Reminds me when I got banned from Amazon for suspected fraud (had an old account, but deleted my email and number since it was in a lot of DB dumps). After I got hired, I reached out to the guy in charge of the anti-fraud team at Amazon, and got unbanned. Emails to support etc. did nothing before I reached out internally (unbanned by 1am the next day).
My biggest Amazon annoyance. I'm often looking for some product, reading Reddit and other reviews. They usually link to amazon.com.
Then it asks me to switch my profile to American/$. But then in order to order I need to switch back to Germany/€.
It's just super cumbersome. Just let me view stuff from any region without switching profiles. If I order from that region you can tell me to switch profiles. But not just for viewing it.
In the same vein. Why is there no, I want this thing, but from a German seller.
Reminds me of when someone links to a product on some (non-Amazon) website, I go there, it says "this is the US site, you should go to the [your country] site!", I click OK, and it takes me to the homepage.
With Amazon it's even worse, if you click on an US Amazon product link on an Android device:
My app is set to use Amazon in Germany. I click a link on a page and this opens the app. The app says it needs to switch to US. If I do that, I'm signed out of my German account and end up in an empty US account. I also think that I didn't even land on the product page afterwards. So I had to sign out of that US account and re-sign in into the German account to use the app normally.
So I basically just can't see the link unless I long-press and select open in new tab.
I get this from the Nothing (phone) emails, the direct links in them go to the UK site, but if I let it redirect me to the US site, it's just the homepage. Extremely annoying.
It doesn't work that way. A link is to an ASIN (unique product in Amazon). The same product might have a different ASIN in a different region. The product description, title etc might vary.
Clearly there exists some higher-level product concept with a mapping of country->ASIN for that product. How hard is it to lookup the correct ASIN?
Maybe you end up on the home page when there is no ASIN for your country? There should be a nicer message telling you that’s what happened. But if it always dumps to the home page, that’s just dumb — Amazon could easily do the lookup.
There's no such mapping, you could do a text search and if the item name matches then maybe US ASIN 2785334 should correspond to UK ASIN 3894948, or maybe some third party just named something ambiguously. They're essentially GUIDs and essentially separate businesses.
> In the same vein. Why is there no, I want this thing, but from a German seller.
eBay used to let you filter results by region, but apparently that ruined some kind of metric and the option is gone. When buying games or books or what have you I only want sellers in the EU. I don’t care that it’s available in the US or UK, shipping and duties are going to be 4x what the product is worth.
It doesn't work in all cases, but I can often replace the .com with .ca. It doesn't guarantee availability, but can at least check if the same listing is available on the Canadian store.
Interesting. I still have a bricked phone from my onboarding at Google, and no internal people cared either. There's a tool I could have used to fix it, but it's accompanied by a message saying that if you use it without permission you'll be fired.
> accompanied by a message saying that if you use it without permission you'll be fired
Probably why none of the internal people cared either. They didn't want to be the person on the line in case it was determined that the usage wasn't valid.
I'm curious how you bricked it beyond repair though. Most devices have a way to enter a recovery/flash mode where you can upload your own firmware from the bootloader. And if you haven't unlocked the bootloader then I don't get how you could have bricked it unless there's an Android bug... which would have probably triggered a more serious look.
Most likely, the tool to upload the firmware from the bootloader required some sort of hardware signing key, which was present within the tool that he wasn’t supposed to use without permission.
It used to be, if you left a provider without paying for the device on time, they would remote-brick it, which would burn a fuse so it wouldn’t work anymore. I used this to get out of paying for the device once (Verizon), since I couldn’t use it anymore, why would I pay for it?
> There's a tool I could have used to fix it, but it's accompanied by a message saying that if you use it without permission you'll be fired.
Sometimes I name certain APIs/function names/whatever with a "do_not_use_or_you_will_be_fired" suffix. Generally for hacks I don't want people to copy-pasta. I can't actually fire anyone, but it gets peoples attention (especially more junior folks).
I'd love for this to happen for me. I lost my Amazon account because they miswrote my phone number - having a single leading digit invalid for international use. So couldn't use text MFA and my phone with the OTP app had died the previous day.
If Google Maps would like to hire me so the km/miles switch can remember I only ever want to see distances in km, my contact details are in my HN profile.
I must have changed that back from miles once a fortnight since Google Maps launched 20 years ago. That's 500 times. Totally ridiculous for a company who core goal is profiling their users...
It's funny cuz is true. Except it'd probably be one long design doc with 10 rounds of review, 15 CLs (PRs) and months of rollouts later ... fails A/B due to declining user engagement.
While you’re at it, can you find and punch the guy who thinks it’s a good idea to zoom the map to “actual size, 1cm = 1cm” mode for your entire trip?
I assume he’s also the one that taught it to spitefully let you drive off the side of the screen if you ever zoom out manually so that you can see more road on the phone than you can in real life. (With a “recenter” button that will zoom you all the way back in).
Also punch the guy that allows Google Maps navigation to flip back to a route that's been specifically rejected.
Earlier this week Google prompted me with "your route may be affected by tsunami warning". Indeed, so I chose the longer, inland route rather than the coast roads.
15 minutes later I realise it's rerouted me "due to traffic conditions" -- obviously the coast road isn't as busy!
(This has happened many times before, but this was the first time I had a safety reason not to take the faster route.)
While you’re there can you add a ‘how much I value my time’ input field for tolls? Google suggests I spend $20 through 3 tolls to save a single minute. Constantly.
Edit: and while you’re there, move the ‘speed camera ahead, is it still there?’ Dialog. IT COVERS THE DAMN SPEED LIMIT ICON.
And also while you're there, if no car ever in the history of your app goes down the road at the speed limit ever it's a good indication you'll never be able to ever do it at that speed. e.g. small narrow single lane country roads which are only theoretically 60mph roads.
Imagine the data they must have on the speeds people actually drive on every mile of every road, they’d easily be able to warn you not that you’re “over the speed limit” as in driving 70 on the freeway, but more usefully, if you’re in the top X percentile of speeds usually or even currently driven, which actually is a decent measure of unsafe was and would also be a great predictor of likelihood to get a ticket.
I saw this video recently where the author set up a camera to record sections of highways and measure the speeds of drivers, and make cool graphs out of it.
I share your intuition that your likelihood of getting a ticket is related to the speed of other vehicles. Presumably, police choose to prioritise ticketing the worst offenders when there are too many offenders to handle.
But I don't share your intuition that safety is also relative in that way. If you're driving dangerously (too fast, or while drunk), you're driving dangerously, even if everyone else is driving dangerously too. If you're in a country where nobody wears a seatbelt, it's still prudent to wear a seatbelt, just as much as in a country where that is the norm. I don't think Google Maps should encourage people to drive as dangerously as everyone else. Quite the opposite!
"dangerously" and "fast" are not synonyms (while drunkess obviously is dangerous regardless). