set about creating an itemized task list for each step of the implementation, broken down into specific tasks and sub-tasks and including which of the acceptance criteria from step 1 that task addressed
4. I could go through the document task by task, ask it to work on it, and then review the results
At one point, it noticed that the compiler had reported a minor issue with the code it had written, but correctly identified that resolving that issue would involve implementing something that was slated for a future task, so it opted to ignore the issue until the appropriate time.
For once, I found myself using an AI tool that handled the part of the job I hate the most, and am the worst at: planning, diagramming, and breaking down tasks. Even if it hadn't been able to write any working code at all, it already created something useful for me that I could have built off of, but it did end up writing something that worked great.
Note that I'm not really familiar with Rust (as most of the code will demonstrate), so it would probably have been far faster for an experienced Rust programmer to implement this. In my case, though, I just let it do its thing in the background and checked in occasionally to validate it was doing what I expected.
What you are describing is effectively my work flow with any LLM tooling, it's just not formalized by the tooling itself. Sounds interesting, I will take it for a spin.
Don't they have their own model for the inline completions? For me I find it really nice to preserve some brain energy by doing one repetitive change and just tab spamming it to get it done everywhere. I could get it done maybe just as fast with a macro in vim, but, Cursor lets me preserve the brain energy for something harder.
Meanwhile other "wrappers" e.g. in nvim or whatever, don't have this feature, they just have slightly better autocomplete than bare LSP.
> I genuinely don't understand what value Cursor itself brings. It's like a wrapper for some APIs, right? As far as I can tell there's like four actual AI firms in the world and everyone else is trying to whitelabel. It reminds me of the hosting industry in the early 2000s.
Yes, there's (maybe?) four, but they're at the very bottom of the value chain.
Things built on top of them will be higher up the value chain and (in theory anyway) command a larger margin, hence a VC rush into betting on which company actually makes it up the value chain.
I mean, the only successes we see now are with coding agents. Nothing else has made it up the value chain except coding agents. Everything else (such as art and literature generation) is still on the bottom rung of the value chain.
That, by definition alone, is where the smallest margins are!
> I genuinely don't understand what value Cursor itself brings. It's like a wrapper for some APIs, right?
By similar token Windows is mostly a wrapper around Intel and AMD and now Qualcomm CPUs. Cursor/Windsurf add a lot of useful functionality. So much so so that Microsoft GitHub Copilot is losing marketshare to these guys.
Are you trying to make a point or just being defensive for no reason? You called out something without having any experience or knowledge and then did the classic “it’s just a wrapper”.
Sorry, I thought the point was clear: Dropbox (file sync) is a feature, not a product. Cursor (AI in your IDE) likewise, is a feature, not a product.
I am old and I remember when you could make a lot of money offering "Get Your Business On The Information Superhighway" (HTML on Apache) and we're in that stage of LLMadness today, but I suspect it will not last.
Don’t be sorry it shows your true colors. The point stands that you continue to step around. Cursor and other tools like it are more than a trivial wrapper but of course you have never used them so you have no idea. At least give yourself some exposure before projecting.
Dropbox is still a $5+bn business. Cursor is still growing, will it work out, I don’t know but lots of folks are seeing the value in these tools and I suspect we have not hit peak yet with the current generation. I am not sure what a service business like a small biz website builder has to do with Cursor or other companies in adjacent spaces.
What "true colors"? I think I've been pretty consistent that I think Cursor is a commodity. You're surprisingly hostile and defensive about your preferred autocomplete plugin. You're right, I haven't used Cursor, but I've used similar tools like Copilot.
Your characterization of hosting as "a small biz website builder" is revealing. https://finance.yahoo.com/quote/GDDY/ is the one that made it and is now a $24B firm, but there were at least dozens of these companies floating around in the early 2000s.
You’ve been dismissive from the start without any real engagement with the product. I pointed out that you haven’t used Cursor, and rather than reflect on that, you’ve responded with sarcasm and condescension.
The only person being defensive here is you. My point was simple: tools like Cursor are more than just “wrappers.” Whether it becomes a massive business or not, revenue is growing, and clearly many users find enough value to justify the subscription. You don’t have to like it but writing it off without firsthand experience just weakens your argument.
At this point, you’re debating a product you haven’t tried, in a market you’re not tracking. Maybe sit this one out unless you have something constructive to say beyond “it’s just a wrapper”.
I think you underestimate the difficulty in getting the tool chain running efficiently in the IDE. It's a significant moat and I suspect their spend is too attractive to cut them off from an API especially when most of the model providers are not exactly competing fully in this space yet or at least not with the same enthusiasm.
Interesting to see multiple posts here saying this. Pretty clearly it isn't true. The IDE is owned by Microsoft. The model is owned by Anthropic or Google or whoever. A business can't be made from a thin sandwich filling between the two.
Hey guys we have a guy here stating pretty clearly it isn’t true. He clearly is the authority on the topic because he said so.
The shell of the IDE is open source. It’s true there is some risk on the supply of models and compute but again none of those, except MSFT which does not even own any of the SOTA models, have any direct competition. OpenAI has codex but it’s half baked and being bundled in ChatGPT. It is in nobodies interest to cut off Cursor as at this point they are a fairly sustained and large customer. The risk exists but feels pretty far fetched until someone is actively competing or Cursor gets bought out by a OpenAI.
Again, what proof do you have that there is zero complexity or most being driven by the sandwich filling. Most of OpenAIs valuation is being driven by the wrapper ChatGPT not API usage. I have written a number of integrations with LLM APIs and while some of it just works, there is a lot of nuance to doing it effectively and efficiently at scale. If it was so simple why would we not see many other active competitors in this space with massive MAUs?
> and the underlying provider can easily cut them out
what? Do you think providers (or their other customers) don’t care about the business implications of a decision like this? All so that cursor can bring their significant customer base to a nearly-indistinguishable competitor?
Anthropic is actually a good point to focus on since Claude is very good proof that it's not about the scaling. We are not quite there yet but we are "programming" through how we shape and filter the input data for training it seems. With time, we'll understand the methods to better represent.
Current situation doesn't sound too good for "scaling hypothesis" itself.
I use github copilot chat right now. First I use ask mode to ask it a question about the state of the codebase outlining my current understanding of the condition of the code. "I'm trying to x, I think the code currently does y." I include a few source files that I am talking about. I correct any misconceptions about the plan the llm may have and suggest stylistic changes to the code. Then once the plan seems correct, I switch to agent mode and ask it to implement the change on the codebase.
Then I'll look through the changes and decide if it is correct. Sometimes can just run the code to decide if it is correct. Any compilation errors are pasted right back in to the chat in agent mode.
Once the feature is done, commit the changes. Repeat for features.
I also do the same. I am on the 200$ maxpro plan. I often let the plan go to pretty fine level of detail, e.g. describe exactly what test conditions to check, what exact code conditions to follow.
Do you write this to a separate plan file? I find myself doing this a lot since after compaction Claude starts to have code drift.
Do you also get it to add to it's to-do list?
I also find that having the o3 model review the plan helps catch gaps. Do you do the same?
Yes, it can't change between edit and ask/agent without losing context but ask <-> agent is no problem. You can also change to your custom chat modes https://code.visualstudio.com/docs/copilot/chat/chat-modes without losing context. At least that's what I just did in VSCode Insiders.
Also, I am using tons of markdown documents for planning, results, research.... This makes it easy to get new agent sessions or yourself up to context.
Yes. I think it used to be separate tabs, but now chat/agent mode is just a toggle. After discussing a concept, you can just switch to agent mode and tell it to "implement the discussed plan."
GitHub Copilot models are intentionally restricted, which unfortunately makes them less capable.
I'm not the original poster, but regarding workflow, I've found it works better to let the LLM create one instead of imposing my own. My current approach is to have 10 instances generate 10 different plans, then I average them out.
This was my answer as well. And I think it just highlights all the serious dangers for the "API wrapper companies" compared to the foundation model companies.
User experience is definitely worth something, and I think Cursor had the first great code integration, but then there is very little stopping the foundation model companies from coming in and deciding they want to cut out the middleman if so desired.
Amp Code is also very good, they released about 2 weeks ago their Oracle feature which leverages o3 to do reviews (https://ampcode.com/news/oracle). Amp leans the closer to Claude Code more than other solutions I’ve seen so far, the team there is really leaning into the agentic approach.
I watch the changes on Kilo Code as well (https://github.com/Kilo-Org/kilocode). Their goal is to merge the best from Cline & Roo Code then sprinkle their own improvements on top.
My problem with Claude code versus Cursor is that with Cursor I could "shop around" the same context with different foundational model providers, often finding bugs this way or or getting insights.
Sometimes one model would get stuck in their thinking and submitting the same question to a different model would resolve the problem
It allows you to have CC shoot out requests to o3, 2.5 pro and more. I was previously bouncing around between different windows to achieve the same thing. With this I can pretty much live in CC with just an editor open to inspect / manually edit files.
Cline is absolutely fantastic when you combine it with Sonnet 4. Always use plan mode first and always have it write tests first (have it do TDD). It changed me from a skeptic to a believer and now I use it full time.
I use Roo Code (Cline fork) and spend roughly $15-30/mo by subscribing to Github Copilot Pro for $10/mo for unlimited use of GPT-4.1 via the VS Code LM API, and a handful of premium credits a month (I use Gemini 2.5 Pro for the most part).
Once I max out the premium credits I pay-as-you-go for Gemini 2.5 Pro via OpenRouter, but always try to one shot with GPT 4.1 first for regular tasks, or if I am certain it's asking too much, use 2.5 Pro to create a Plan.md and then switch to 4.1 to implement it which works 90% of the time for me (web dev, nothing too demanding).
With the different configurable modes Roo Code adds to Cline I've set up the model defaults so it's zero effort switching between them, and have been playing around with custom rules so Roo could best guess whether it should one shot with 4.1 or create a plan with 2.5 Pro first but haven't nailed it down yet.
Looking at Cline, wondering what the real selling points for Roo Code are. Any chance you can say what exactly made you go with Roo Code instead of Cline?
Cline has two modes (Plan and Act) which work pretty well but Roo Code has 5 modes by default. (Code, Ask, Architect, Orchestrator, Debug) and it's designed so that users can add custom modes. e.g. I added a Code (simple) mode with instructions about the scale/complexity of tasks it can handle or decide to pass it to Code for a better model. I also changed the Architect mode to evaluate whether to redirect the user to Code or Code (simple) after generating a plan.
Roo Code just has a lot more config exposed to the user which I really appreciate. When I was using Cline I would run into minor irritating quirks that I wished I can change but couldn't vs. Roo where the odds are pretty good there are some knobs you could turn to modify that part of your workflow.
We still use Windsurf at work, but I've cancelled my personal subscription as (frankly) the code which Claude Sonnet 3.7 (thinking) generates is just not good enough, and much of the time its analysis of issues is also not correct. :(
That being said _sometimes_ its analysis is actually correct, so it's not a total miss. Just not something I'm willing to pay for when Ollama and free models exist.
Feels nothing like the same. The .com bubble was largely companies with no business, unchanged revenue but still having massive swings in price in private and public markets.
Cursor has a $500mm ARR your anecdote might be meaningful in the medium turn but so far growth as not slowed down.
> The .com bubble was largely companies with no business
Ah, yes, companies like Amazon.com, eBay, PayPal, Expedia, and Google. Never heard of those losers again. Not to mention those crazy kids at Kozmo foolishly thinking that people would want to have stuff delivered same-day.
The two lessons you should learn from the .com bubble are that the right idea won’t save you from bad execution, and that boom markets–especially when investors are hungry for big returns–can stay inflated longer than you think. You can be early to market, have a big share, and still end up like Netscape because Microsoft decided to take the money from under the couch cushions and destroy your revenue stream. That seems especially relevant for AI as long as model costs are high and nobody has a moat: even if you’re right on the market, if someone else can train users to expect subsidized low prices long enough you’ll run out of runway.
That’s what people predicted bit, for example, on Amazon’s case it was less than 3 years because they just kept posting solid numbers. The thing which all of those companies have in common is that they stood out from the Pets.com types in having profitable revenue - they didn’t need a miracle to be profitable, only for customers to keep buying.
You’re right that many .com companies lacked fundamentals but you’re cherry-picking survivors. For every Amazon, there were dozens of Pets.coms. The current AI wave does feel different in terms of revenue traction (e.g., Cursor’s $500M ARR), but the broader lesson still applies: hype cycles don’t discriminate between good and bad execution in the short term.
Cursor’s growth is impressive, but sustained dominance isn’t guaranteed. Distribution, margins, and defensibility still matter and we haven’t seen how durable any of that is once incentives tighten and infra costs stop being subsidized.
My point in listing survivors was simply to make the point that while there were plenty of doomed businesses, there were also many giants which were big at the time and could be told apart by looking at their fundamentals — they had real people paying them money for tangible things at a price which could be profitable. Amazon famously reported low numbers due to reinvestment but they were profitable in most business segments a few years after entering, which was quite different from the “lose money on every sale, make it up on volume” plays many dotcoms made.
How does that refute the statement you quoted? I said the vast majority of companies during the bubble had no business, were run on hype dollars and had insane P/E ratios. That supports a handful of companies making it through the bloodbath, but also a cherry picked examples that neither refutes my claim or supports yours.
You said “largely” and I think that’s painting with too broad a brush. The dotcom world included a bunch of companies which are still around (or were acquired later after surviving the collapse), and it wasn’t hard to tell who those were even at the time. There was a lot of lazy boosterism and criticism painting the whole field as the same, and that was a disservice to readers who could’ve used a more thoughtful triage approach. That’s especially the case for companies like Kozmo which actually had a popular idea and had the potential to be profitable (they were in most urban markets) but made the mistake of expanding too quickly or taking on more debt than they could service.
You’re sidestepping the core point. Of course some companies had fundamentals, even Kozmo had product market fit in a narrow sense. But the broader ecosystem was bloated with capital chasing flimsy ideas, and most dot-coms had no viable path to profit. That’s not “too broad a brush”, it’s backed by the collapse itself.
Kozmo is a great case study: decent demand, terrible unit economics, and zero pricing power. They didn’t just scale too fast, they scaled a structurally unprofitable model. There was no markup, thin margins, and they held inventory without enough throughput.
Many of these companies may fail but it’s a much different environment and the path to profitability is moving a lot quicker.
> The .com bubble was largely companies with no business, unchanged revenue but still having massive swings in price in private and public markets.
There also were companies like Sun and Cisco who had real, roaring business and lots of revenue that depended on loose start-up purse-strings, and VC exuberance...
Sun and Cisco both survived the .com bust, but were never the same, nor did theu ever reach their high-water marks again. They were shovel-sellers, much like Amazon and Nvidia in 2025.
Or yahoo- they were the premier sellers of ad space online (like google today) and made a lot of money from over-funded tech companies overpaying for online advertising during the boom years.
For sure but unlike then we are in a very different buying environment. Investors are more discerning even though folks here would like tot think differently. Cisco had something at peak like a 179x pe. That is a vastly different world than what we see Nvidia at today. I am not saying it cannot fail or collapse but to say this feels like the .com bubble is wrong.
>Feels nothing like the same. The .com bubble was largely companies with no business, unchanged revenue but still having massive swings in price in private and public markets.
I'm an attorney that got pitched the leading legal AI service and it was nothing but junk... so I'm not sure why you think that's different from what's going on right now.
Never heard Vincent to be leading in the legal space but I know vlex is more popular outside of the US so perhaps that’s why.
I am not sure why you would think your single anecdote is defensible or evidence to prove much. My perspective is valuations that are going on right now don’t have multiples that are that wild especially if we aren’t compare it to the com bubble.
>I am not sure why you would think your single anecdote is defensible or evidence to prove much.
Evidence? Prove? What are you talking about. This is just a discussion between people, not some courtroom melodrama you are making it out to be.
>My perspective is valuations that are going on right now don’t have multiples that are that wild especially if we aren’t compare it to the com bubble.
I responded in the exact same format as yourself and adding my additional thoughts. If that counts a being rude then were you being rude? I don’t believe in the US market Vincent is considered the best tool. That said I don’t believe a lot of the tools in the legal space that try to do everything are that powerful and the ones on my lens that are going well have focused on specific domains and problems. Even Harvey tried to do too much from the start.
As for valuations, when looking at current VC multiples and equity markets, I don’t see the same bubble from a qualitative perspective. Absolutely there is over hype coming from CEOs in public markets but there is a lot of value being driven. I don’t believe the giants are going to do well, maybe the infrastructure plays will but I think we will see a carve out of a new generation of companies driving the change. Unlike ‘99, I am seeing a lot more startups and products with closer to the ground roadmaps to profitability. In 99 so many were running off of hopes and dreams.
If you would actually like to converse I would love to see your perspective but if all you can be is mad please please don’t respond. Nobody is having a courtroom drama other than what’s playing out in your head.
Was the leading legal AI service Harvey.ai by any chance? I feel like big VC money goes to solving legal analysis, but I'm seeing a lot of wins with document drafting/templating.
Briefpoint.ai, casely.ai, eve.legal etc. I work with an attorney who trained his paralegals to use chatgpt + some of these drafting tools, says it's significantly faster than what they could've done previously.
> I feel like big VC money goes to solving legal analysis, but I'm seeing a lot of wins with document drafting/templating.
What do you mean "wins?" Like motions won with AI drafted papers? I'm skeptical.
>I work with an attorney who trained his paralegals to use chatgpt + some of these drafting tools, says it's significantly faster than what they could've done previously.
I'd be concerned about malpractice, personally. The case reviews I've seen from Vincent (which is ChatGPT + the entire federal docket) are shocking in how facially wrong they can be. It's one thing for an attorney to use ChatGPT when they do know the law and issues (hasn't seemed to help the various different partners getting sanctioned for filing AI drafted briefs) but to leave the filtering to a paralegal? That's insane, imo.
Not really much a of stuck bubble this time, though. Besides Nvidia and a handful of other HW companies, at least. Almost all of the very high valuations are for private companies and usually the amount of actual money involved in is relatively low.
I’ve always dwelled over $5 a month subscriptions for iPhone apps due to subscription fatigue. I find myself signing up for $200 AI subscriptions without a moments hesitation.
What do you do with $200/mo subscription to Anthropic? I’d consider myself a power user and I’ve never come close to a rate limit on the $20 subscription.
If you discusses a plan with CC well upfront, covering all integration points where things might go off rail, perhaps checkpoint the plan in a file then start a fresh CC session for coding, then CC is usually going to one shot a 2k-LoC feature uninterrupted, which is very token efficient.
If the plan is not crystal clear, people end up arguing with CC over this and that. Token usage will be bad.
I personally find gemini 2.5 pro and o4.1 mini to handle complexity better than claude code. i was a power user of claude code for a couple months but its bias to action repeatedly led me down the wrong path. what am i missing?
Are there available numbers to support this? Software engineering in the U.S. is well-compensated. $200/mo is a small amount to pay if it makes a big difference in productivity.
My day job in talks to do that. I'm partly responsible for that decision, and i'm using my personal $200/m plan to test the idea.
My assessment so far is that it is well worth it, but only if you're invested in using the tool correctly. It can cause as much harm as it can increase productivity and i'm quite fearful of how we'll handle this at day-job.
I also think it's worth saying that imo, this is a very different fear than what drives "butts in seats" arguments. Ie i'm not worried that $Company will not get their value out of the Engineer and instead the bot will do the work for them. I'm concerned that Engineer will use the tool poorly and cause more work for reviewers having to deal with high LOC.
Reviews are difficult and "AI" provides a quick path to slop. I've found my $200 well worth it, but the #1 difficulty i've had is not getting features to work, but in getting the output to be scalable and maintainable code.
Sidenote, one of the things i've found most productive is deterministic tooling wrapping the LLM. Eg robust linters like Rust Clippy set to automatically run after Claude Code (via hooks) helps bend the LLM away from many bad patterns. It's far from perfect of course, but it's the thing i think we need most atm. Determinism around the spaghetti-chaos-monkeys.
Yes, but that doesn't mean they aren't finding real value
The challenge with the bubble/not bubble framing is the question of long term value.
If the labs stopped spending money today, they would recoup their costs. Quickly.
There are possible risks (could prices go to zero because of a loss leader?), but I think anthropic and OpenAI are both sufficiently differentiated that they would be profitable/extremely successful companies by all accounts if they stopped spending today.
So the question is: at what point does any of this stop being true?
> I think anthropic and OpenAI are both sufficiently differentiated that they would be profitable/extremely successful companies by all accounts if they stopped spending today.
Maybe. But that would probably be temporary. The market is sufficiently dynamic that any advantages they have right now, probably isn't stable defensible longer term. Hence the need to keep spending. But what do I know? I'm not a VC.
Have we seen any examples of any of these companies turning a profit yet even at $200+/mo? My understanding is that most, if not all, are still deeply in the red. Please feel free to correct me (not sarcastic - being genuine).
If that is the case at some point the music is going to stop and they will either perish or they will have to crank up their subscription costs.
I am absolutely benefitting from them subsidizing my usage to give me Claude Code at $200/month. However, even if they 10x the price its still going to be worth it for me personally.
I totally get that but that’s not really what I asked/am driving at. Though I certainly question how many people are willing to spend $2k/mo on this. I think it’s pretty hard for most folks to justify basically a mortgage for an AI tool.
My napkin math is that I can now accomplish 10x more in a day than I could even one year ago, which means I don't need to hire nearly as many engineers, and I still come out ahead.
I use claude code exclusively for the initial version of all new features, then I review and iterate. With the Max plan I can have many of these loops going concurrently in git worktrees. I even built a little script to make the workflow better: http://github.com/jarredkenny/cf
Again I understand and I don’t doubt you’re getting insane value out of it but if they believed people would spend $2000 a month for it they would be charging $2000 a month, not 1/10th of that, which is undoubtedly not generating a profit.
As I said above, I don’t think a single AI company is remotely in the black yet. They are driven by speculation and investment and they need to figure out real quick how they’re going to survive when that money dries up. People are not going to fork out 24k a year for these tools. I don’t think they’ll spend even $10k. People scoff at paying $70+ for internet, a thing we all use basically all the time.
I have found it rather odd that they have targeted individual consumers for the most part. These all seem like enterprise solutions that need to charge large sums and target large companies tbh. My guess is a lot of them think it will get cheaper and easier to provide the same level of service and that they won’t have to make such dramatic increases in their pricing. Time will tell, but I’m skeptical
It looks like their revenue has indeed increased dramatically this year but I can’t find anything saying they’re profitable, which I assume they’d be loudly proclaiming if it had happened. That being said looking at the charts in some of these articles it looks like they might pull it off! I need to look more closely at their pricing model, I wonder what they’re doing differently
I guess my genuine question in response is can you tell investors "Please give us billions of dollars - we never plan on being profitable, just endlessly growing and raising money from outside sources"? Unless the goal is to be sold off eventually that seems a bit like a hard sell.
> My napkin math is that I can now accomplish 10x more in a day than I could even one year ago, which means I don't need to hire nearly as many engineers, and I still come out ahead.
The only answer that matters is the one to the question "how much more are you making per month from your $200/m spend?"
I'm curious, how are you accounting this? Does the productivity improvement from Claude's product let you get your work done faster, which buys you more free time? Does it earn you additional income, presumably to the tune of somewhere north of $2k/month?
Anecdotally, I can take on and complete the side projects I've always wanted to do but didn't due to the large amounts of yak shaving or unfamiliarity with parts of the stack. It's the difference between "hey wouldn't it be cool to have a Monte Carlo simulator for retirement planning with multidimensional search for the safe withdrawal rate depending on savings rate, age of retirement, and other assumptions" and doing it in an afternoon with some prompts.
For curiosity, how complex are these side projects? My experience is that Claude Code can absolutely nail simple apps. But as the complexity increases it seems to lose its ability to work through things without having to burn tokens on constantly reminding it of the patterns it needs to follow. At the very least it diminishes the enjoyment of it.
It varies, but they're not necessarily very complex projects. The most complex project that I'm still working on is a Java swing UI to run multiple instances of Claude code in parallel with different chat histories and the ability to have them make progress in the background.
If you need to repeatedly remind it to do something though, you can store it in claude.md so that it is part of every chat. For example, in mine I have asked it to not invoke git commit but to review the git commit message with me before committing, since I usually need to change it.
There may be a maximum amount of complexity it can handle. I haven't reached that limit yet, but I can see how it could exist.
Simple apps are the majority of use-cases though - to me this feels like what programming/using a computer should have been all along: if I want to do something I’m curious about I just try with Claude whereas in the past I’d mostly be too lazy/tired to program after hours in my free time (even though my programming ability exceed Claude’s).
Well that's why I'm curious. I've been reading a lot of people talking about how the Max plan has 100x their productivity and they're getting a ton of value out of Claude Code. I too have had moments where Claude Code did amazing things for me. But I find myself in a bit of a valley of despair at the moment as I'm trying to force it to do things I'm finding out that it's not good at.
There are definitely things it can't do, and things it hilariously gets wrong.
I've found though that if you can steer it in the right direction it usually works out okay. It's not particularly good at design, but it's good at writing code, so one thing you can do is say write classes and some empty methods with // Todo Claude: implement, then ask it to implement the methods with Todo Claude in file foo. So this way you get the structure that you want, but without having to implement all the details.
I work at an Amazon subsidiary so I kinda have unlimited gpu budgets. I agree with siblings, I'm working on 5 side projects I have wanted to do as a framework lead for 7 years. I do them in my meetings. None of them are taking production traffic from customers, they're all nice to haves for developers. These tools have dropped the costs of building these tools massively. It's yet to be seen if they'll also make maintaining them the same, or spinning back up on them. But given AI built several of them in a few hours I'm less worried about that cost than I was a year ago (and not building them).
The point is that if a minority is prepared to pay $200 per month, then what is the majority prepared to pay? I also don’t think this is such an extreme priority, I also know multiple people in real life with these kinds of selections.
Yeah, cause we want to be in control of software, understandably. It's hard to charge for software users have full control of - except for donations. That's #1 reason for me to not use any gen AI at the moment - I'm keeping an eye on when (if) open-weight models become useful on consumer hardware though.
So $415m revenue per month, annualized $5 billion / yr. Let's say we use a revenue multiple of 4x, that means OpenAI should be valued at $20 billion USD just based on this. Then one obviously has several other factors, given the nature of OpenAI and future potential. Maybe 10x more.
Which puts the current valuations I've heard pretty much in the right ballpark. Crazy, but it could make sense.
I'm easily getting $10K/month of value from my Anthropic subscription. (Rough estimate of how much I would have paid someone else to create the things I've (co)created with Claude Code so far.) If this is a bubble, I just hope I can finish all the projects I want to finish before it pops (or before they raise their prices to $9K/month because they read this comment.)
> I'm easily getting $10K/month of value from my Anthropic subscription.
Are those things created by Claude actually making you that much in real money every month? Because the amount of money it would cost to pay someone to create something, and the value that something brings to you once it's made are largely unrelated.
They are tools I want/need for my business (like creating software libraries for various things). My $10K number is how much I would have paid a contractor in the past to code it for me.
> I'm easily getting $10K/month of value from my Anthropic subscription. (Rough estimate of how much I would have paid someone else to create the things I've (co)created with Claude Code so far.)
I know it's hard to place a value on how much a utility saves a business, but honestly this math is like the piracy math and we didn't buy it back then either.
Some teenager downloading 20k songs does not mean that they saved $20k[1], nor does it mean that the record labels lost $20k.
In your case, the relevant question is "how much did your revenue increase by after you started 10x your utility code?"
[1] Assuming the songs are sold on the market for $1 each.
OP wanted a thing. in the past, they've been OK paying $10k for similar things. now they're paying $200/month + a bunch of their time wrangling it and they're also OK with that.
seems reasonable to consider that "$10k of value" in very rough terms which is of course how all value is measured.
> OP wanted a thing. in the past, they've been OK paying $10k for similar things.
Okay, then their costs should have come down similarly, no? OP said they were a business and that these weren't luxury hobby things but business needs. In which case, it must reflect on the bottom line.
I operate as a business myself (self-employed), and I can generally correlate purchases with the bottom line almost immediately for some things (Jetbrains, VPSes for self-hosted git, etc) and correlate it with other things in the near future (certifications, conferences, etc).
The idea that "here is something I recently started paying a non-trivial amount for but it does not reflect on the bottom line" is a new and alien concept to me.
When people made studio ghibli versions of themselves for free, were they creating hundreds of dollars worth of value since that's how much it would've cost a freelancer to commission such a picture? I would say rather the value of the pictures themselves became very cheap.
I have a feeling you'd have had better results if you actually paid $10k/month to a good dev or three in a LCoL geo.
You can actually hire a few excellent devs for very little money. You just can't hire 20k of them and convince them to move to a certain coastal peninsula with high rent and $20 shawarmas, for very little money each.
Your first link is (in my opinion) highly biased in the samples they choose, they hired maintainers from open-source repos (people with multi years of experience, on their specific repo).
So indeed, IF you are in that case: Many years on the same project with multiple years experience then it is not usefull, otherwise it might be.
This means it might be usefull for junior and for experienced devs who are switching projects.
It is a tool like any other, indeed if you have a workflow that you optimized through years of usage it won't help.
Exactly. I think the study is a good reminder that we really have to be careful about the productivity gains attributed to AI. Main takeaway imo, despite limitations from the study, is AI is not a panacea, it can increase productivity, but only if used 'well' and with the good workflows in place, and in the right context.
You are welcomed to your point of view, but for me while one agent is finding an obscure bug, I have another agent optimising or refactoring, while I am working on something else. Its hard to believe I am deluded in thinking I am spending more time on a task.
I think the research does highlight that training is important. I don't throws devs agents and expect them to be productive.
1. If you're maxing out your subscription, they're burning money on you.
2. They don't have a moat. DeepSeek and Kimi are already good enough to destroy any high margins they're hoping to generate from compute.
Just because something is highly useful doesn't mean it's highly profitable. Water is essential to life, but it's dirt cheap in most of the world. Same goes for food.
I agree. I was experimenting with tool use with Kimi K2 APIs yesterday - very effective, and so incredibly inexpensive. I am retired, now doing independent research, so my requirements are very different than most people here who are still in the job market or growing their own business.
I find a combination of local Ollama models, with very inexpensive APis like Moonshot’s Kimi with occasional Gemini 2.5 Pro use, and occasionally using gemini-cli provides extraordinary value. Am I missing out by not using one or more $200-$300 a month subscriptions? Probably but I don’t care.
growing ARR is easy when you are selling dollars for cents. people hyping ARR as an meaningful investment indicator are a dead giveaway that we are in fact in a bubble.
For sure, but then again - Nvidia $4T?! I can't shake the feeling though that with Nvidia we're looking at another Sun type of situation from _the bubble_. Remember the dot in dot com?
NVidia is being propped up by the US government. Huawei’s new chips are lower tech but would probably hit a good ‘practical sweet spot’ for AI data centers in many countries around the world but our current administration is threatening economic violence against any countries who choose to use more cost effective Huawei AI chips.
I don’t want to descend into talking politics, but I want to say that geopolitics, the rising geopolitical ‘south’, etc., is fascinating stuff - much more interesting and entertaining than anything fictional on Netflicks or HBO!
Economic bubbles _usually_ aren't based on things which are entirely worthless; there's generally _something_ there (just not enough something to sustain the valuations). There are exceptions (NFTs, arguably cryptocurrency as a whole, and of course tulips), but those _are_ the exceptions.
The big question is to what extent they hit a plateau and are commoditized. What happens when there is a fully open stack that gets Claude Code level results but at a fraction of the cost? Not saying that will happen, but that seems to be the scenario for a bubble bursting.
I decide to try out the agent built into VS Code. It basically matches most of these fly by night "agent" ides which are mostly just VS Code forks anyway.
But it's weird. Because Microsoft can use Anthropic's API, funnel them revenue and take a loss on Copilot.
We're all getting this stuff heavily subsidized by either VC money or big corp money.
Microsoft can eat billions in losses on this if they become *the* provider of choice.
This stuff isn't perfect, but this is the worst it'll ever be. In 2 years it'll be able to replace many of us.
I pay $0 and that's already enough for me. Genuinely, what are you getting for your $200? I cannot fathom paying that much for what seems like I get basically for free anyway.
So how much is it worth? what if it balloons to 5x the amount? what if "they" just decide this is too powerful so only corporate accounts worth $xx,000 a month, and plebs get limited per month for $x,000?
As with any investing there's a risk appetite/timescale component to thinking about this stuff. Lots of companies went to zero in the dot-com bubble. Even Amazon was down over 90% between the end of 1999 and late 2001, and took until 2007 to recover to its high. NASDAQ overall took 15 years to return to its March 2000 high. Some incredible returns to be had if you waited it all out, to be sure, but it's hard to know what the interim looks like.
Yeah, only those evolved a lot from the initial products everyone hyped and products people hyped in 2000 are extinct or free. And I still don't understand where Facebook makes money. :)
Regarding LLMs there are two concerns - current products don't have any killer feature to lock in customers, so people can easily jump ship. And diminishing returns, if there won't be a clear progress with models, then free/small, maybe even local models will fill most of people needs.
People are speculating that even OAI is burning more money than they make, it's hard to say what will happen if customer churn will increase. Like for example me - I never paid for LLMs specifically, and didn't use them in any major way, but I used free Claude for testing how it works, maybe incorporating in the workflow. I may transitioned to the paid tier in the future. But recently someone noted that Google cloud storage includes "free" Gemini Pro and I've switched to it, because why not, I'm already paying for the storage part. And there was nothing keeping me with Anthropic. Actually that name alone is revolting imo. I wrote this as an example that when monsters like Google or Microsoft or Apple would start bundling their solutions (and advertise them properly, unlike Google), then specialized companies including OAI will feel very very bad, with their insane expenses and investments.
> And I still don't understand where Facebook makes money. :)
If that's a genuine question: Facebooks sells ads, information and influence (eg. to political parties). It's a very profitable enterprise. In 2024 Meta made $164B in revenue, and they're still growing at ~16% year-over-year.
Well, LLMs are themselves very broad. They encompass everything from web search to everything that you could automate yourself but don't have the time.
I don't LLM capacities have to reach human-equivalent for their uses to multiply for years to come.
I don't LLM technology as it exists can reach AGI by the simple addition of more compute power and moreover, I don't think adding computer necessarily is going to provide proportionate benefit (indeed, someone pointed-out that the current talent race acknowledges that brute-force has likely had it's day and some other "magic" is needed. Unlike brute-force, technical advances can't be summoned at will).
"Brute force" is only held back by economics and hardware limitations.
There are still massive gains to be had from scaling up - but frontier training runs have converged on "about the largest model that we can fit into our existing hardware for training and inference". Going bigger than that comes with non-linear cost increases. The next generations of AI hardware are expected to push that envelope.
The reason why major AI companies prioritize things like reasoning modes and RLVR over scaling the base models up is that reasoning and RLVR give real world performance gains cheaper and faster. Once scaling up becomes cheaper, or once the gains you can squeeze out of RLVR deplete, they'll get back to scaling up once again.
I think overstating their broad-ness is core to the hype-cycle going on. Everyone wants to believe—or wants a buyer to believe—that a machine which can grow documents about X is just as good (and reliable) as actually creating X.
Since APIs exist (including APIs that allow you to get a human to execute the task), a machine which can grow specs about X is able to create real-world X in many, many situations.
A machine which can define a valid CAD document can get the actual product built (even if the building requires manual assembly).
The dot-com was a bubble because investors pulled money and belief at the first sign of trouble.
The landscape has changed dramatically now. Investors and VCs have learnt if we stick with winners and growth companies, the payoffs are massive.
We also have more automatic, retail and foreign money flowing into the market. Buy the dip is a phenomenon that didn't exist at the scale it is now.
Pre-2015 if Big Money pulled out, the market was guaranteed to fail, but now retailers sometimes have longer views and belief (on people like Musk, Altman) than institutions and they continue to prop it.
So, it's foolish to apply 2000 parallels to now. Yes, history repeats, but doesn't with the exact time or price points
Also there are no early IPOs. Very few people can buy stocks in these companies which changes the dynamics significantly. Note sure what's the point of talking about the stock market this much when for almost everyone the only way to get any exposureis through Nvidia or other hardware companies and maybe MS/Google/AWS.
> Investors and VCs have learnt if we stick with winners and growth companies, the payoffs are massive.
Well... yes and no. 2021 wasn't that long ago.
> So, it's foolish to apply 2000 parallels to now
The stock market and other financial stuff is of course different. The fundamental trend not necessarily though. It took awhile for anyone to figure out how to directly build a highly profitable internet based business back then for AI it seems more or less the same so far.
dot-com bubble companies were not good companies. They either built something that was not novel so it could be copied, or had insufficient value to monetize. We'll see the same with current AI.
Similar to the invention of the web, AI is not a bubble. Real value has been created.
Cisco was the quintessential dot-com bubble company. Back then, it was what Nvidia is today: at the very spearhead of investors rallying behind the Internet.
"Good company" is subjective, but to argue that the company that built the backbone of modern web didn't make anything novel or monetizable is a bit short-sighted, don't you find?
And if you'd invested in Cisco then, it would have gone very badly for you. It was _wildly_ overvalued; at peak it had a P/E ratio of ~200 (even Nvidia's only about 50).
how so? Many internet companies that went bust like Webvan and Pets.com have successful equivalents in Instacart and Chewy?
I'm pretty sure many internet companies would be given a longer rope to survive now. E.g OpenAI and Anthropic will probably take years to get profitable but investors are OK with it
lol is a coping mechanism for the poor. If you really think top VCs / investors haven't learnt the long-term importance of staying the course, then you know nothing about the industry and mostly being influenced by popular social media posts shitting on the investor class.
There is a reason Anthropic/OpenAI and many startups are given much much longer ropes to be profitable than in the 2000 era when VCs pulled the rug the first opportunity of trouble
The thing that was most disillusioning for me here was SVB -- failure to apply basic principles of banking (i.e. they never really had a plan for central bank interest rates to change more than +-1%). Not just that the VC types running a bank weren't able to do so, but that such a large number of tech companies held all their cash on hand in a bank account (and didn't deposit anything in another bank, or a money-market fund/t-bills).
I have yet to be pleasantly surprised by the alleged collective wisdom of Wallstreet. I would hope that you are right, and that our corporate masters are smarter than I give them credit for, but I'm not going to hold my breath
It is foolish to compare to the dot com boom and bust. At least when that bubble burst we still had the global broadband internet that it built. When this bubble bursts, we'll have next to nothing to show for it.
I'm starting to think that making a bunch of tech companies the most valuable companies on Earth, and tying their value to everyone's ability to retire so the number must always go up was perhaps not the wisest thing to have done.
Of course we are in a massive bubble. At least in the US where I live, everything is about driving short term profits, with less than a little thought to what is good for the non-special interests/insider class.
I have lived and worked through two previous ‘AI winters’ and I expect the current bubble to eventually pop in a dramatic way. There will be good things produced by AI, but I am skeptical of the panic FOMO rush to AGI or super intelligence.
Look at the process of shifting manufacturing out of the USA: that was all about driving extreme wealth for special interest insiders. Sadly, most people look to their particular little political party for some form of relief - how is that working out?
I've begun to think some bubbles are good for the economy overall. In the dotcom days anyone with an idea and a domain name could get funding. I myself worked for a company that nabbed 7x more funding than needed but still failed due to poor leadership. I had reservations about the founder but thought I could help drive things, but he was even more absent than I ever anticipated.
A lot of VCs and PEs lost a lot of money during the crash. This means a lot of capital was spent in the economy, generating a lot of good activity, and the companies that failed then also put a lot more capital back into the economy through bankruptcies. Other businesses can pick up talent, IP, and assets for cheap, and everyone can learn from the failures. While losing that money isn't great for VCs, what they got was a very valuable education to be better stewards of their investments, and pick better companies. The next rounds of companies have to hit metrics, milestones, have to prove their value, etc.
Never waste a perfectly good crisis: learn if nothing else.
The problem you don't know which innings we are on and it may very well be the 4th/5th. More often than not you lose more by not participating in the run up and waiting for the collapse. I'd rather ride the wave (as intelligently as I can) than be on the sidelines just so I can say I told ya so.
Not in a bubble yet. Wait till your aunt takes out a second mortgage to chase NVDA stock to the moon. My guess is this will continue until later this decade with some bumps in between. That said, it’s absurd these guys are paying people so much money for what I think amounts to being context management off of some markdowns.
Bubble or not these kinds of deals could put a chill on the tech startup sector as well.
Things move fast in tech because there isn't nearly as much red tape and litigation as there are in other mature industries. This is because there's an agreed 'way of doing things'. Take funding, grow like crazy, sell/merge or IPO. Everyone wins or loses together (even if things are stacked in favour for some over others).
Once trust in this process is broken and founders or VCs start stacking the deck in their favour the game becomes rigged to the point where other people don't want to play anymore. Once that trust is gone red tape and litigation appears.
Yes and no. Are AI companies overvalued? Yes. Will most of them crash and burn? Yes, they will. Are the words "intelligence" or "reasoning" misused? Absolutely. Nonetheless, nobody can deny that some of these tools are useful and have demonstrated they can generate revenue like no other tool, app, or device before. There is something different from the dot-com bubble; many barriers have come down since 2000. Everyone can be connected 24/7 for a few tens of dollars a month. People trust the internet as a medium to perform transactions and access data. The real bubble is in the private market valuations, especially in the pre-seed stage. Many young entrepreneurs don't understand that raising their first round at $30M, $50M, or even $100M post-money will put a heavy weight on their shoulders approaching the Series A. Raising a funding round is a promise you make to the market. Increasingly high expectations will burn many wanna be entrepreneurs whose contribution to make things better for everyone will be lost forever. I'm deeply convinced that the reality check for all those companies is the public market, and in today's world, you can't go public after 3-5 years if your initial valuation has been built on a 20-year-long promise. All the trillion-dollar companies we know today went public a few years after their creation: Apple, Google, Microsoft, NVIDIA, etc. OpenAI, Anthropic, and Cursor are black swans, not evidence of the power law.
> the possibility that we’re in a massive bubble thats quite divorced from any kind of fundamentals
I'm astounded that anyone still genuinely believes we're not in a massive bubble. Of course AI company CEOs are going to say we're not and that AGI is just around the corner, it's deeply in their financial interest to keep inflating the bubble as long as possible.
Unpopular, controversial take: there should be an LSP extension that lets CLI agents, like Claude Code show diffs in the editor, and also one for completions, and sending snippets back to the CLI.
That, by itself, would obliterate the entire value of Windsurf or Cursor or whatever. The fact that Google has this kind of money and spends it on dubious "talent" (though none of these people are known in the community) is a testament to how overfunded tech companies are compared to the value that they provide.
> For other IDEs: The protocol is editor-agnostic. Any editor that can run a WebSocket server and implement the MCP tools can integrate with Claude Code.
Claude Code can already show diffs in JetBrains IDEs and VSCode ('/ide' command connects the CLI/TUI to plugin/extension running in the IDE-side).
It can also access the IDEs' real-time errors and warnings, not just compile output ('ideDiagnostics' tool), see your active editor selection, cursor position, etc.
It is known there were many developers without really much work to do that were hired only to be denied to competitors. Maybe it was cleaned up in the meantime
Big companies hired a lot, but I don’t think this specifically is true? In theory a high-value engineer would be productive, or else they aren’t worth stealing.
The simpler explanation seems more correct here — there was a lot of product fluff and a lot of headcount allocated to build that fluff.
Revenue is a funding source. Companies with too much money sometimes made really dumb decisions with that money.
The fact that one division of Google is wildly profitable does not exempt other parts of the company from criticism of their financially dubious choices.
Net profit can't be used as a measure of both "funding" and "value generation" while saying that a company is "overfunded" because it doesn't provide enough "value". Come back to your senses.
I'm genuinely not sure if you're not understanding, or deliberately ignoring my point.
I'll assume the former and try again. Maybe you didn't realize I'm not the person you originally replied to?
If a company is profitable, they have funds. The funds generated by the profits, can be used to fund additional internal projects. If the bucket of funds from profits gets ridiculously large, then it may begin to be used for vanity projects, like gutting an AI company, or building a gold statue of the founder. It seems reasonable to call companies spending on mostly-useless excesses "overfunded".
It's kind of their prerogative to spend their money how they like. If it's anti-competitive then you can hope that regulation exists to prevent that behaviour. As a shareholder you can complain that they will stop this value igniting behaviour. As an investor if you believe this behaviour is irrational you can short their stock and hope that the market is efficient and the share price will reflect the igniting value in their multiple.
As a bystander and outsider it is hard to isolate the value igniting behaviour from the moonshot behaviour. Shareholders love to gut a business of its risk taking and excess behaviour for predictable and inflated margins (and dividends) but the past 20+ years of our megacap companies is that they have continued to "innovate" in spite of all their inefficiencies.
I always have a chuckle when I recall how shareholders tried to oust Zuck for buying Instagram for 1B...
These vanity hires do seem frothy and reminiscent of dotcom style behaviour. But "AI" clearly will be game changing much like the internet was, and who at this stage can say what is worth recruiting people at the forefront of commercialising the tech right now
I'm not misunderstanding anything. It's just that the idea of a massively profitable company being labeled "over-funded" when excess profit is used in a way that you find disagreeable is stupid. It doesn't matter how well they allocate their revenue.
Stop looking at the entire world through the eyes of VC, because it doesn't work.
You're not funding google by paying for youtube, you're buying a service.
You didn't "overfund" your pizza shop that hired a stripper for friday night vibes, and neither did 99.99% of customers that paid google for their service offerings.
Gotcha, you're just being mean. You've got a bone to pick with VCs, I guess, so you've declared I listen to VC podcasts, see the world like one, and am using a stupid idiom.
That's fine. I'm going to continue referring to corporations that blow lots of money on random intra-industry dick measuring matches because they can as "overfunded", and you can continue expressing your opinion to anyone who will listen that this one person on the internet used an idiom that you think is dumb because it implies something other than "that person profited, therefore they did something right and therefore whatever they do with that money is correct and intelligent and never ever wasteful or dumb."
Windsurf is the first company that moved from IDE to training and trained good models. Unfortunately in code good models are not enough to win against claude 4. Any wrapper constraining the model is also doomed to fail.
> there should be an LSP extension that lets CLI agents, like Claude Code show diffs in the editor, and also one for completions, and sending snippets back to the CLI.
aren't you reviewing diffs in whatever diff tool you like? I find magit to be superlative for this (and for correcting and committing things).
I use Rust and found it's better to let the AI hallucinate function names, then let the compiler correct them. Rust's compiler is significantly better than TypeScript's at this, so it works well.
There should be an LSP integration that lets Cursor look up symbols and definitions the same way I do, instead of trying and failing to use grep / its fuzzy vector index thing!
> That, by itself, would obliterate the entire value of Windsurf or Cursor or whatever.
Have used Cursor and I know that there is quite a bit of value between the model and the chat input box and it will be similar to Claude Code or Codex, it's what makes this agentic, it's just accessed through a different interface. So from that perspective, Cursor makes sense for folks that are already in the VSCode environment.
On the contrary, I'm already in the VSCode environment and I would hate to have to switch to Cursor. Doesn't matter that it is a fork, I don't want a fork for every extension of my workflow. I used Cline and now Claude code integrated to VSCode, and I don't think I'm missing much by not switching to Cursor.
I symlinked my vscode settings into the cursor config directory and the transition was strictly additive; everything worked exactly the same as before, with two exceptions so far: a .NET extension wasn't supported, and Cursor recently forced a switch to their own version of Pyright (but that seemed like an improvement once the type checker diagnostics were tamed a little).
I don't see a justification for high valuations of companies that aim to build an "AI Software engineer". If something like Devin really succeeds, then anyone can use their product to simply build their own competing AI engineer. There's no moat, it's just another LLM wrapper SaaS.
Yep. The reality is folks building these types of companies are trying to get acquired as quickly as possible before the house of cards fall. This has led to a huge speculative rush of acquisitions to avoid FOMO later.
The technology is nowhere close to what they're hoping for and incremental progress isn't getting us there.
If we get true AGI agents, anyone can also build a multi-billion dollar tech companies on the cheap.
You're right - AGI would be unfathomable, it would be more productive than a quadrillion earths entirely populated by MIT valedictorians who just drank 2 espressos each. "Multi-billion dollar" would be a silly valuation.
I don't see this. The ai software engineer that succeeds, maybe it's because of a mixture of very complicated architecture derived from novel research etc. You can't replicate that with just hiring more human engineers, it takes time and effort and elite hiring. Plus enterprise support etc.
Devin etc will give you let's say 10x more engineering power, but not necessarily elite one.
i advise you to not take marketing lines too literally and be so casually dismissive as a result. you will miss a lot of good investments and startups this way and (worse) be lulled into a false sense of comfort and security.
Is it? Last I checked when you trained an LLM on another's output, at best you got the same performance as the original, and it was more likely you significantly degraded usefulness. (I'm not talking about distillation, where that tradeoff is known in return for a smaller, more efficient parameter set)
This is my exact takeaway too, and I'm always surprised it doesn't get mentioned often. If AI is truly groundbreaking, then shouldn't AI be able to re-implement itself? Which, to me, would imply that every AI company is not only full of software devs cannibalizing themselves, but the companies themselves also are.
This is my watershed for true AGI. It should be able to create a smarter version of itself.
Last I checked, feeding the output of an LLM back into its training data leads to a progressively worse LLM. (Note I'm not talking about distillation, which involves training a smaller model, by sacrificing accuracy. I'm referring to an equal or greater number of model parameters)
If the LLM is given the code for its training and is able to improve that, does that count? Because it seems like a safe bet that we're already there, the only problem is latency of training runs.
I was a bit confused as to what "Cognition" was, but they're the makers of Devin (edit: that just got added to the title, for reference), so that makes sense. Just buying the competition, the only surprise is they had more money to spend than the big ones.
Well Google did also just pay $2.5B to license Windsurf in perpetuity. Cognition is probably spending a lot less than that for just whatever it left after that type of a deal. Remaining team members, etc.
Circumvents regulatory oversight and also shafts 99% of the employees. Seems to be a backdoor way to acquire the key founders/leaders and IP (via a perpetual license) while leaving behind a desiccated husk of rank and file employees, customers, and obligations.
No, as some portion of the $2.5b goes to Windsurf investors.
Basically, Google bought the top talent from the company. This cash was used (according to articles I read this morning) in part to pay directly out to shareholders, and in exchange Google got the top talent from the company and a license for the software (probably mostly so their new talent didn't have to worry about NDA, non-compete, and patent challenges).
Since this money went to shareholders, not to the company bank, and since top talent fleeing the company reduces the value of the company the overall value of Windsurf likely went down as part of the Google deal. This in turn likely made it cheap enough for the remainder to be purchased by Cognition.
Let this be a learning lesson in judging these deals based on partial information. Kudos to the Google, Windsurf, and Cognition teams for keeping all of these deals under wraps until announcement (OpenAI could learn something...), but even so it's likely that we the public will never learn every detail of what transpired. I've seen a lot of harsh, misguided takes over the past few days, like that the Windsurf founders screwed over their employees, or that OpenAI reneged on the deal. In this case, this seems like a happy ending for all parties involved: congrats to the Windsurf team!
> I've seen a lot of harsh, misguided takes over the past few days, like that the Windsurf founders screwed over their employees [...] In this case, this seems like a happy ending for all parties involved
There is no evidence at all in the announcement that is the case. It just says "100% of Windsurf employees will participate financially in this deal". What "participate financially" looks like is not elaborated upon.
It is possible you're right. It's also equally possible that the founders have still screwed over their employees, we just don't know. Nothing in this post supports either position.
>It is possible you're right. It's also equally possible that the founders have still screwed over their employees, we just don't know. Nothing in this post supports either position.
In the lack of evidence, its okay to assume the most likely scenario, which is the executives & shareholders will make out like bandits and everyone else is likely to at best, get pennies.
If my understanding is correct, this is still a much worse deal for employees than if Windsurf's exec team had negotiated a "standard" "accelerated vesting, common conversion" acquisition with Google.
Presumably the "payout" from Cognition is at a lower nominal value and in illiquid (and IMO overvalued) shares in Cognition rather than cash.
Did OpenAI ever actually announce anything publicly regarding a potential windsurf acquisition?
AFAICT most of the reporting was based on rumors or leaks. But they never actually announced an acquisition. Seems like Bloomberg may have made an oopsie here.
Unfortunately this seems to be true. I like Windsurf, but these days I just use it as a harness for running Claude Code while still retaining decent code completions.
Start with a clean git (even if you're just committing WIP locally). Run the tool. See the diffs in your IDE. They show up immediately in IntelliJ; use the "commit" pane. Same as if I had edited a bunch of files.
not sure of windsurf's ToS, but Google could be after the usage data (e.g. did the user accept the suggested edit? if not, what did they change it to? etc.)
VC's probably did fine. Their preferred shares get paid out on the dividend out of the Google 'license' fee. They might get some stock in Cognition as part of the acquisition of the remains of the company, but they've likely already been made whole and made a tidy profit on money invested in early rounds.
I’m really confused now. Also, is there really that much of a transformative difference between Windsurf and say OpenAI/Claude etc to warrant this crazy valuations?
I'd love to learn more about that arrangement. Maybe Google's terms indicated that the "licensing fees" portion of "$2.4 billion in licensing fees and for compensation" are void if Windsurf gets acquired.
I just started trying out Codeium in Nvim on a recommendation from a video that's like, less than 6 months old. Oops, actually, codeium is Windsurf now. But the tool is still codeium.nvim at least.
It's weird being a Windsurf fan on HN. You get the feeling that you're in the minority or doing it wrong because the dialog is dominated by Cursor and (more recently) Claude Code.
It's also been a lot of random stuff recently with their 3 separate Ross and Rachel acquisition storylines.
Some takeaways:
1. Devin/Cognition definitely have a legit AI dev agent now
2. It's crazy what Google passed on. The fact that it was worth it to them without the traditional best assets is wild. Guess that's what happens when you play on ultra hard mode with an infinite money glitch.
3. I am worried/pre-emptively sad that Windsurf will likely go away or get nerfed, more expensive etc.
"Acquiring" is often synonymous with "merger". Technically the surviving company acquires the other, but effectively it's a merger. Not always, but it's not uncommon to see two weaker companies merge in a competitive landscape to survive. It also counts as a liquidity event allowing employees some financial levers.
They probably got what got left of it in a cashless deal. Basically, the shareholders got to exchange X shares in a fatally wounded company into Y shares in a still-alive startup. The economic sense depends on the ratio between X and Y, but if the board was close to panicking due to recent events, Cognition probably got a good deal.
I think google is buying windsurf, they are leaving 100% of assets and some employees, and windsurf is buying that 100m in assets + taking the leftover employee liability.
Microsoft poached the talent, devin Co. Picks up the scraps
I have been testing Devin for a long time, early access and all. I'm not impressed by it at all, a decent developer with their LLM of choice does a far better job.
I mean isn't that amazing for an 1 year old product? If it's already better than a terrible dev with an LLM, or better than a decent dev without an LLM, it's not hard to imagine in 2-3-5 years Devin is better and cheaper than most hires you could do. Without having to do HR, equity etc.
Does anyone actually use Windsurf? I know a ton of people on Cursor, Cline, Roo, Claude Code... but zero people in my engineering circles have even mentioned trying Windsurf.
For me it was substantially better than Cursor. Their RAG/code indexing/whatever implementation is better (Cursor seems to completely fall apart on larger codebases, unless you keep a very detailed cursorrules file). The cmd+tab was also substantially better. Plus it's cheaper than using Cline.
But with all this changing of hands, I'm not sure I can trust it going forward at all, so I guess it's back to looking for alternatives.
I tried it once and couldn’t understand why it exists. Neither windsurf nor cursor seem to offer anything Cline doesn’t, and the real Microsoft developed tools like Pylance are also broken and can’t be replaced with shitty AI.
I was using Windsurf in Pycharm, until I switched to Claude Code / Gemini CLI last week. I am also finding that Copilot with Sonnet 4 is pretty on par to Windsurf.
Very curious how we'll look back on Google spending 2 billion dollars to "license IP" and hire a handful of people.
If there's 47m software engineers in the world, at $200/month, and 50% gross profit that's a $56 billion TAM. Not crazy to think it's more if we include the adjacent space of analyst roles that write software (sql, advanced excel, etc).
They'll have to crush it to make a $2 billion acquihire look reasonable, but it's possible.
It’s more about not getting embarrassed ever again. The breakthrough papers for LLMs came from Google, but not the product. This was embarrassing. Even if you have to spend double on everything from now on you do it, because Google effectively got sucker punched by ChatGPT.
that's not the market they're thinking of, they're thinking of the total amount of money spent on developers per year, globally, and capturing a percentage of that
They are doing what they've always done skirting the edges of antitrust laws by buying up competitors before they become competitors so that they can shut them down.
> This sounds like a happy ending for the employees of Windsurf and a good deal for Cognition
The employees were robbed from having a big cash exit. Illiquid stock options from Windsurf were converted to illiquid stock options of Devin.
What's worse is that the well is now poisoned. I would advise against joining startups from now on, because I think that there's no upside for employees anymore.
"work to" is key.. sounds like they aren't fully accelerating, just accelerating the time they worked that hasn't vested. So if you were hired a week ago, you get a week's worth of participation.
Great news...saw Windsurf vibing on YC podcast several months ago...and it's my go-to now. Google Gemini Pro 2.5 works best for me (Python, C++, Fortran).
Just finished a weather analytics code translation test from Fortran to C++ and Python. It was simple temperature high/low averaging, so to amp up the processing, I worked with Gemini to add a Monte Carlo simulation for a 7-day weather forecast, and a 2D heat diffusion/advection physics model to calculate the temperatures across a NxN grid over T time divisions.
I've never heard of Cognition (not a slight on them, its just me being disconnected). How much was the deal for and how do they have this type of money? Wasn't there just an article out yesterday or so saying that Windsurf's leadership was going to google?
I totally missed the puck with this one. There was a time where Cursor did not have the agent feel of a true AI pair-programming buddy. Windsurf had that magical aspect, and I totally thought they would destroy Cursor. But it took Cursor about 2 weeks by my recollection to add agent mode, and ultimately I went back to Cursor because of their better WSL2 integration.
For those brief 2 weeks, Windsurf felt like the SOTA tool. Crazy how the winds change.
Same experience. I was using Windsurf for a couple of weeks, but felt like the editor wasn't as nice to use as Cursor, so features an agent could handle went to Windsurf and pair coding was more with Cursor. Once Cursor got agent mode, I haven't really touched Windsurf.
Feels like a new SOTA tool every couple weeks. Heck, the post below this is about a new agentic IDE.
As someone who stuck with it - it's good. And it's not actually just a thin wrapper. Their tab model is really, really good now (I think that was Cursor's big advantage for a long time). And they also price very fairly (read: heavily subsidized, minus Claude).
I writing this just after the night end llm cannot even replace to using Trans i18n react component if it is not in context of the file. We are still far from agi
> To that end, Jeff and I worked together to ensure that every single employee is treated with respect and well taken care of in this transaction. Specifically:
> 100% of Windsurf employees will participate financially in this deal
> 100% of Windsurf employees will have vesting cliffs waived for their work to date
> 100% of Windsurf employees will receive fully accelerated vesting for their work to date
Not clearly called out. That call out is as clear as mud. We have no idea how much if any money the left behind employees will get and it's almost certainly far less than they would have got if their execs had held out for a deal that benefited everyone instead of their selfish and damaging to the entire startup ecosystem deal with Google.
Having been acquired by Google, there is always a leeway for the execs to take the employees with them. Google is a weak negotiator when they NEED something.
On the other hand, I can imagine the execs taking Google golden handcuffs while trying to close the Cognition deal so the employees are made whole or maybe even on better terms than if they all went to Google.
> This transaction is structured so that 100% of Windsurf employees will participate financially. They will also have all vesting cliffs waived and will receive fully accelerated vesting for their work to date.
Not important, marketing speak unless it comes with numbers. "We'll split 100% of $10 across every employee" is just as much a possibility as the windfall their PR team has convinced you of.
Uhh so where does the 2.4B go to? For the "licensing" rights but without equity to Windsurf? Does the whole 2.4B get distributed amongst the talent that Google acquired or is it shared amongst all Windsurf employees?
Windsurf employees left behind by their executives get none of that Google money. It went to the selfish exec team and a few top employees that abandoned their teams and projects for Google's cash, and what didn't go to those selfish jerks went to Windsurf's investors. The rank and file that were abandoned by their leadership got totally screwed. Startup executives will never again be trusted to deal fairly with the people who spend their opportunity trying to make that particular startup work rather than some other, or simply going to work for an incumbent. This deal probably did more damage to the tech startup ecosystem than anything I've seen in my 25 years in Silicon Valley.
did the team that joined deepmind forfeit their equity or will the cognition acquisition basically allow them to get a double payout from their equity?
The Devin name/branding is so toxic that nobody will try their current product offerings, so the hollowed out shell of a respected company is actually fine for their needs.
The company was intact after key employees moved to Alphabet, and importantly there was a service contract to use Windsurf, so apparently Google will be paying Cognition, ultimately, now.
I'm just speculating but you'd get to use the IP without needing to negotiate overmuch about it, if they're using GCP or can be retrofitted to then you could set prices at a discount but claim a market rate when discussing growth of usage, which is great at quarterly earnings report time.. or even just as an ease-of-transition to expand into Windsurf's existing user base? There are plenty of reasons, including negotiation leverage, optics, or even just appeasment to the founders or board that the original project doesn't get immediately destroyed.
Why? Why presume that? These employees got screwed. They were left with a shell of a company that got bought for pennies on the dollar while their executives and executive pets all got massive payouts moving to Google. It was a total dick move and the result will be a flight from startups because who wants to bust their asses for an executive team that will leave them behind when the payout arrives.
This is the logical, satisfying, and probably best conclusion to an un-ideal and optically terrible situation all parties were placed in.
First, OpenAI wanted to acquire Windsurf. Terrific move! Win-win for OpenAI (who needs more AI product) and Windsurf (for the deal price). But this fell apart because Windsurf didn't want the IP to go to Microsoft (which imo should not have been not a big deal, especially if you knew what would have happened next). Big loss for all parties for this to have fallen apart.
My biggest question still is why not continue on as an independent company? Perhaps losing access to Claude doomed signups; perhaps employees/investors had a taste of an exit and still wanted it; perhaps due to fiduciary duty to maximize returns; perhaps their growth stalled due to the announcement? In any case, the founders got a similarly equivalent deal from Google, and were arguably wise to pursue it.
But Google's Corp Dev team here is the most maddening. Why not fully acquire the entire company, instead of doing the same "acquihire and license" deal that was done to Character AI, Adept, Scale, etc.? Risk of FTC antitrust review is a thing, but Google's not even competitive in the coding market, so I doubt there is a review (though I do hear that all acquisitions by large tech companies these days are reviewed by default). If there's anyone to blame in this situation, it's the FTC and Google for pursuing this strategy, instead of a full acquisition. Win-win for Google (for the team) and Windsurf (for getting a similar acquisition price, but liquid!).
Imo, the founders did a good job ensuring that close to the $3B acquisition price was reflected in the $2.5B Google deal--all existing investors and vested employee/equity holders are paid out; the company also retained $100M which was suspiciously similar to the amount needed to pay out all unvested employee/equity holders [1]. So theoretically the remaining company could pay accelerate vesting, then pay out the cash to their remaining employees, and then shut down, to give everyone the same exit as an acquisition, or better. This might have been the best scenario, because the brand damage to Windsurf as an IDE that happened over the weekend was pretty close to unrecoverable for them as an independent company.
But instead, the company leadership decided to field acquisition offers for the remaining company and IP, and got one from Cognition. (I'm actually surprised this acquisition isn't under FTC review; it's more plainly an agentic coding company acquiring a competitor agentic coding company). In taking the offer, it reinforces that the Windsurf IDE will continue to exist, that they have a R&D team backing the IDE again, and can marry Windsurf's enterprise sales chops with Cognition's product [3]. Win-win for both Cognition and Windsurf.
So overall, win-win-win all around, except for OpenAI, Varun's public reputation (imo, undeserved), and startups hiring employees (who might think they might not get a proper exit) [2].
> -all existing investors and vested employee/equity holders are paid out;
But the statement from Cognition was:
> 100% of Windsurf employees will participate financially in this deal
> 100% of Windsurf employees will have vesting cliffs waived for their work to date
> 100% of Windsurf employees will receive fully accelerated vesting for their work to date
The details matter. "vesting cliffs waived" meaning what? Windsurf shares exchanged for Cognition shares? at what ratio?
"Participate financially" means what exactly? They could all get a coupon for a free doughnut, and that statement would be true.
I'm not saying the employees are getting nothing, or even a raw deal. I'm saying we have no idea if the deal is good for them, without details.
> theoretically the remaining company could pay accelerate vesting, then pay out the cash to their remaining employees, and then shut down, to give everyone the same exit as an acquisition, or better.
unlikely that will happen. More likely the investors and VCs will take the lion's share of the $2.5B, that is what they do. That is why they exist. And they'll distribute as thin a slice as possible to the employees.
Historically, when a company gets acquired, the terms of the acquisition vary wildly. Many acquisitions over the past few years have led to a payout of all equity holders, pennies for employees, and layoffs for much of the existing team. This outcome is no worse than an existing acquisition--people just want details because of the new structure. My point is that this is financially no worse than an existing acquisition. It just _feels_ worse because of how it's structured.
And to your last paragraph, read reference [1]. The distribution of that 2.5B is in accordance to the existing cap table; it will make sense once you read that tweet. You must allocate money according to the cap table, and so that allocation is already determined in a company's previous funding rounds.
I wonder if we'll ever stop to ask ourselves if faster and faster output of software is actually a good thing for the world. Or will we just continue because it's just what we do nowadays in civilization to get ahead?
A lot of people are asking that question, and the answer is emphatically yes. All improvements to the human condition are rooted in technology, and software is technology. Who's to say the latest advancements aren't some tech tree precursor to cure an ailment impacting millions - how could you argue against that? The genie is out of the bottle.
I would argue against it if the downside is even more technological enslavement for billions.
And while many improvements to the human condition are rooted in technology, many of the problems of humanity are rooted in it as well. There might very well be an optimal point that we've already past.
ahead of what? ahead of generating ugly and mostly usable masks over the same data. I'm in favor of AI but it seems to me that no one really stopped asking himself what real problems people have and how to actually fix them
See, most closed source software really just pisses me off of ideological reasons, I just like to tinker with things and just having the possibility to do so by being provided the source code really helps my mind feel happy I guess.
So I "vibe coded" a game that I used to play and some projects that I was curious about and I just wanted to tinker too. sure the game and code have bugs.
Also with the help of AI, I feel like I can tinker about things that I don't know too much about and get a decent distance ahead. You might think that I am an AI advocate by reading this comment, but quite the contrary, I personally think that this is the only positive quality that AI helped in quite substantially.
But at what cost? The job market has sunk a large hole and nobody's hiring the junior devs because everybody feels better doing some AI deals than hiring junior devs.
My hunch is that senior devs are extremely in demand and are paid decently and so will retire on average early too. Then, there would be a huge gap b/w senior and juniors, because nobody's hiring the junior engineers now, so who will become the senior engineers if nobody got hired in the first place. I really hope that most companies actually realize that the AI game is quite a funny game really, most companies are too invested into it to realize that really, open source AI will catch up and there is just no moat with AI and building with AI or just doing stuff with AI isn't that meaningfully significant as they think it is as shown by recent studies.
Sorry I guess, I may have been incorrect in that regards. I actually just meant as in comparison to juniors really. And I personally felt that way from what I've heard from all the people, I am not sure too about salary rising but still I always thought that seniors are getting on with more and more responsibility since juniors aren't getting hired and so I thought that they were more compensated and I am pretty sure that I heard it somewhere and I think I just repeated that.
Also maybe I felt this way because of 100 Million $ and the 30 Billion $ acquisition by Zuckerberg I guess
And I would like to quote a part from the hackerrank ie.
Taken in isolation, this might suggest a cautious but healthy rebound. But viewed through a 2025 lens, a deeper pattern emerges: teams are leaning hard into experience, and leaving early-career talent behind.
I don't even remember anymore. I guess hitting myself in the head worked..?
But in general my reason for hitting myself in the head is that I'm an IDE author and people keep tripping over themselves to place great value on forks of VSCode
Could you please stop posting unsubstantive comments and flamebait? You've unfortunately been doing it repeatedly. It's not what this site is for, and destroys what it is for.
My name is Devin; it has been for many decades now. I'm embarrassed to see you've named your product after me. It has already prompted uncomfortable jokes at my expense, and I'm sure there will be more. I now have newfound empathy for people named Alexa.
For instance, people have made jokes about my name in interviews, and it's embarrassing for me, and thus awkward for everyone, and awkward interactions make it objectively less likely that I will get job offers.
I don't think any product should be named after people. Please change the name of Devin.
Apart from this news I haven't heard about Devin in a long time. Not since the very start of the AI hype. Considering how foolish a purchase this seems to be I don't think you're going to have to worry about them Bogarting your name for too much longer.
I think the amount of turmoil around these deals is giving more weight to the possibility that we’re in a massive bubble thats quite divorced from any kind of fundamentals. Sooner or later the bubbles gonna burst.
> divorced from any kind of fundamentals
Anthropic ARR went $1B -> $4B in the first half of this year. They're getting my $200 a month and it's easily the best money I spend. There's definitely something there.
"Sooner or later the bubble's gonna burst" and "There's definitely something there" aren't mutually exclusive - in fact they often go together.
It makes me perhaps a little sad to say that "I'm showing my age" by bringing up the .com boom/bust, but this feels exactly the same. The late 90s/early 00s were the dawn of the consumer Internet, and all of that tech vastly changed global society and brought you companies like Google and Amazon. It also brought you Pets.com, Webvan, and the bajillion other companies chronicled in "Fucked Company".
You mention Anthropic, which I think is in a good a position as any to be one of the winners. I'm much less convinced about tons of the others. Look at Cursor - they were a first moving leader, but I know tons of people (myself included) who have cancelled their subscription because there are now better options.
> It makes me perhaps a little sad to say that "I'm showing my age"
Please don't say stuff like that.
As a 20-something who was in diapers during the dot-com boom, I really appreciate your insight. Thanks for sticking around on HN!
Thanks, that's very kind!
I probably mean it less as "I'm too old" and more of "Wow, time really flies".
To me, who started my career in the very late 90s, the .com boom doesn't really seem that long ago. But then I realize that there is about the same amount of time between now and the .com boom, and the .com boom and the Altair 8800, and I think "OK, that was a loooong time ago". It really is true what they say, that the perception of time speeds up the older you get, which is both a blessing and a curse.
Regarding AI, it's a bit fascinating to me to think that we really only had enough data to get generative AI to work in the very recent past, and nearly as soon as we had enough data, the tech appeared. In the past I would have probably guessed that the time between having enough data and the development of AI would have been a lot longer.
Seconded
> It also brought you Pets.com, Webvan, and the bajillion other companies chronicled in "Fucked Company".
The irony with Webvan, they had the right idea about 15 years too early. Now we have InstaCart, DoorDash, etc. You really needed the mobile revolution circa 2010 for it to work.
Pets.com is essentially Chewy (successful pet focused online retailer)
So, neither of those ideas were really terrible in the same vain as say Juicera, or outright frauds like Theranos. Overvalued and ill-timed, sure
Don't forget that Amazon's market cap was higher in 1999 than 2009!
What
March 1999: ~$27.7B
Jan 2009: ~$25B (back to $27.7B & rising by Feb)
Huh.
I genuinely don't understand what value Cursor itself brings. It's like a wrapper for some APIs, right? As far as I can tell there's like four actual AI firms in the world and everyone else is trying to whitelabel. It reminds me of the hosting industry in the early 2000s.
The value is the UX/DevX. Though, they are essentially just a fork of VS Code so it's hard to justify using instead of VS Code + Copilot or Continue which is almost the same UX now.
That's the problem with most "AI" products/companies that still isn't being answered. Why do people use your tool/service if you don't own the LLM which is most of the underlying "engine"? And further, how do you stay competitive when your LLM provider starts to scale RL with whatever prompting tricks you're doing, making your product obsolete?
Cursor isn't my preferred environment for development, but for me, it yields much better results than Copilot. Continue and Cline bugged out on me so badly, and so frequently, that I didn't find them worth using. YMMV
If you're interested, I recommend checking out Kiro from Amazon.
Every time I've tried Copilot or Cursor, it's happily gone off and written or rewritten code into a state it seemed very proud of, and which didn't even work, let alone solve the problem I put to it.
Meanwhile, Kiro:
1. Created a requirements document, with user stories and acceptance criteria, so that we could be on the same page about the goals
2. Once I signed off on that, it then created a design document, with code examples, error handling cases, and an architecture diagram, for me to review
3. After that looked good, it set about creating an itemized task list for each step of the implementation, broken down into specific tasks and sub-tasks and including which of the acceptance criteria from step 1 that task addressed
4. I could go through the document task by task, ask it to work on it, and then review the results
At one point, it noticed that the compiler had reported a minor issue with the code it had written, but correctly identified that resolving that issue would involve implementing something that was slated for a future task, so it opted to ignore the issue until the appropriate time.
For once, I found myself using an AI tool that handled the part of the job I hate the most, and am the worst at: planning, diagramming, and breaking down tasks. Even if it hadn't been able to write any working code at all, it already created something useful for me that I could have built off of, but it did end up writing something that worked great.
In case anyone is curious about the files it created, you can see them here: https://github.com/danudey/rust-downloader/pull/4
Note that I'm not really familiar with Rust (as most of the code will demonstrate), so it would probably have been far faster for an experienced Rust programmer to implement this. In my case, though, I just let it do its thing in the background and checked in occasionally to validate it was doing what I expected.
What you are describing is effectively my work flow with any LLM tooling, it's just not formalized by the tooling itself. Sounds interesting, I will take it for a spin.
Wow, that sounds like it might be useful!
Don't they have their own model for the inline completions? For me I find it really nice to preserve some brain energy by doing one repetitive change and just tab spamming it to get it done everywhere. I could get it done maybe just as fast with a macro in vim, but, Cursor lets me preserve the brain energy for something harder.
Meanwhile other "wrappers" e.g. in nvim or whatever, don't have this feature, they just have slightly better autocomplete than bare LSP.
> I genuinely don't understand what value Cursor itself brings. It's like a wrapper for some APIs, right? As far as I can tell there's like four actual AI firms in the world and everyone else is trying to whitelabel. It reminds me of the hosting industry in the early 2000s.
Yes, there's (maybe?) four, but they're at the very bottom of the value chain.
Things built on top of them will be higher up the value chain and (in theory anyway) command a larger margin, hence a VC rush into betting on which company actually makes it up the value chain.
I mean, the only successes we see now are with coding agents. Nothing else has made it up the value chain except coding agents. Everything else (such as art and literature generation) is still on the bottom rung of the value chain.
That, by definition alone, is where the smallest margins are!
> I genuinely don't understand what value Cursor itself brings. It's like a wrapper for some APIs, right?
By similar token Windows is mostly a wrapper around Intel and AMD and now Qualcomm CPUs. Cursor/Windsurf add a lot of useful functionality. So much so so that Microsoft GitHub Copilot is losing marketshare to these guys.
These four AI companies are nothing more than re-packaging GPUs and training data.
Amazing how folks make comments without even trying it and especially making a comment similar to how Dropbox is simply rsync, right?
It is a lot less trivial than people like yourself make it out to be to get an effective tool chain and especially do it efficiently.
That's why Dropbox is a trillion dollar company and not a feature called iCloud, right?
Are you trying to make a point or just being defensive for no reason? You called out something without having any experience or knowledge and then did the classic “it’s just a wrapper”.
Sorry, I thought the point was clear: Dropbox (file sync) is a feature, not a product. Cursor (AI in your IDE) likewise, is a feature, not a product.
I am old and I remember when you could make a lot of money offering "Get Your Business On The Information Superhighway" (HTML on Apache) and we're in that stage of LLMadness today, but I suspect it will not last.
“It's like a wrapper for some APIs, right?”
Don’t be sorry it shows your true colors. The point stands that you continue to step around. Cursor and other tools like it are more than a trivial wrapper but of course you have never used them so you have no idea. At least give yourself some exposure before projecting.
Dropbox is still a $5+bn business. Cursor is still growing, will it work out, I don’t know but lots of folks are seeing the value in these tools and I suspect we have not hit peak yet with the current generation. I am not sure what a service business like a small biz website builder has to do with Cursor or other companies in adjacent spaces.
What "true colors"? I think I've been pretty consistent that I think Cursor is a commodity. You're surprisingly hostile and defensive about your preferred autocomplete plugin. You're right, I haven't used Cursor, but I've used similar tools like Copilot.
Your characterization of hosting as "a small biz website builder" is revealing. https://finance.yahoo.com/quote/GDDY/ is the one that made it and is now a $24B firm, but there were at least dozens of these companies floating around in the early 2000s.
Why are you so sure Cursor is the new GoDaddy and not the new Tripod? https://www.tripod.lycos.com/
You’ve been dismissive from the start without any real engagement with the product. I pointed out that you haven’t used Cursor, and rather than reflect on that, you’ve responded with sarcasm and condescension.
The only person being defensive here is you. My point was simple: tools like Cursor are more than just “wrappers.” Whether it becomes a massive business or not, revenue is growing, and clearly many users find enough value to justify the subscription. You don’t have to like it but writing it off without firsthand experience just weakens your argument.
At this point, you’re debating a product you haven’t tried, in a market you’re not tracking. Maybe sit this one out unless you have something constructive to say beyond “it’s just a wrapper”.
It's a very well done wrapper that improves your coding productivity a lot.
The problem is that they have no moat and the underlying provider can easily cut them out.
I think you underestimate the difficulty in getting the tool chain running efficiently in the IDE. It's a significant moat and I suspect their spend is too attractive to cut them off from an API especially when most of the model providers are not exactly competing fully in this space yet or at least not with the same enthusiasm.
Interesting to see multiple posts here saying this. Pretty clearly it isn't true. The IDE is owned by Microsoft. The model is owned by Anthropic or Google or whoever. A business can't be made from a thin sandwich filling between the two.
Hey guys we have a guy here stating pretty clearly it isn’t true. He clearly is the authority on the topic because he said so.
The shell of the IDE is open source. It’s true there is some risk on the supply of models and compute but again none of those, except MSFT which does not even own any of the SOTA models, have any direct competition. OpenAI has codex but it’s half baked and being bundled in ChatGPT. It is in nobodies interest to cut off Cursor as at this point they are a fairly sustained and large customer. The risk exists but feels pretty far fetched until someone is actively competing or Cursor gets bought out by a OpenAI.
Again, what proof do you have that there is zero complexity or most being driven by the sandwich filling. Most of OpenAIs valuation is being driven by the wrapper ChatGPT not API usage. I have written a number of integrations with LLM APIs and while some of it just works, there is a lot of nuance to doing it effectively and efficiently at scale. If it was so simple why would we not see many other active competitors in this space with massive MAUs?
Not really, it's pretty hard to get the editor and code editing via AI working as well as they did.
If it's so hard, then why are there multiple open source projects that are just as good?
They're not just as good unless you are willing to spend huge amounts in API credits.
> and the underlying provider can easily cut them out
what? Do you think providers (or their other customers) don’t care about the business implications of a decision like this? All so that cursor can bring their significant customer base to a nearly-indistinguishable competitor?
Anthropic is actually a good point to focus on since Claude is very good proof that it's not about the scaling. We are not quite there yet but we are "programming" through how we shape and filter the input data for training it seems. With time, we'll understand the methods to better represent.
Current situation doesn't sound too good for "scaling hypothesis" itself.
> Current situation doesn't sound too good for "scaling hypothesis" itself.
But the “scaling hypothesis” is the easiest, fastest story to raise money. So it will be leveraged until conclusively broken by the next advancement.
what're you finding better than cursor now?
Claude Code with Pro, Max100, or Max200 subscriptions. Works with any IDE including none.
For the time being, nothing comes close, at least for me.
Can you please share your Claude usage workflow?
I use Github copilot and often tend to be frustrated. It messes up old things while making new. I use Claude 4 model in GH CP.
I use github copilot chat right now. First I use ask mode to ask it a question about the state of the codebase outlining my current understanding of the condition of the code. "I'm trying to x, I think the code currently does y." I include a few source files that I am talking about. I correct any misconceptions about the plan the llm may have and suggest stylistic changes to the code. Then once the plan seems correct, I switch to agent mode and ask it to implement the change on the codebase.
Then I'll look through the changes and decide if it is correct. Sometimes can just run the code to decide if it is correct. Any compilation errors are pasted right back in to the chat in agent mode.
Once the feature is done, commit the changes. Repeat for features.
I also do the same. I am on the 200$ maxpro plan. I often let the plan go to pretty fine level of detail, e.g. describe exactly what test conditions to check, what exact code conditions to follow. Do you write this to a separate plan file? I find myself doing this a lot since after compaction Claude starts to have code drift.
Do you also get it to add to it's to-do list?
I also find that having the o3 model review the plan helps catch gaps. Do you do the same?
Does it remember context from chat mode and when you switch to agent mode?
Yes, it can't change between edit and ask/agent without losing context but ask <-> agent is no problem. You can also change to your custom chat modes https://code.visualstudio.com/docs/copilot/chat/chat-modes without losing context. At least that's what I just did in VSCode Insiders.
Here are some nice copilot resources: https://github.com/github/awesome-copilot
Also, I am using tons of markdown documents for planning, results, research.... This makes it easy to get new agent sessions or yourself up to context.
Yes. I think it used to be separate tabs, but now chat/agent mode is just a toggle. After discussing a concept, you can just switch to agent mode and tell it to "implement the discussed plan."
GitHub Copilot models are intentionally restricted, which unfortunately makes them less capable.
I'm not the original poster, but regarding workflow, I've found it works better to let the LLM create one instead of imposing my own. My current approach is to have 10 instances generate 10 different plans, then I average them out.
Github Copilot is weirdly bad, and all the alternatives are better. Sometimes people think "they have the same model, must be the same," but it's not.
This was my answer as well. And I think it just highlights all the serious dangers for the "API wrapper companies" compared to the foundation model companies.
User experience is definitely worth something, and I think Cursor had the first great code integration, but then there is very little stopping the foundation model companies from coming in and deciding they want to cut out the middleman if so desired.
Amp Code is also very good, they released about 2 weeks ago their Oracle feature which leverages o3 to do reviews (https://ampcode.com/news/oracle). Amp leans the closer to Claude Code more than other solutions I’ve seen so far, the team there is really leaning into the agentic approach.
I watch the changes on Kilo Code as well (https://github.com/Kilo-Org/kilocode). Their goal is to merge the best from Cline & Roo Code then sprinkle their own improvements on top.
My problem with Claude code versus Cursor is that with Cursor I could "shop around" the same context with different foundational model providers, often finding bugs this way or or getting insights.
Sometimes one model would get stuck in their thinking and submitting the same question to a different model would resolve the problem
I’m unaffiliated, but I’ve really been enjoying this - https://github.com/BeehiveInnovations/zen-mcp-server
It allows you to have CC shoot out requests to o3, 2.5 pro and more. I was previously bouncing around between different windows to achieve the same thing. With this I can pretty much live in CC with just an editor open to inspect / manually edit files.
Cline is absolutely fantastic when you combine it with Sonnet 4. Always use plan mode first and always have it write tests first (have it do TDD). It changed me from a skeptic to a believer and now I use it full time.
How much is it costing you?
I use Roo Code (Cline fork) and spend roughly $15-30/mo by subscribing to Github Copilot Pro for $10/mo for unlimited use of GPT-4.1 via the VS Code LM API, and a handful of premium credits a month (I use Gemini 2.5 Pro for the most part).
Once I max out the premium credits I pay-as-you-go for Gemini 2.5 Pro via OpenRouter, but always try to one shot with GPT 4.1 first for regular tasks, or if I am certain it's asking too much, use 2.5 Pro to create a Plan.md and then switch to 4.1 to implement it which works 90% of the time for me (web dev, nothing too demanding).
With the different configurable modes Roo Code adds to Cline I've set up the model defaults so it's zero effort switching between them, and have been playing around with custom rules so Roo could best guess whether it should one shot with 4.1 or create a plan with 2.5 Pro first but haven't nailed it down yet.
Looking at Cline, wondering what the real selling points for Roo Code are. Any chance you can say what exactly made you go with Roo Code instead of Cline?
Cline has two modes (Plan and Act) which work pretty well but Roo Code has 5 modes by default. (Code, Ask, Architect, Orchestrator, Debug) and it's designed so that users can add custom modes. e.g. I added a Code (simple) mode with instructions about the scale/complexity of tasks it can handle or decide to pass it to Code for a better model. I also changed the Architect mode to evaluate whether to redirect the user to Code or Code (simple) after generating a plan.
Roo Code just has a lot more config exposed to the user which I really appreciate. When I was using Cline I would run into minor irritating quirks that I wished I can change but couldn't vs. Roo where the odds are pretty good there are some knobs you could turn to modify that part of your workflow.
As much as you theoretically want to spend, since it's pay-per-use.
I spend $200/month by using Sonnet 4. Could be higher if you want to use Opus.
You can use Claude Code as a provider if you want it subscription based
https://docs.cline.bot/provider-config/claude-code
Devin is light years ahead of Cursor. It’s not even the same category!
I stopped writing code by hand almost entirely and my output (measured in landed PRs) has been 10x
And when I write code myself then it’s gnarly stuff and I want AI to get out of my way…so I just use Webstorm
>there are now better options.
Care to share your opinions on which options are better?
We still use Windsurf at work, but I've cancelled my personal subscription as (frankly) the code which Claude Sonnet 3.7 (thinking) generates is just not good enough, and much of the time its analysis of issues is also not correct. :(
That being said _sometimes_ its analysis is actually correct, so it's not a total miss. Just not something I'm willing to pay for when Ollama and free models exist.
Feels nothing like the same. The .com bubble was largely companies with no business, unchanged revenue but still having massive swings in price in private and public markets.
Cursor has a $500mm ARR your anecdote might be meaningful in the medium turn but so far growth as not slowed down.
> The .com bubble was largely companies with no business
Ah, yes, companies like Amazon.com, eBay, PayPal, Expedia, and Google. Never heard of those losers again. Not to mention those crazy kids at Kozmo foolishly thinking that people would want to have stuff delivered same-day.
The two lessons you should learn from the .com bubble are that the right idea won’t save you from bad execution, and that boom markets–especially when investors are hungry for big returns–can stay inflated longer than you think. You can be early to market, have a big share, and still end up like Netscape because Microsoft decided to take the money from under the couch cushions and destroy your revenue stream. That seems especially relevant for AI as long as model costs are high and nobody has a moat: even if you’re right on the market, if someone else can train users to expect subsidized low prices long enough you’ll run out of runway.
Thing is that it took 10-15 years for the stocks of these companies to reach the same marketcap again.
That’s what people predicted bit, for example, on Amazon’s case it was less than 3 years because they just kept posting solid numbers. The thing which all of those companies have in common is that they stood out from the Pets.com types in having profitable revenue - they didn’t need a miracle to be profitable, only for customers to keep buying.
You’re right that many .com companies lacked fundamentals but you’re cherry-picking survivors. For every Amazon, there were dozens of Pets.coms. The current AI wave does feel different in terms of revenue traction (e.g., Cursor’s $500M ARR), but the broader lesson still applies: hype cycles don’t discriminate between good and bad execution in the short term.
Cursor’s growth is impressive, but sustained dominance isn’t guaranteed. Distribution, margins, and defensibility still matter and we haven’t seen how durable any of that is once incentives tighten and infra costs stop being subsidized.
My point in listing survivors was simply to make the point that while there were plenty of doomed businesses, there were also many giants which were big at the time and could be told apart by looking at their fundamentals — they had real people paying them money for tangible things at a price which could be profitable. Amazon famously reported low numbers due to reinvestment but they were profitable in most business segments a few years after entering, which was quite different from the “lose money on every sale, make it up on volume” plays many dotcoms made.
How does that refute the statement you quoted? I said the vast majority of companies during the bubble had no business, were run on hype dollars and had insane P/E ratios. That supports a handful of companies making it through the bloodbath, but also a cherry picked examples that neither refutes my claim or supports yours.
You said “largely” and I think that’s painting with too broad a brush. The dotcom world included a bunch of companies which are still around (or were acquired later after surviving the collapse), and it wasn’t hard to tell who those were even at the time. There was a lot of lazy boosterism and criticism painting the whole field as the same, and that was a disservice to readers who could’ve used a more thoughtful triage approach. That’s especially the case for companies like Kozmo which actually had a popular idea and had the potential to be profitable (they were in most urban markets) but made the mistake of expanding too quickly or taking on more debt than they could service.
You’re sidestepping the core point. Of course some companies had fundamentals, even Kozmo had product market fit in a narrow sense. But the broader ecosystem was bloated with capital chasing flimsy ideas, and most dot-coms had no viable path to profit. That’s not “too broad a brush”, it’s backed by the collapse itself.
Kozmo is a great case study: decent demand, terrible unit economics, and zero pricing power. They didn’t just scale too fast, they scaled a structurally unprofitable model. There was no markup, thin margins, and they held inventory without enough throughput.
Many of these companies may fail but it’s a much different environment and the path to profitability is moving a lot quicker.
> The .com bubble was largely companies with no business, unchanged revenue but still having massive swings in price in private and public markets.
There also were companies like Sun and Cisco who had real, roaring business and lots of revenue that depended on loose start-up purse-strings, and VC exuberance...
Sun and Cisco both survived the .com bust, but were never the same, nor did theu ever reach their high-water marks again. They were shovel-sellers, much like Amazon and Nvidia in 2025.
Or yahoo- they were the premier sellers of ad space online (like google today) and made a lot of money from over-funded tech companies overpaying for online advertising during the boom years.
For sure but unlike then we are in a very different buying environment. Investors are more discerning even though folks here would like tot think differently. Cisco had something at peak like a 179x pe. That is a vastly different world than what we see Nvidia at today. I am not saying it cannot fail or collapse but to say this feels like the .com bubble is wrong.
>Feels nothing like the same. The .com bubble was largely companies with no business, unchanged revenue but still having massive swings in price in private and public markets.
I'm an attorney that got pitched the leading legal AI service and it was nothing but junk... so I'm not sure why you think that's different from what's going on right now.
Never heard Vincent to be leading in the legal space but I know vlex is more popular outside of the US so perhaps that’s why.
I am not sure why you would think your single anecdote is defensible or evidence to prove much. My perspective is valuations that are going on right now don’t have multiples that are that wild especially if we aren’t compare it to the com bubble.
>I am not sure why you would think your single anecdote is defensible or evidence to prove much.
Evidence? Prove? What are you talking about. This is just a discussion between people, not some courtroom melodrama you are making it out to be.
>My perspective is valuations that are going on right now don’t have multiples that are that wild especially if we aren’t compare it to the com bubble.
Okay, I could be equally rude to you, but I wont.
I responded in the exact same format as yourself and adding my additional thoughts. If that counts a being rude then were you being rude? I don’t believe in the US market Vincent is considered the best tool. That said I don’t believe a lot of the tools in the legal space that try to do everything are that powerful and the ones on my lens that are going well have focused on specific domains and problems. Even Harvey tried to do too much from the start.
As for valuations, when looking at current VC multiples and equity markets, I don’t see the same bubble from a qualitative perspective. Absolutely there is over hype coming from CEOs in public markets but there is a lot of value being driven. I don’t believe the giants are going to do well, maybe the infrastructure plays will but I think we will see a carve out of a new generation of companies driving the change. Unlike ‘99, I am seeing a lot more startups and products with closer to the ground roadmaps to profitability. In 99 so many were running off of hopes and dreams.
If you would actually like to converse I would love to see your perspective but if all you can be is mad please please don’t respond. Nobody is having a courtroom drama other than what’s playing out in your head.
Was the leading legal AI service Harvey.ai by any chance? I feel like big VC money goes to solving legal analysis, but I'm seeing a lot of wins with document drafting/templating.
Briefpoint.ai, casely.ai, eve.legal etc. I work with an attorney who trained his paralegals to use chatgpt + some of these drafting tools, says it's significantly faster than what they could've done previously.
It was Vincent.
> I feel like big VC money goes to solving legal analysis, but I'm seeing a lot of wins with document drafting/templating.
What do you mean "wins?" Like motions won with AI drafted papers? I'm skeptical.
>I work with an attorney who trained his paralegals to use chatgpt + some of these drafting tools, says it's significantly faster than what they could've done previously.
I'd be concerned about malpractice, personally. The case reviews I've seen from Vincent (which is ChatGPT + the entire federal docket) are shocking in how facially wrong they can be. It's one thing for an attorney to use ChatGPT when they do know the law and issues (hasn't seemed to help the various different partners getting sanctioned for filing AI drafted briefs) but to leave the filtering to a paralegal? That's insane, imo.
Not really much a of stuck bubble this time, though. Besides Nvidia and a handful of other HW companies, at least. Almost all of the very high valuations are for private companies and usually the amount of actual money involved in is relatively low.
I’ve always dwelled over $5 a month subscriptions for iPhone apps due to subscription fatigue. I find myself signing up for $200 AI subscriptions without a moments hesitation.
What do you do with $200/mo subscription to Anthropic? I’d consider myself a power user and I’ve never come close to a rate limit on the $20 subscription.
Depends a lot on the way people use them.
If you discusses a plan with CC well upfront, covering all integration points where things might go off rail, perhaps checkpoint the plan in a file then start a fresh CC session for coding, then CC is usually going to one shot a 2k-LoC feature uninterrupted, which is very token efficient.
If the plan is not crystal clear, people end up arguing with CC over this and that. Token usage will be bad.
Anecdotally, usage rises precipitously when you are building a system from scratch with unlimited ai access.
If you're using Claude Code with any regularity then the $200/m plan is better than a Costco membership in value.
I personally find gemini 2.5 pro and o4.1 mini to handle complexity better than claude code. i was a power user of claude code for a couple months but its bias to action repeatedly led me down the wrong path. what am i missing?
I'm finding myself agreeing with you... After also being a Max plan power user.
Now I just find myself exasperated at its choices and constant forgetfulness.
how do you integrate that with a code editor ?
I hope both of you know that you're in the extreme minority, right?
Are there available numbers to support this? Software engineering in the U.S. is well-compensated. $200/mo is a small amount to pay if it makes a big difference in productivity.
Which raises the question: If the productivity gains are realized by the employer, is the employer not paying this subscription?
My day job in talks to do that. I'm partly responsible for that decision, and i'm using my personal $200/m plan to test the idea.
My assessment so far is that it is well worth it, but only if you're invested in using the tool correctly. It can cause as much harm as it can increase productivity and i'm quite fearful of how we'll handle this at day-job.
I also think it's worth saying that imo, this is a very different fear than what drives "butts in seats" arguments. Ie i'm not worried that $Company will not get their value out of the Engineer and instead the bot will do the work for them. I'm concerned that Engineer will use the tool poorly and cause more work for reviewers having to deal with high LOC.
Reviews are difficult and "AI" provides a quick path to slop. I've found my $200 well worth it, but the #1 difficulty i've had is not getting features to work, but in getting the output to be scalable and maintainable code.
Sidenote, one of the things i've found most productive is deterministic tooling wrapping the LLM. Eg robust linters like Rust Clippy set to automatically run after Claude Code (via hooks) helps bend the LLM away from many bad patterns. It's far from perfect of course, but it's the thing i think we need most atm. Determinism around the spaghetti-chaos-monkeys.
Perceived productivity or actual productivity?
Yes, but that doesn't mean they aren't finding real value
The challenge with the bubble/not bubble framing is the question of long term value.
If the labs stopped spending money today, they would recoup their costs. Quickly.
There are possible risks (could prices go to zero because of a loss leader?), but I think anthropic and OpenAI are both sufficiently differentiated that they would be profitable/extremely successful companies by all accounts if they stopped spending today.
So the question is: at what point does any of this stop being true?
> I think anthropic and OpenAI are both sufficiently differentiated that they would be profitable/extremely successful companies by all accounts if they stopped spending today.
Maybe. But that would probably be temporary. The market is sufficiently dynamic that any advantages they have right now, probably isn't stable defensible longer term. Hence the need to keep spending. But what do I know? I'm not a VC.
A very productive minority.
Have we seen any examples of any of these companies turning a profit yet even at $200+/mo? My understanding is that most, if not all, are still deeply in the red. Please feel free to correct me (not sarcastic - being genuine).
If that is the case at some point the music is going to stop and they will either perish or they will have to crank up their subscription costs.
It's possible Anthropic is cash-flow positive now.
Claude 3.7 Sonnet supposedly cost "a few tens of millions of dollars"[1], and they recently hit $4B ARR[2].
Those numbers seem to give a fair bit of room for salaries, and it would be surprising if there wasn't a sustainable business in there.
[1] https://techcrunch.com/2025/02/25/anthropics-latest-flagship...
[2] https://www.theinformation.com/articles/anthropic-revenue-hi...
Cost to train and cost to operate are two very different things
I am absolutely benefitting from them subsidizing my usage to give me Claude Code at $200/month. However, even if they 10x the price its still going to be worth it for me personally.
I totally get that but that’s not really what I asked/am driving at. Though I certainly question how many people are willing to spend $2k/mo on this. I think it’s pretty hard for most folks to justify basically a mortgage for an AI tool.
My napkin math is that I can now accomplish 10x more in a day than I could even one year ago, which means I don't need to hire nearly as many engineers, and I still come out ahead.
I use claude code exclusively for the initial version of all new features, then I review and iterate. With the Max plan I can have many of these loops going concurrently in git worktrees. I even built a little script to make the workflow better: http://github.com/jarredkenny/cf
Again I understand and I don’t doubt you’re getting insane value out of it but if they believed people would spend $2000 a month for it they would be charging $2000 a month, not 1/10th of that, which is undoubtedly not generating a profit.
As I said above, I don’t think a single AI company is remotely in the black yet. They are driven by speculation and investment and they need to figure out real quick how they’re going to survive when that money dries up. People are not going to fork out 24k a year for these tools. I don’t think they’ll spend even $10k. People scoff at paying $70+ for internet, a thing we all use basically all the time.
I have found it rather odd that they have targeted individual consumers for the most part. These all seem like enterprise solutions that need to charge large sums and target large companies tbh. My guess is a lot of them think it will get cheaper and easier to provide the same level of service and that they won’t have to make such dramatic increases in their pricing. Time will tell, but I’m skeptical
> As I said above, I don’t think a single AI company is remotely in the black yet.
As I note above, Anthropic probably is in the black. $4B ARR, and spending less than $100M on training models.
It looks like their revenue has indeed increased dramatically this year but I can’t find anything saying they’re profitable, which I assume they’d be loudly proclaiming if it had happened. That being said looking at the charts in some of these articles it looks like they might pull it off! I need to look more closely at their pricing model, I wonder what they’re doing differently
Why would they want to be profitable? Genuine question.
Profit is for companies that don't have anything else to spend money on, not ones trying to grow.
I guess my genuine question in response is can you tell investors "Please give us billions of dollars - we never plan on being profitable, just endlessly growing and raising money from outside sources"? Unless the goal is to be sold off eventually that seems a bit like a hard sell.
I know very little about this. But isn't the inference cost the big one. Not the training?
> My napkin math is that I can now accomplish 10x more in a day than I could even one year ago, which means I don't need to hire nearly as many engineers, and I still come out ahead.
The only answer that matters is the one to the question "how much more are you making per month from your $200/m spend?"
In terms of revenue for my startup, plenty more.
I'm curious, how are you accounting this? Does the productivity improvement from Claude's product let you get your work done faster, which buys you more free time? Does it earn you additional income, presumably to the tune of somewhere north of $2k/month?
You would honestly pay 2k a month to an AI tool? Do you not have other costs like a mortgage or rent?
Are there studies to show those paying $200/month to openai/claude are more productive?
Anecdotally, I can take on and complete the side projects I've always wanted to do but didn't due to the large amounts of yak shaving or unfamiliarity with parts of the stack. It's the difference between "hey wouldn't it be cool to have a Monte Carlo simulator for retirement planning with multidimensional search for the safe withdrawal rate depending on savings rate, age of retirement, and other assumptions" and doing it in an afternoon with some prompts.
For curiosity, how complex are these side projects? My experience is that Claude Code can absolutely nail simple apps. But as the complexity increases it seems to lose its ability to work through things without having to burn tokens on constantly reminding it of the patterns it needs to follow. At the very least it diminishes the enjoyment of it.
It varies, but they're not necessarily very complex projects. The most complex project that I'm still working on is a Java swing UI to run multiple instances of Claude code in parallel with different chat histories and the ability to have them make progress in the background.
If you need to repeatedly remind it to do something though, you can store it in claude.md so that it is part of every chat. For example, in mine I have asked it to not invoke git commit but to review the git commit message with me before committing, since I usually need to change it.
There may be a maximum amount of complexity it can handle. I haven't reached that limit yet, but I can see how it could exist.
Simple apps are the majority of use-cases though - to me this feels like what programming/using a computer should have been all along: if I want to do something I’m curious about I just try with Claude whereas in the past I’d mostly be too lazy/tired to program after hours in my free time (even though my programming ability exceed Claude’s).
Well that's why I'm curious. I've been reading a lot of people talking about how the Max plan has 100x their productivity and they're getting a ton of value out of Claude Code. I too have had moments where Claude Code did amazing things for me. But I find myself in a bit of a valley of despair at the moment as I'm trying to force it to do things I'm finding out that it's not good at.
I'm just worried that I'm doing it wrong.
There are definitely things it can't do, and things it hilariously gets wrong.
I've found though that if you can steer it in the right direction it usually works out okay. It's not particularly good at design, but it's good at writing code, so one thing you can do is say write classes and some empty methods with // Todo Claude: implement, then ask it to implement the methods with Todo Claude in file foo. So this way you get the structure that you want, but without having to implement all the details.
What kind of things are you having issues with?
This has nothing to do with AI, but might help: All complex software programs are compositions of simpler programs.
I work at an Amazon subsidiary so I kinda have unlimited gpu budgets. I agree with siblings, I'm working on 5 side projects I have wanted to do as a framework lead for 7 years. I do them in my meetings. None of them are taking production traffic from customers, they're all nice to haves for developers. These tools have dropped the costs of building these tools massively. It's yet to be seen if they'll also make maintaining them the same, or spinning back up on them. But given AI built several of them in a few hours I'm less worried about that cost than I was a year ago (and not building them).
It's subjective, but the high monthly fee would suggest so. At the very least, they're getting an experience that those without are not.
The point is that if a minority is prepared to pay $200 per month, then what is the majority prepared to pay? I also don’t think this is such an extreme priority, I also know multiple people in real life with these kinds of selections.
>if a minority is prepared to pay $200 per month, then what is the majority prepared to pay?
Nothing. Most people will not pay for a chat bot unless forced to by cramming it into software that they already have to use
Forget chat bots, most people will not pay for Software, period.
This is _especially_ true for developers in general, which is very ironic considering how our livelihood is dependent on Software.
Yeah, cause we want to be in control of software, understandably. It's hard to charge for software users have full control of - except for donations. That's #1 reason for me to not use any gen AI at the moment - I'm keeping an eye on when (if) open-weight models become useful on consumer hardware though.
> Forget chat bots, most people will not pay for Software, period.
Apple says their App Store did $53B in "digital goods and services" the US alone last year. Thats not 100% software, but its definitely more than 0%
Games are a big exception here, as is anything in the app store.
But productivity software in general, only a few large companies seem to be able to get away with it. The Office Suite, CRM such as SalesForce.
In the graphics world, Maya and 3DS Max. Adobe has been holding on.
It's a generic chat LLM product, but ChatGPT now has over 20 million paid subscribers. https://www.theverge.com/openai/640894/chatgpt-has-hit-20-mi...
So $415m revenue per month, annualized $5 billion / yr. Let's say we use a revenue multiple of 4x, that means OpenAI should be valued at $20 billion USD just based on this. Then one obviously has several other factors, given the nature of OpenAI and future potential. Maybe 10x more.
Which puts the current valuations I've heard pretty much in the right ballpark. Crazy, but it could make sense.
A fool and his money are soon parted.
I'm easily getting $10K/month of value from my Anthropic subscription. (Rough estimate of how much I would have paid someone else to create the things I've (co)created with Claude Code so far.) If this is a bubble, I just hope I can finish all the projects I want to finish before it pops (or before they raise their prices to $9K/month because they read this comment.)
> I'm easily getting $10K/month of value from my Anthropic subscription.
Are those things created by Claude actually making you that much in real money every month? Because the amount of money it would cost to pay someone to create something, and the value that something brings to you once it's made are largely unrelated.
They are tools I want/need for my business (like creating software libraries for various things). My $10K number is how much I would have paid a contractor in the past to code it for me.
> I'm easily getting $10K/month of value from my Anthropic subscription. (Rough estimate of how much I would have paid someone else to create the things I've (co)created with Claude Code so far.)
I know it's hard to place a value on how much a utility saves a business, but honestly this math is like the piracy math and we didn't buy it back then either.
Some teenager downloading 20k songs does not mean that they saved $20k[1], nor does it mean that the record labels lost $20k.
In your case, the relevant question is "how much did your revenue increase by after you started 10x your utility code?"
[1] Assuming the songs are sold on the market for $1 each.
that seems like a silly way to think about it.
OP wanted a thing. in the past, they've been OK paying $10k for similar things. now they're paying $200/month + a bunch of their time wrangling it and they're also OK with that.
seems reasonable to consider that "$10k of value" in very rough terms which is of course how all value is measured.
> OP wanted a thing. in the past, they've been OK paying $10k for similar things.
Okay, then their costs should have come down similarly, no? OP said they were a business and that these weren't luxury hobby things but business needs. In which case, it must reflect on the bottom line.
I operate as a business myself (self-employed), and I can generally correlate purchases with the bottom line almost immediately for some things (Jetbrains, VPSes for self-hosted git, etc) and correlate it with other things in the near future (certifications, conferences, etc).
The idea that "here is something I recently started paying a non-trivial amount for but it does not reflect on the bottom line" is a new and alien concept to me.
When people made studio ghibli versions of themselves for free, were they creating hundreds of dollars worth of value since that's how much it would've cost a freelancer to commission such a picture? I would say rather the value of the pictures themselves became very cheap.
I have a feeling you'd have had better results if you actually paid $10k/month to a good dev or three in a LCoL geo.
You can actually hire a few excellent devs for very little money. You just can't hire 20k of them and convince them to move to a certain coastal peninsula with high rent and $20 shawarmas, for very little money each.
and people are still saying vibe coding is overrated? nonsense: https://www.lycee.ai/blog/why-vibe-coding-is-overrated
Or so you think..
[1] https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o...
[2] https://futurism.com/companies-fixing-ai-replacement-mistake...
Your first link is (in my opinion) highly biased in the samples they choose, they hired maintainers from open-source repos (people with multi years of experience, on their specific repo).
So indeed, IF you are in that case: Many years on the same project with multiple years experience then it is not usefull, otherwise it might be. This means it might be usefull for junior and for experienced devs who are switching projects. It is a tool like any other, indeed if you have a workflow that you optimized through years of usage it won't help.
> This means it might be usefull for junior and for experienced devs who are switching projects.
In other words: it might be useful for people who don't understand the generated code well enough to know that it's incorrect or unmaintainable.
Exactly. I think the study is a good reminder that we really have to be careful about the productivity gains attributed to AI. Main takeaway imo, despite limitations from the study, is AI is not a panacea, it can increase productivity, but only if used 'well' and with the good workflows in place, and in the right context.
> Or so you think.. > [1] https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o...
You are welcomed to your point of view, but for me while one agent is finding an obscure bug, I have another agent optimising or refactoring, while I am working on something else. Its hard to believe I am deluded in thinking I am spending more time on a task.
I think the research does highlight that training is important. I don't throws devs agents and expect them to be productive.
I mean, hacker news is still the same aren't they using AI to completely make this website more of whatever it was before ????
1. If you're maxing out your subscription, they're burning money on you.
2. They don't have a moat. DeepSeek and Kimi are already good enough to destroy any high margins they're hoping to generate from compute.
Just because something is highly useful doesn't mean it's highly profitable. Water is essential to life, but it's dirt cheap in most of the world. Same goes for food.
I agree. I was experimenting with tool use with Kimi K2 APIs yesterday - very effective, and so incredibly inexpensive. I am retired, now doing independent research, so my requirements are very different than most people here who are still in the job market or growing their own business.
I find a combination of local Ollama models, with very inexpensive APis like Moonshot’s Kimi with occasional Gemini 2.5 Pro use, and occasionally using gemini-cli provides extraordinary value. Am I missing out by not using one or more $200-$300 a month subscriptions? Probably but I don’t care.
growing ARR is easy when you are selling dollars for cents. people hyping ARR as an meaningful investment indicator are a dead giveaway that we are in fact in a bubble.
For sure, but then again - Nvidia $4T?! I can't shake the feeling though that with Nvidia we're looking at another Sun type of situation from _the bubble_. Remember the dot in dot com?
NVidia is being propped up by the US government. Huawei’s new chips are lower tech but would probably hit a good ‘practical sweet spot’ for AI data centers in many countries around the world but our current administration is threatening economic violence against any countries who choose to use more cost effective Huawei AI chips.
I don’t want to descend into talking politics, but I want to say that geopolitics, the rising geopolitical ‘south’, etc., is fascinating stuff - much more interesting and entertaining than anything fictional on Netflicks or HBO!
Then why didn't DeepSeek use the Huawei chips, and as opposed to the H800s ?
Economic bubbles _usually_ aren't based on things which are entirely worthless; there's generally _something_ there (just not enough something to sustain the valuations). There are exceptions (NFTs, arguably cryptocurrency as a whole, and of course tulips), but those _are_ the exceptions.
Why are you paying for that? Are you employed as a dev and paying out of your pocket or are you a hobbyist or ?
The big question is to what extent they hit a plateau and are commoditized. What happens when there is a fully open stack that gets Claude Code level results but at a fraction of the cost? Not saying that will happen, but that seems to be the scenario for a bubble bursting.
Microsoft is obviously the elephant in the room.
I decide to try out the agent built into VS Code. It basically matches most of these fly by night "agent" ides which are mostly just VS Code forks anyway.
But it's weird. Because Microsoft can use Anthropic's API, funnel them revenue and take a loss on Copilot.
We're all getting this stuff heavily subsidized by either VC money or big corp money.
Microsoft can eat billions in losses on this if they become *the* provider of choice.
This stuff isn't perfect, but this is the worst it'll ever be. In 2 years it'll be able to replace many of us.
They're still gonna be an estimated $3 billion in the hole though. Jury still out of there is really "something there".
Genuinely curious for the value add with Claude code here. Some perspective and/or data is appreciated.
> They're getting my $200 a month and it's easily the best money I spend
Can you explain? I don't see how $200 makes that much difference than what I get from paying $20/month with OpenAI? What's the use case?
I pay $0 and that's already enough for me. Genuinely, what are you getting for your $200? I cannot fathom paying that much for what seems like I get basically for free anyway.
It's so funny to see people in this industry get super excited about revenue while never mentioning costs.
The bubble isn't the value available to buyers, it's the x00% multiplier created by speculation.
Yeah, the difference between the "something there" and the $4b valuation is the bubble.
Unless they are paying annually, the next big thing could see those $200 a month premium users gone.
Agree. But I think Anthropic is the outlier. Maybe ElevenLabs, too.
So how much is it worth? what if it balloons to 5x the amount? what if "they" just decide this is too powerful so only corporate accounts worth $xx,000 a month, and plebs get limited per month for $x,000?
Remember that the web also had a bubble that popped and look at where are we now with Google, Amazon, Meta...
I think that there is a bubble but it's shaped more like the web bubble and less like the crypto bubble.
As with any investing there's a risk appetite/timescale component to thinking about this stuff. Lots of companies went to zero in the dot-com bubble. Even Amazon was down over 90% between the end of 1999 and late 2001, and took until 2007 to recover to its high. NASDAQ overall took 15 years to return to its March 2000 high. Some incredible returns to be had if you waited it all out, to be sure, but it's hard to know what the interim looks like.
It's taken Cisco 25 years to recover
Intel never recovered. Well they did if you count dividends but still..
Yeah, only those evolved a lot from the initial products everyone hyped and products people hyped in 2000 are extinct or free. And I still don't understand where Facebook makes money. :)
Regarding LLMs there are two concerns - current products don't have any killer feature to lock in customers, so people can easily jump ship. And diminishing returns, if there won't be a clear progress with models, then free/small, maybe even local models will fill most of people needs.
People are speculating that even OAI is burning more money than they make, it's hard to say what will happen if customer churn will increase. Like for example me - I never paid for LLMs specifically, and didn't use them in any major way, but I used free Claude for testing how it works, maybe incorporating in the workflow. I may transitioned to the paid tier in the future. But recently someone noted that Google cloud storage includes "free" Gemini Pro and I've switched to it, because why not, I'm already paying for the storage part. And there was nothing keeping me with Anthropic. Actually that name alone is revolting imo. I wrote this as an example that when monsters like Google or Microsoft or Apple would start bundling their solutions (and advertise them properly, unlike Google), then specialized companies including OAI will feel very very bad, with their insane expenses and investments.
> And I still don't understand where Facebook makes money. :)
If that's a genuine question: Facebooks sells ads, information and influence (eg. to political parties). It's a very profitable enterprise. In 2024 Meta made $164B in revenue, and they're still growing at ~16% year-over-year.
[0] https://investor.atmeta.com/investor-news/press-release-deta...
> And I still don't understand where Facebook makes money. :)
Its Meta now, and they own alot of "brands" besides Facebook. Instagram, WhatsApp, Oculus, Giphy, etc.
You don’t understand how the world’s 5th largest company by market cap makes money and this is evidence of… something?
That was a joke, mostly unrelated to the main point - about LLM corporations' finances.
“Web” is such a broad category. Quite a leap from LLM wrappers.
Well, LLMs are themselves very broad. They encompass everything from web search to everything that you could automate yourself but don't have the time.
I don't LLM capacities have to reach human-equivalent for their uses to multiply for years to come.
I don't LLM technology as it exists can reach AGI by the simple addition of more compute power and moreover, I don't think adding computer necessarily is going to provide proportionate benefit (indeed, someone pointed-out that the current talent race acknowledges that brute-force has likely had it's day and some other "magic" is needed. Unlike brute-force, technical advances can't be summoned at will).
"Brute force" is only held back by economics and hardware limitations.
There are still massive gains to be had from scaling up - but frontier training runs have converged on "about the largest model that we can fit into our existing hardware for training and inference". Going bigger than that comes with non-linear cost increases. The next generations of AI hardware are expected to push that envelope.
The reason why major AI companies prioritize things like reasoning modes and RLVR over scaling the base models up is that reasoning and RLVR give real world performance gains cheaper and faster. Once scaling up becomes cheaper, or once the gains you can squeeze out of RLVR deplete, they'll get back to scaling up once again.
> Well, LLMs are themselves very broad.
I think overstating their broad-ness is core to the hype-cycle going on. Everyone wants to believe—or wants a buyer to believe—that a machine which can grow documents about X is just as good (and reliable) as actually creating X.
Since APIs exist (including APIs that allow you to get a human to execute the task), a machine which can grow specs about X is able to create real-world X in many, many situations.
A machine which can define a valid CAD document can get the actual product built (even if the building requires manual assembly).
LLMs don't seem very broad. You're literally just talking about predictive text engines.
The dot-com was a bubble because investors pulled money and belief at the first sign of trouble.
The landscape has changed dramatically now. Investors and VCs have learnt if we stick with winners and growth companies, the payoffs are massive.
We also have more automatic, retail and foreign money flowing into the market. Buy the dip is a phenomenon that didn't exist at the scale it is now.
Pre-2015 if Big Money pulled out, the market was guaranteed to fail, but now retailers sometimes have longer views and belief (on people like Musk, Altman) than institutions and they continue to prop it.
So, it's foolish to apply 2000 parallels to now. Yes, history repeats, but doesn't with the exact time or price points
Also there are no early IPOs. Very few people can buy stocks in these companies which changes the dynamics significantly. Note sure what's the point of talking about the stock market this much when for almost everyone the only way to get any exposureis through Nvidia or other hardware companies and maybe MS/Google/AWS.
> Investors and VCs have learnt if we stick with winners and growth companies, the payoffs are massive.
Well... yes and no. 2021 wasn't that long ago.
> So, it's foolish to apply 2000 parallels to now
The stock market and other financial stuff is of course different. The fundamental trend not necessarily though. It took awhile for anyone to figure out how to directly build a highly profitable internet based business back then for AI it seems more or less the same so far.
dot-com bubble companies were not good companies. They either built something that was not novel so it could be copied, or had insufficient value to monetize. We'll see the same with current AI.
Similar to the invention of the web, AI is not a bubble. Real value has been created.
Cisco was the quintessential dot-com bubble company. Back then, it was what Nvidia is today: at the very spearhead of investors rallying behind the Internet.
"Good company" is subjective, but to argue that the company that built the backbone of modern web didn't make anything novel or monetizable is a bit short-sighted, don't you find?
And if you'd invested in Cisco then, it would have gone very badly for you. It was _wildly_ overvalued; at peak it had a P/E ratio of ~200 (even Nvidia's only about 50).
how so? Many internet companies that went bust like Webvan and Pets.com have successful equivalents in Instacart and Chewy?
I'm pretty sure many internet companies would be given a longer rope to survive now. E.g OpenAI and Anthropic will probably take years to get profitable but investors are OK with it
> not novel so it could be copied
AI Agents can't be copied in a race to the bottom market to resell inference compute?
> Investors and VCs have learnt
lol. Investors and VCs have no idea what they're doing
lol is a coping mechanism for the poor. If you really think top VCs / investors haven't learnt the long-term importance of staying the course, then you know nothing about the industry and mostly being influenced by popular social media posts shitting on the investor class.
There is a reason Anthropic/OpenAI and many startups are given much much longer ropes to be profitable than in the 2000 era when VCs pulled the rug the first opportunity of trouble
The thing that was most disillusioning for me here was SVB -- failure to apply basic principles of banking (i.e. they never really had a plan for central bank interest rates to change more than +-1%). Not just that the VC types running a bank weren't able to do so, but that such a large number of tech companies held all their cash on hand in a bank account (and didn't deposit anything in another bank, or a money-market fund/t-bills).
If VCs knew what they were doing, they'd have real jobs
Allocating capital might be the "realest" job in capital...ism.
There are always shitty 20% operators in every industry. They won't make money and get weeded out.
Delusional to apply this to top operators (and at the same breath complain about Rich getting Richer)
I have yet to be pleasantly surprised by the alleged collective wisdom of Wallstreet. I would hope that you are right, and that our corporate masters are smarter than I give them credit for, but I'm not going to hold my breath
It is foolish to compare to the dot com boom and bust. At least when that bubble burst we still had the global broadband internet that it built. When this bubble bursts, we'll have next to nothing to show for it.
Nothing except massive data centers full of GPU compute resources paid for by VC money. Wait, that's actually pretty similar...
I'm starting to think that making a bunch of tech companies the most valuable companies on Earth, and tying their value to everyone's ability to retire so the number must always go up was perhaps not the wisest thing to have done.
They could close shop and you could print the money and give to the retirements fonds and everyone would be better off. Maybe Apple would be missed.
gpus go obsolete faster than fiber backbone equipment
Do they? I figured each speed increase requires new optical equipment, but I guess I was just making assumptions.
We have AI, a marvel that might change the arc of humanity and an epoch in our timeline. Fire, wheel etc. and AI.
I'll choose the wheel over using a country's worth of electricity to parrot unusable AI slop to gullible fools.
Is AI not useful to you? I've sped up my SWE work significantly (10x). Not sure why the cynicism.
If you're just talking about SWE work, thats only one segment of an economy and is the "virtual world". But humans have to live in the real world.
I believe the true revolution is going to be when AI can start living / interacting with the physical world. Driverless cars might be the start here.
We will have a mountain of GPUs!
OP didn't reference the dot-com bubble though...
Of course we are in a massive bubble. At least in the US where I live, everything is about driving short term profits, with less than a little thought to what is good for the non-special interests/insider class.
I have lived and worked through two previous ‘AI winters’ and I expect the current bubble to eventually pop in a dramatic way. There will be good things produced by AI, but I am skeptical of the panic FOMO rush to AGI or super intelligence.
Look at the process of shifting manufacturing out of the USA: that was all about driving extreme wealth for special interest insiders. Sadly, most people look to their particular little political party for some form of relief - how is that working out?
I've begun to think some bubbles are good for the economy overall. In the dotcom days anyone with an idea and a domain name could get funding. I myself worked for a company that nabbed 7x more funding than needed but still failed due to poor leadership. I had reservations about the founder but thought I could help drive things, but he was even more absent than I ever anticipated.
A lot of VCs and PEs lost a lot of money during the crash. This means a lot of capital was spent in the economy, generating a lot of good activity, and the companies that failed then also put a lot more capital back into the economy through bankruptcies. Other businesses can pick up talent, IP, and assets for cheap, and everyone can learn from the failures. While losing that money isn't great for VCs, what they got was a very valuable education to be better stewards of their investments, and pick better companies. The next rounds of companies have to hit metrics, milestones, have to prove their value, etc.
Never waste a perfectly good crisis: learn if nothing else.
Good point, it transfers capital from the investor class to the working class.
The problem you don't know which innings we are on and it may very well be the 4th/5th. More often than not you lose more by not participating in the run up and waiting for the collapse. I'd rather ride the wave (as intelligently as I can) than be on the sidelines just so I can say I told ya so.
Not in a bubble yet. Wait till your aunt takes out a second mortgage to chase NVDA stock to the moon. My guess is this will continue until later this decade with some bumps in between. That said, it’s absurd these guys are paying people so much money for what I think amounts to being context management off of some markdowns.
Bubble or not these kinds of deals could put a chill on the tech startup sector as well.
Things move fast in tech because there isn't nearly as much red tape and litigation as there are in other mature industries. This is because there's an agreed 'way of doing things'. Take funding, grow like crazy, sell/merge or IPO. Everyone wins or loses together (even if things are stacked in favour for some over others).
Once trust in this process is broken and founders or VCs start stacking the deck in their favour the game becomes rigged to the point where other people don't want to play anymore. Once that trust is gone red tape and litigation appears.
Yes and no. Are AI companies overvalued? Yes. Will most of them crash and burn? Yes, they will. Are the words "intelligence" or "reasoning" misused? Absolutely. Nonetheless, nobody can deny that some of these tools are useful and have demonstrated they can generate revenue like no other tool, app, or device before. There is something different from the dot-com bubble; many barriers have come down since 2000. Everyone can be connected 24/7 for a few tens of dollars a month. People trust the internet as a medium to perform transactions and access data. The real bubble is in the private market valuations, especially in the pre-seed stage. Many young entrepreneurs don't understand that raising their first round at $30M, $50M, or even $100M post-money will put a heavy weight on their shoulders approaching the Series A. Raising a funding round is a promise you make to the market. Increasingly high expectations will burn many wanna be entrepreneurs whose contribution to make things better for everyone will be lost forever. I'm deeply convinced that the reality check for all those companies is the public market, and in today's world, you can't go public after 3-5 years if your initial valuation has been built on a 20-year-long promise. All the trillion-dollar companies we know today went public a few years after their creation: Apple, Google, Microsoft, NVIDIA, etc. OpenAI, Anthropic, and Cursor are black swans, not evidence of the power law.
No, MS went public in 1986, so more than a decade there.
> the possibility that we’re in a massive bubble thats quite divorced from any kind of fundamentals
I'm astounded that anyone still genuinely believes we're not in a massive bubble. Of course AI company CEOs are going to say we're not and that AGI is just around the corner, it's deeply in their financial interest to keep inflating the bubble as long as possible.
Yeah, I wonder if there are parallels to server-less tools of a few years ago.
it’s not bubble so much as urgency, the trillion dollar club are pricing in the risk of losing their position
How about that Anthropic revenue though.
Recent and related:
OpenAI’s Windsurf deal is off, and Windsurf’s CEO is going to Google - https://news.ycombinator.com/item?id=44536988 - July 2025 (679 comments))
Attended Windsurf's Build Night 18 hours before founders joined Google DeepMind - https://news.ycombinator.com/item?id=44539884 - July 2025 (1 comment)
Unpopular, controversial take: there should be an LSP extension that lets CLI agents, like Claude Code show diffs in the editor, and also one for completions, and sending snippets back to the CLI.
That, by itself, would obliterate the entire value of Windsurf or Cursor or whatever. The fact that Google has this kind of money and spends it on dubious "talent" (though none of these people are known in the community) is a testament to how overfunded tech companies are compared to the value that they provide.
This already exists via MCP
> For other IDEs: The protocol is editor-agnostic. Any editor that can run a WebSocket server and implement the MCP tools can integrate with Claude Code.
https://github.com/anthropics/claude-code/issues/1234 https://github.com/coder/claudecode.nvim/blob/da78309eaa2ca2...
Example in Emacs, this is how I use claude-code: https://github.com/manzaltu/claude-code-ide.el
This is great!
Claude Code can already show diffs in JetBrains IDEs and VSCode ('/ide' command connects the CLI/TUI to plugin/extension running in the IDE-side).
It can also access the IDEs' real-time errors and warnings, not just compile output ('ideDiagnostics' tool), see your active editor selection, cursor position, etc.
"overfunded" is a weird way to talk about a tech company that is incredibly profitable.
It is known there were many developers without really much work to do that were hired only to be denied to competitors. Maybe it was cleaned up in the meantime
Big companies hired a lot, but I don’t think this specifically is true? In theory a high-value engineer would be productive, or else they aren’t worth stealing.
The simpler explanation seems more correct here — there was a lot of product fluff and a lot of headcount allocated to build that fluff.
Revenue is a funding source. Companies with too much money sometimes made really dumb decisions with that money.
The fact that one division of Google is wildly profitable does not exempt other parts of the company from criticism of their financially dubious choices.
Net profit can't be used as a measure of both "funding" and "value generation" while saying that a company is "overfunded" because it doesn't provide enough "value". Come back to your senses.
I'm genuinely not sure if you're not understanding, or deliberately ignoring my point.
I'll assume the former and try again. Maybe you didn't realize I'm not the person you originally replied to?
If a company is profitable, they have funds. The funds generated by the profits, can be used to fund additional internal projects. If the bucket of funds from profits gets ridiculously large, then it may begin to be used for vanity projects, like gutting an AI company, or building a gold statue of the founder. It seems reasonable to call companies spending on mostly-useless excesses "overfunded".
It's kind of their prerogative to spend their money how they like. If it's anti-competitive then you can hope that regulation exists to prevent that behaviour. As a shareholder you can complain that they will stop this value igniting behaviour. As an investor if you believe this behaviour is irrational you can short their stock and hope that the market is efficient and the share price will reflect the igniting value in their multiple.
As a bystander and outsider it is hard to isolate the value igniting behaviour from the moonshot behaviour. Shareholders love to gut a business of its risk taking and excess behaviour for predictable and inflated margins (and dividends) but the past 20+ years of our megacap companies is that they have continued to "innovate" in spite of all their inefficiencies.
I always have a chuckle when I recall how shareholders tried to oust Zuck for buying Instagram for 1B...
These vanity hires do seem frothy and reminiscent of dotcom style behaviour. But "AI" clearly will be game changing much like the internet was, and who at this stage can say what is worth recruiting people at the forefront of commercialising the tech right now
I'm not misunderstanding anything. It's just that the idea of a massively profitable company being labeled "over-funded" when excess profit is used in a way that you find disagreeable is stupid. It doesn't matter how well they allocate their revenue.
Stop looking at the entire world through the eyes of VC, because it doesn't work.
You're not funding google by paying for youtube, you're buying a service.
You didn't "overfund" your pizza shop that hired a stripper for friday night vibes, and neither did 99.99% of customers that paid google for their service offerings.
You just bought a pizza. Put down the VC podcasts
Gotcha, you're just being mean. You've got a bone to pick with VCs, I guess, so you've declared I listen to VC podcasts, see the world like one, and am using a stupid idiom.
That's fine. I'm going to continue referring to corporations that blow lots of money on random intra-industry dick measuring matches because they can as "overfunded", and you can continue expressing your opinion to anyone who will listen that this one person on the internet used an idiom that you think is dumb because it implies something other than "that person profited, therefore they did something right and therefore whatever they do with that money is correct and intelligent and never ever wasteful or dumb."
It's fair to suggest a different word or phrasing, but you're coming off as hostile, not constructive.
Windsurf is the first company that moved from IDE to training and trained good models. Unfortunately in code good models are not enough to win against claude 4. Any wrapper constraining the model is also doomed to fail.
> there should be an LSP extension that lets CLI agents, like Claude Code show diffs in the editor, and also one for completions, and sending snippets back to the CLI.
aren't you reviewing diffs in whatever diff tool you like? I find magit to be superlative for this (and for correcting and committing things).
There is MCP-LSP and Context7.
I use Rust and found it's better to let the AI hallucinate function names, then let the compiler correct them. Rust's compiler is significantly better than TypeScript's at this, so it works well.
There should be an LSP integration that lets Cursor look up symbols and definitions the same way I do, instead of trying and failing to use grep / its fuzzy vector index thing!
Except the Windsurf team is already moving in that direction...
Unpopular, controversial take...
Could you please avoid juicing a random comment this way?
> That, by itself, would obliterate the entire value of Windsurf or Cursor or whatever.
Have used Cursor and I know that there is quite a bit of value between the model and the chat input box and it will be similar to Claude Code or Codex, it's what makes this agentic, it's just accessed through a different interface. So from that perspective, Cursor makes sense for folks that are already in the VSCode environment.
On the contrary, I'm already in the VSCode environment and I would hate to have to switch to Cursor. Doesn't matter that it is a fork, I don't want a fork for every extension of my workflow. I used Cline and now Claude code integrated to VSCode, and I don't think I'm missing much by not switching to Cursor.
I symlinked my vscode settings into the cursor config directory and the transition was strictly additive; everything worked exactly the same as before, with two exceptions so far: a .NET extension wasn't supported, and Cursor recently forced a switch to their own version of Pyright (but that seemed like an improvement once the type checker diagnostics were tamed a little).
> Cursor recently forced a switch to their own version of Pyright
Microsoft forced Cursor to stop using their versions of various plugins https://forum.cursor.com/t/the-c-dev-kit-extension/76226/18
I don't see a justification for high valuations of companies that aim to build an "AI Software engineer". If something like Devin really succeeds, then anyone can use their product to simply build their own competing AI engineer. There's no moat, it's just another LLM wrapper SaaS.
Yep. The reality is folks building these types of companies are trying to get acquired as quickly as possible before the house of cards fall. This has led to a huge speculative rush of acquisitions to avoid FOMO later.
The technology is nowhere close to what they're hoping for and incremental progress isn't getting us there.
If we get true AGI agents, anyone can also build a multi-billion dollar tech companies on the cheap.
> If we get true AGI agents, anyone can also build a multi-billion dollar tech companies on the cheap.
That's not how the economy works...
You're right - AGI would be unfathomable, it would be more productive than a quadrillion earths entirely populated by MIT valedictorians who just drank 2 espressos each. "Multi-billion dollar" would be a silly valuation.
I can't tell if you're joking or serious.
AGI is a distant dream in sci fi novels. Don't confuse it with today's AI technology.
I don't see this. The ai software engineer that succeeds, maybe it's because of a mixture of very complicated architecture derived from novel research etc. You can't replicate that with just hiring more human engineers, it takes time and effort and elite hiring. Plus enterprise support etc.
Devin etc will give you let's say 10x more engineering power, but not necessarily elite one.
i advise you to not take marketing lines too literally and be so casually dismissive as a result. you will miss a lot of good investments and startups this way and (worse) be lulled into a false sense of comfort and security.
This is true for LLMs themselves. If a new LLM is really better than all the other ones then it can be used to help improve other LLMs.
Is it? Last I checked when you trained an LLM on another's output, at best you got the same performance as the original, and it was more likely you significantly degraded usefulness. (I'm not talking about distillation, where that tradeoff is known in return for a smaller, more efficient parameter set)
This is my exact takeaway too, and I'm always surprised it doesn't get mentioned often. If AI is truly groundbreaking, then shouldn't AI be able to re-implement itself? Which, to me, would imply that every AI company is not only full of software devs cannibalizing themselves, but the companies themselves also are.
This is my watershed for true AGI. It should be able to create a smarter version of itself.
Last I checked, feeding the output of an LLM back into its training data leads to a progressively worse LLM. (Note I'm not talking about distillation, which involves training a smaller model, by sacrificing accuracy. I'm referring to an equal or greater number of model parameters)
If the LLM is given the code for its training and is able to improve that, does that count? Because it seems like a safe bet that we're already there, the only problem is latency of training runs.
There are any number of tools that already make that promise. Turns out it’s still hard to complete projects and bring them to market.
I was a bit confused as to what "Cognition" was, but they're the makers of Devin (edit: that just got added to the title, for reference), so that makes sense. Just buying the competition, the only surprise is they had more money to spend than the big ones.
Well Google did also just pay $2.5B to license Windsurf in perpetuity. Cognition is probably spending a lot less than that for just whatever it left after that type of a deal. Remaining team members, etc.
This looks to me like the smoking gun on a type of acquisition that circumvents regulatory oversight, primarily driven by the "need for speed":
https://medium.com/@villispeaks/the-blitzhire-acquisition-e3...
which I first saw here
https://news.ycombinator.com/item?id=44553257
Circumvents regulatory oversight and also shafts 99% of the employees. Seems to be a backdoor way to acquire the key founders/leaders and IP (via a perpetual license) while leaving behind a desiccated husk of rank and file employees, customers, and obligations.
it's super shitty by management to flee, but given that most startups fail, this startup maybe not-failing really isn't a "shafting".
> The acquisition includes Windsurf’s IP, product, trademark and brand, and strong business.
So, Google will be paying $2.5B to Devin guys?
No, as some portion of the $2.5b goes to Windsurf investors.
Basically, Google bought the top talent from the company. This cash was used (according to articles I read this morning) in part to pay directly out to shareholders, and in exchange Google got the top talent from the company and a license for the software (probably mostly so their new talent didn't have to worry about NDA, non-compete, and patent challenges).
Since this money went to shareholders, not to the company bank, and since top talent fleeing the company reduces the value of the company the overall value of Windsurf likely went down as part of the Google deal. This in turn likely made it cheap enough for the remainder to be purchased by Cognition.
> Google did also just pay $2.5B to license Windsurf in perpetuity
Could there have been a clause that made this invalid in case of acquisition?
IANAL but you'd have to be pretty dumb to include that clause.
Can the new buyers revoke that license?
> the only surprise is they had more money to spend than the big ones
The sale price for Windsurf was likely significantly lower than the original acquisition plans.
It didn't go to $0 like some predicted, but it was never going to be as valuable as it was before the executives bailed on it.
kind of funny that no one seems to know them by name, only by the infamously panned reception of their main product
One benefit of separating your brands from your company is you can try again without the stigma of your failures :)
As far as I recall they were first to market with the "AI software engineer" promise.
“Had more money to spend” => it could be a little money and a large amount of stock.
Had it been the reverse, they would have announced the purchase price.
Let this be a learning lesson in judging these deals based on partial information. Kudos to the Google, Windsurf, and Cognition teams for keeping all of these deals under wraps until announcement (OpenAI could learn something...), but even so it's likely that we the public will never learn every detail of what transpired. I've seen a lot of harsh, misguided takes over the past few days, like that the Windsurf founders screwed over their employees, or that OpenAI reneged on the deal. In this case, this seems like a happy ending for all parties involved: congrats to the Windsurf team!
> I've seen a lot of harsh, misguided takes over the past few days, like that the Windsurf founders screwed over their employees [...] In this case, this seems like a happy ending for all parties involved
There is no evidence at all in the announcement that is the case. It just says "100% of Windsurf employees will participate financially in this deal". What "participate financially" looks like is not elaborated upon.
It is possible you're right. It's also equally possible that the founders have still screwed over their employees, we just don't know. Nothing in this post supports either position.
>It is possible you're right. It's also equally possible that the founders have still screwed over their employees, we just don't know. Nothing in this post supports either position.
In the lack of evidence, its okay to assume the most likely scenario, which is the executives & shareholders will make out like bandits and everyone else is likely to at best, get pennies.
pizza party
If my understanding is correct, this is still a much worse deal for employees than if Windsurf's exec team had negotiated a "standard" "accelerated vesting, common conversion" acquisition with Google.
Presumably the "payout" from Cognition is at a lower nominal value and in illiquid (and IMO overvalued) shares in Cognition rather than cash.
You're talking like the founders orchestrated this deal to cognition all along.
that's absolutely not the case. they ejected and the remaining executive team dealt with the sale over the weekend.
I can't imagine this deal was done over just the weekend. It takes time to set this up, check the financials etc.
> OpenAI could learn something...
Did OpenAI ever actually announce anything publicly regarding a potential windsurf acquisition?
AFAICT most of the reporting was based on rumors or leaks. But they never actually announced an acquisition. Seems like Bloomberg may have made an oopsie here.
> Kudos to the Google, Windsurf, and Cognition teams for keeping all of these deals under wraps until announcement
Geez is the cognitive distortion field active again? Even Grok could figure this one out.
Wait, so Google picks up the talent, and Devin picks up the brand/product? This is so confusing!
I think it's safe to say don't use Windsurf. There are so many other options.
Unfortunately this seems to be true. I like Windsurf, but these days I just use it as a harness for running Claude Code while still retaining decent code completions.
Try Gemini CLI too. Free and I like it more than Claude Code.
This is the final fix. I'm confident this will resolve all the errors.
[API Error: got status: INTERNAL. {"error":{"code":500,"message":"An internal error has occurred. Please retry or report in https://developers.generativeai.google/guide/troubleshooting..."}}]```
How do you do the nice line-by-line (or section-by-section) before/after handling in a CLI tool?
Start with a clean git (even if you're just committing WIP locally). Run the tool. See the diffs in your IDE. They show up immediately in IntelliJ; use the "commit" pane. Same as if I had edited a bunch of files.
Works fine.
It has an external editor option (which I havent tried) the cli interface by default is pretty bad for medium to large code changes.
The talent stays with Devin. Google has got the CEO. Not sure why they need a CEO, maybe Pichai wants to retire
Is this purely the rump company left over from the Google pseudo-acquistion? Or does this mean that deal fell through?
Does this represent confirmation that there was no pro-rata compensation to common share holders in the Google deal?
I just have so many questions.
Windsurf founding team is already at Google
And what is Google "paying $2.4 billion in license fees" to use Windsurf's technology for, and to who? Does Windsurf have any technology?
I suspect it is more a "licence" for the things in the staff's heads for those who were poached.
And this highlights why intellectual property is really gross to me
not sure of windsurf's ToS, but Google could be after the usage data (e.g. did the user accept the suggested edit? if not, what did they change it to? etc.)
Windsurf is the soham parekh of acquisitions at this point. Is there anyone not acquiring windsurf?
so devin gets the leftover remains of windsurf to fix their agentic AI ide that wasn't working in the first place?
yes.
So some VCs got rinsed for Google’s leftovers.
VC's probably did fine. Their preferred shares get paid out on the dividend out of the Google 'license' fee. They might get some stock in Cognition as part of the acquisition of the remains of the company, but they've likely already been made whole and made a tidy profit on money invested in early rounds.
ok so it's not all bad then
Must have been acquired for an extreme discount far from the $3B offer from OpenAI and Windsurf's (alleged) valuation of $1.3BN.
Cognition being worth $4B with Devin being raced to zero by Claude Code also undercutting both Windsurf and Cursor have a very steep hill to climb.
Having both Devin and Windsurf will just make them raise more money as they burn through their operational costs.
Google already paid $2.4B out of the $3B OpenAI deal, to what I imagine is Cognition now. Cognition must have spent low 100s of Millions max.
> Google already paid $2.4B
This is unclear. $2.4B was for licensing and compensation. Why would Google have agreed to pay any significant amount to the Windsurf leftovers?
I’m really confused now. Also, is there really that much of a transformative difference between Windsurf and say OpenAI/Claude etc to warrant this crazy valuations?
I'm also confused why Devin is worth billions
For real. According to the marketing material. Couldn’t Devon just whip up a windsurf competitor in minutes?
They are hiring software engineers! Just look at the careers page! hahahaha
Devin's whole business model is predicated upon replacing programmers with minimum wage AI agents.
That the generally unspoken business model for all ai products. How can we get rid of these annoying humans.
Reminds me of of some quip where a doctor says to a resident. Yeah this job would be so cool if it wasn't for all these sick people.
That'll be cool only if Devin actually worked.
the founder went to harvard, is basically the answer
Who decides for much a company is worth?
> crazy valuations
I haven't seen anything to indicate what was paid for what's left of Windsurf.
I was also referring to the $2.5B Google paid. I can’t imagine what could possibly be the value they’re hoping to get
I'd love to learn more about that arrangement. Maybe Google's terms indicated that the "licensing fees" portion of "$2.4 billion in licensing fees and for compensation" are void if Windsurf gets acquired.
They got the top talent from there with the technology and they can continue on that or be used for other AI projects, right?
So the nobility goes to Google and rank and file engineers are subject to consolidation, am I reading this right?
"Consolidation" aka "we got the IP we wanted, you are all now redundant"
The worst AI coding tool I've tried aquiring the best is interesting.
Isn't Devin AI basically a scam, selling an "AI Software Engineer" when no such thing exists.
We're using it consistently to put out (albeit smaller) features.
Any examples? Curious what you let it rip on and what it can actually do.
I just started trying out Codeium in Nvim on a recommendation from a video that's like, less than 6 months old. Oops, actually, codeium is Windsurf now. But the tool is still codeium.nvim at least.
I guess it's about to happen again!
It's weird being a Windsurf fan on HN. You get the feeling that you're in the minority or doing it wrong because the dialog is dominated by Cursor and (more recently) Claude Code.
It's also been a lot of random stuff recently with their 3 separate Ross and Rachel acquisition storylines.
Some takeaways:
1. Devin/Cognition definitely have a legit AI dev agent now
2. It's crazy what Google passed on. The fact that it was worth it to them without the traditional best assets is wild. Guess that's what happens when you play on ultra hard mode with an infinite money glitch.
3. I am worried/pre-emptively sad that Windsurf will likely go away or get nerfed, more expensive etc.
https://www.reddit.com/r/ChatGPTCoding/comments/1m05ar4/it_s...
"If the founder/CEO and best engineers at Windsurf don't believe in that product/company, why should you?"
^ Any company that competes with them could say that and it would create some pause.
Edit: nvm, see replies
"Acquiring" is often synonymous with "merger". Technically the surviving company acquires the other, but effectively it's a merger. Not always, but it's not uncommon to see two weaker companies merge in a competitive landscape to survive. It also counts as a liquidity event allowing employees some financial levers.
But these seem to be two well-funded companies? I.e. both with $100M+ funding within the last 12 months
> How does this transaction make sense?
The fact that it doesn't make sense with those numbers almost surely indicates those numbers are misleading.
> Google paid a $2.4 billion licensing fee
This is the reported number for licensing and compensation, but who knows what the terms really were.
> Cognition’s valuation is $4 billion
Doubtful
They probably got what got left of it in a cashless deal. Basically, the shareholders got to exchange X shares in a fatally wounded company into Y shares in a still-alive startup. The economic sense depends on the ratio between X and Y, but if the board was close to panicking due to recent events, Cognition probably got a good deal.
I think google is buying windsurf, they are leaving 100% of assets and some employees, and windsurf is buying that 100m in assets + taking the leftover employee liability.
Microsoft poached the talent, devin Co. Picks up the scraps
Cognition also works with Goldman Sachs now.
I have been testing Devin for a long time, early access and all. I'm not impressed by it at all, a decent developer with their LLM of choice does a far better job.
I mean isn't that amazing for an 1 year old product? If it's already better than a terrible dev with an LLM, or better than a decent dev without an LLM, it's not hard to imagine in 2-3-5 years Devin is better and cheaper than most hires you could do. Without having to do HR, equity etc.
Does anyone actually use Windsurf? I know a ton of people on Cursor, Cline, Roo, Claude Code... but zero people in my engineering circles have even mentioned trying Windsurf.
For me it was substantially better than Cursor. Their RAG/code indexing/whatever implementation is better (Cursor seems to completely fall apart on larger codebases, unless you keep a very detailed cursorrules file). The cmd+tab was also substantially better. Plus it's cheaper than using Cline.
But with all this changing of hands, I'm not sure I can trust it going forward at all, so I guess it's back to looking for alternatives.
I tried it once and couldn’t understand why it exists. Neither windsurf nor cursor seem to offer anything Cline doesn’t, and the real Microsoft developed tools like Pylance are also broken and can’t be replaced with shitty AI.
I was using Windsurf in Pycharm, until I switched to Claude Code / Gemini CLI last week. I am also finding that Copilot with Sonnet 4 is pretty on par to Windsurf.
https://sweep.dev/ is better than Copilot/Windsurf in PyCharm (disclaimer I am one of the founders)
It has been my go to IDE over all others.
They had released their own model which was free and good enough a couple of weeks back.
Obviously will need to look for alternatives.
Very curious how we'll look back on Google spending 2 billion dollars to "license IP" and hire a handful of people.
If there's 47m software engineers in the world, at $200/month, and 50% gross profit that's a $56 billion TAM. Not crazy to think it's more if we include the adjacent space of analyst roles that write software (sql, advanced excel, etc).
They'll have to crush it to make a $2 billion acquihire look reasonable, but it's possible.
It’s more about not getting embarrassed ever again. The breakthrough papers for LLMs came from Google, but not the product. This was embarrassing. Even if you have to spend double on everything from now on you do it, because Google effectively got sucker punched by ChatGPT.
It’s sad that the organizational failures that led to that embarrassment will definitely not change. Too much eagerness for subjugation
I don't think its just SDEs they're targeting.
Basically anyone that inputs and outputs goods which can be digitized. So writers, graphic artists, accountants, legal work, etc.
that's not the market they're thinking of, they're thinking of the total amount of money spent on developers per year, globally, and capturing a percentage of that
They are doing what they've always done skirting the edges of antitrust laws by buying up competitors before they become competitors so that they can shut them down.
Will require a lot of barriers to entry to charge %. Not disagreeing with your statement of their thinking, but will be v surprised if that pans out.
> 100% of Windsurf employees will participate financially in this deal
> 100% of Windsurf employees will have vesting cliffs waived for their work to date
> 100% of Windsurf employees will receive fully accelerated vesting for their work to date
This sounds like a happy ending for the employees of Windsurf and a good deal for Cognition
> This sounds like a happy ending for the employees of Windsurf and a good deal for Cognition
The employees were robbed from having a big cash exit. Illiquid stock options from Windsurf were converted to illiquid stock options of Devin.
What's worse is that the well is now poisoned. I would advise against joining startups from now on, because I think that there's no upside for employees anymore.
I'm not so sure. No telling what the purchase price was.
Or how much of it is in Cognition stock.
I don't think so. If your shares dropped 90% overnight you wouldn't be excited to have your vesting cliff waived
"work to" is key.. sounds like they aren't fully accelerating, just accelerating the time they worked that hasn't vested. So if you were hired a week ago, you get a week's worth of participation.
Are the cliffs waived or accelerated? Makes me not trust this
Great news...saw Windsurf vibing on YC podcast several months ago...and it's my go-to now. Google Gemini Pro 2.5 works best for me (Python, C++, Fortran).
Just finished a weather analytics code translation test from Fortran to C++ and Python. It was simple temperature high/low averaging, so to amp up the processing, I worked with Gemini to add a Monte Carlo simulation for a 7-day weather forecast, and a 2D heat diffusion/advection physics model to calculate the temperatures across a NxN grid over T time divisions.
I've never heard of Cognition (not a slight on them, its just me being disconnected). How much was the deal for and how do they have this type of money? Wasn't there just an article out yesterday or so saying that Windsurf's leadership was going to google?
All these agentic tools are great for Nvidia. They just hoover up tokens like no tomorrow.
This is so confusing
I totally missed the puck with this one. There was a time where Cursor did not have the agent feel of a true AI pair-programming buddy. Windsurf had that magical aspect, and I totally thought they would destroy Cursor. But it took Cursor about 2 weeks by my recollection to add agent mode, and ultimately I went back to Cursor because of their better WSL2 integration.
For those brief 2 weeks, Windsurf felt like the SOTA tool. Crazy how the winds change.
Same experience. I was using Windsurf for a couple of weeks, but felt like the editor wasn't as nice to use as Cursor, so features an agent could handle went to Windsurf and pair coding was more with Cursor. Once Cursor got agent mode, I haven't really touched Windsurf.
Feels like a new SOTA tool every couple weeks. Heck, the post below this is about a new agentic IDE.
As someone who stuck with it - it's good. And it's not actually just a thin wrapper. Their tab model is really, really good now (I think that was Cursor's big advantage for a long time). And they also price very fairly (read: heavily subsidized, minus Claude).
I writing this just after the night end llm cannot even replace to using Trans i18n react component if it is not in context of the file. We are still far from agi
"Devin" is making itself fatter and prepping itself for another acquisition. Google or MS? Maybe Apple?
The "world-class GTM" is a joke.
Will the remaining employees not bought out and brought over to Microsoft get a windfall in this deal even if they’re stakes haven’t vested?
I thought OpenAI already acquired Windsurf for $3BN? And Anthropic refused to give them discounted version of Claude 4 because of that?
I hope Windsurf employees made some money in this whole kerfuffle. Would be terrible if they got left out of the payday.
Very clearly called out in the article:
Not clearly called out. That call out is as clear as mud. We have no idea how much if any money the left behind employees will get and it's almost certainly far less than they would have got if their execs had held out for a deal that benefited everyone instead of their selfish and damaging to the entire startup ecosystem deal with Google.
The ones that went to Google did! I don't know if the remaining employees are getting anything out of this acquisition.
Wait, is this final? Has Cursor or Augment Code or the next big Coding Assistant made their offer already?
that the founders went one way and the engineers went another suggests a failure of leadership IMHO.
Having been acquired by Google, there is always a leeway for the execs to take the employees with them. Google is a weak negotiator when they NEED something.
On the other hand, I can imagine the execs taking Google golden handcuffs while trying to close the Cognition deal so the employees are made whole or maybe even on better terms than if they all went to Google.
What does this actually mean for the product? Huge fan of the plugin on Jetbrains products...
Transparently I'm a founder building a plugin that's like "Cursor for JetBrains IDEs", if you're an windsurf user you might like us: https://plugins.jetbrains.com/plugin/26860-sweep-ai
How would you describe your stance towards security? I'm an enterprise user so it'd have to clear a high bar with our security team
I think the two big things are privacy mode on by default and zero data retention by default.
No code or prompts are stored unless you opt-in. We also have on-prem deployment options but it's much more expensive.
Which plugin?
The Windsurf plugin!
This is fantastic news. Especially the way they are structuring the equity payout.
Galaxy brain move by cognition
"Devin's on first" just doesn't work the same
Important context from the tweet:
> This transaction is structured so that 100% of Windsurf employees will participate financially. They will also have all vesting cliffs waived and will receive fully accelerated vesting for their work to date.
Not important, marketing speak unless it comes with numbers. "We'll split 100% of $10 across every employee" is just as much a possibility as the windfall their PR team has convinced you of.
So they will get stocks of a company that ceases to exist basically
Uhh so where does the 2.4B go to? For the "licensing" rights but without equity to Windsurf? Does the whole 2.4B get distributed amongst the talent that Google acquired or is it shared amongst all Windsurf employees?
Windsurf employees left behind by their executives get none of that Google money. It went to the selfish exec team and a few top employees that abandoned their teams and projects for Google's cash, and what didn't go to those selfish jerks went to Windsurf's investors. The rank and file that were abandoned by their leadership got totally screwed. Startup executives will never again be trusted to deal fairly with the people who spend their opportunity trying to make that particular startup work rather than some other, or simply going to work for an incumbent. This deal probably did more damage to the tech startup ecosystem than anything I've seen in my 25 years in Silicon Valley.
did the team that joined deepmind forfeit their equity or will the cognition acquisition basically allow them to get a double payout from their equity?
What a waste of money!
Devin/Cognition?
All right, cancelled.
Sounds like one of those "Two turkeys don't make an eagle." situations.
What's left ???.
The Devin name/branding is so toxic that nobody will try their current product offerings, so the hollowed out shell of a respected company is actually fine for their needs.
The company was intact after key employees moved to Alphabet, and importantly there was a service contract to use Windsurf, so apparently Google will be paying Cognition, ultimately, now.
> there was a service contract to use Windsurf
I wonder what the terms were there. Hard for me to imagine why Google would've included that in the deal.
I'm just speculating but you'd get to use the IP without needing to negotiate overmuch about it, if they're using GCP or can be retrofitted to then you could set prices at a discount but claim a market rate when discussing growth of usage, which is great at quarterly earnings report time.. or even just as an ease-of-transition to expand into Windsurf's existing user base? There are plenty of reasons, including negotiation leverage, optics, or even just appeasment to the founders or board that the original project doesn't get immediately destroyed.
the hosed employees that actually built a great product while their leadership sings "go on take the money and run" on their mega yacht.
Presumably the employees were compensated fairly for their work while they worked there?
Why? Why presume that? These employees got screwed. They were left with a shell of a company that got bought for pennies on the dollar while their executives and executive pets all got massive payouts moving to Google. It was a total dick move and the result will be a flight from startups because who wants to bust their asses for an executive team that will leave them behind when the payout arrives.
And they're also now compensated by the purchase by Cognition as post says
Users and enterprise contracts?
What a thoughtfully written letter. You have to respect great leaders when they communicate in this eloquent, respectful manner.
This is the logical, satisfying, and probably best conclusion to an un-ideal and optically terrible situation all parties were placed in.
First, OpenAI wanted to acquire Windsurf. Terrific move! Win-win for OpenAI (who needs more AI product) and Windsurf (for the deal price). But this fell apart because Windsurf didn't want the IP to go to Microsoft (which imo should not have been not a big deal, especially if you knew what would have happened next). Big loss for all parties for this to have fallen apart.
My biggest question still is why not continue on as an independent company? Perhaps losing access to Claude doomed signups; perhaps employees/investors had a taste of an exit and still wanted it; perhaps due to fiduciary duty to maximize returns; perhaps their growth stalled due to the announcement? In any case, the founders got a similarly equivalent deal from Google, and were arguably wise to pursue it.
But Google's Corp Dev team here is the most maddening. Why not fully acquire the entire company, instead of doing the same "acquihire and license" deal that was done to Character AI, Adept, Scale, etc.? Risk of FTC antitrust review is a thing, but Google's not even competitive in the coding market, so I doubt there is a review (though I do hear that all acquisitions by large tech companies these days are reviewed by default). If there's anyone to blame in this situation, it's the FTC and Google for pursuing this strategy, instead of a full acquisition. Win-win for Google (for the team) and Windsurf (for getting a similar acquisition price, but liquid!).
Imo, the founders did a good job ensuring that close to the $3B acquisition price was reflected in the $2.5B Google deal--all existing investors and vested employee/equity holders are paid out; the company also retained $100M which was suspiciously similar to the amount needed to pay out all unvested employee/equity holders [1]. So theoretically the remaining company could pay accelerate vesting, then pay out the cash to their remaining employees, and then shut down, to give everyone the same exit as an acquisition, or better. This might have been the best scenario, because the brand damage to Windsurf as an IDE that happened over the weekend was pretty close to unrecoverable for them as an independent company.
But instead, the company leadership decided to field acquisition offers for the remaining company and IP, and got one from Cognition. (I'm actually surprised this acquisition isn't under FTC review; it's more plainly an agentic coding company acquiring a competitor agentic coding company). In taking the offer, it reinforces that the Windsurf IDE will continue to exist, that they have a R&D team backing the IDE again, and can marry Windsurf's enterprise sales chops with Cognition's product [3]. Win-win for both Cognition and Windsurf.
So overall, win-win-win all around, except for OpenAI, Varun's public reputation (imo, undeserved), and startups hiring employees (who might think they might not get a proper exit) [2].
[1] https://x.com/haridigresses/status/1944406541064433848
[2] https://stratechery.com/2025/google-and-windsurf-stinky-deal...
[3] https://x.com/russelljkaplan/status/1944845868273709520
> -all existing investors and vested employee/equity holders are paid out;
But the statement from Cognition was:
The details matter. "vesting cliffs waived" meaning what? Windsurf shares exchanged for Cognition shares? at what ratio?"Participate financially" means what exactly? They could all get a coupon for a free doughnut, and that statement would be true.
I'm not saying the employees are getting nothing, or even a raw deal. I'm saying we have no idea if the deal is good for them, without details.
> theoretically the remaining company could pay accelerate vesting, then pay out the cash to their remaining employees, and then shut down, to give everyone the same exit as an acquisition, or better.
unlikely that will happen. More likely the investors and VCs will take the lion's share of the $2.5B, that is what they do. That is why they exist. And they'll distribute as thin a slice as possible to the employees.
Historically, when a company gets acquired, the terms of the acquisition vary wildly. Many acquisitions over the past few years have led to a payout of all equity holders, pennies for employees, and layoffs for much of the existing team. This outcome is no worse than an existing acquisition--people just want details because of the new structure. My point is that this is financially no worse than an existing acquisition. It just _feels_ worse because of how it's structured.
And to your last paragraph, read reference [1]. The distribution of that 2.5B is in accordance to the existing cap table; it will make sense once you read that tweet. You must allocate money according to the cap table, and so that allocation is already determined in a company's previous funding rounds.
I wonder if we'll ever stop to ask ourselves if faster and faster output of software is actually a good thing for the world. Or will we just continue because it's just what we do nowadays in civilization to get ahead?
A lot of people are asking that question, and the answer is emphatically yes. All improvements to the human condition are rooted in technology, and software is technology. Who's to say the latest advancements aren't some tech tree precursor to cure an ailment impacting millions - how could you argue against that? The genie is out of the bottle.
> how could you argue against that?
I would argue against it if the downside is even more technological enslavement for billions.
And while many improvements to the human condition are rooted in technology, many of the problems of humanity are rooted in it as well. There might very well be an optimal point that we've already past.
ahead of what? ahead of generating ugly and mostly usable masks over the same data. I'm in favor of AI but it seems to me that no one really stopped asking himself what real problems people have and how to actually fix them
I think so but not for the reason that you think.
See, most closed source software really just pisses me off of ideological reasons, I just like to tinker with things and just having the possibility to do so by being provided the source code really helps my mind feel happy I guess.
So I "vibe coded" a game that I used to play and some projects that I was curious about and I just wanted to tinker too. sure the game and code have bugs.
Also with the help of AI, I feel like I can tinker about things that I don't know too much about and get a decent distance ahead. You might think that I am an AI advocate by reading this comment, but quite the contrary, I personally think that this is the only positive quality that AI helped in quite substantially.
But at what cost? The job market has sunk a large hole and nobody's hiring the junior devs because everybody feels better doing some AI deals than hiring junior devs.
My hunch is that senior devs are extremely in demand and are paid decently and so will retire on average early too. Then, there would be a huge gap b/w senior and juniors, because nobody's hiring the junior engineers now, so who will become the senior engineers if nobody got hired in the first place. I really hope that most companies actually realize that the AI game is quite a funny game really, most companies are too invested into it to realize that really, open source AI will catch up and there is just no moat with AI and building with AI or just doing stuff with AI isn't that meaningfully significant as they think it is as shown by recent studies.
> that senior devs are extremely in demand
Is this true? I am not seeing salaries rising, the demand seems to be met. But maybe I'm wrong.
Sorry I guess, I may have been incorrect in that regards. I actually just meant as in comparison to juniors really. And I personally felt that way from what I've heard from all the people, I am not sure too about salary rising but still I always thought that seniors are getting on with more and more responsibility since juniors aren't getting hired and so I thought that they were more compensated and I am pretty sure that I heard it somewhere and I think I just repeated that.
Also maybe I felt this way because of 100 Million $ and the 30 Billion $ acquisition by Zuckerberg I guess
I might ask AI (Oh the irony) and here is the chat https://chatgpt.com/share/68756188-d374-8011-9f23-6860d6b1db... and here is one of the major source of this I suppose
https://www.hackerrank.com/blog/senior-hiring-is-surging-wil...
And I would like to quote a part from the hackerrank ie. Taken in isolation, this might suggest a cautious but healthy rebound. But viewed through a 2025 lens, a deeper pattern emerges: teams are leaning hard into experience, and leaving early-career talent behind.
I use cursor
Two products that nobody uses, together at last
I hit myself in the head as hard as I could without causing permanent damage
Care to explain? I am honestly clueless
I don't even remember anymore. I guess hitting myself in the head worked..?
But in general my reason for hitting myself in the head is that I'm an IDE author and people keep tripping over themselves to place great value on forks of VSCode
Windsurf already had sold their code and most of their devs to DeepMind. The company is worthless, idk what Devin thinks they’re buying.
Clearly you are more informed wrt to remaining IP and talent than the CEO of Devin.
Clearly.
[flagged]
Could you please stop posting unsubstantive comments and flamebait? You've unfortunately been doing it repeatedly. It's not what this site is for, and destroys what it is for.
If you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.
To be fair OpenAI didn’t screw windsurf yet. Or did they
Windsurf screwed OpenAI from the sound of it.
Hello Cognition AI,
My name is Devin; it has been for many decades now. I'm embarrassed to see you've named your product after me. It has already prompted uncomfortable jokes at my expense, and I'm sure there will be more. I now have newfound empathy for people named Alexa.
For instance, people have made jokes about my name in interviews, and it's embarrassing for me, and thus awkward for everyone, and awkward interactions make it objectively less likely that I will get job offers.
I don't think any product should be named after people. Please change the name of Devin.
Apart from this news I haven't heard about Devin in a long time. Not since the very start of the AI hype. Considering how foolish a purchase this seems to be I don't think you're going to have to worry about them Bogarting your name for too much longer.
Meanwhile, Claudes are getting ahead!
Plenty of men would love to have their name be a euphemism for eroticism.
Reading this made me wish I hadn't named my son Windsurf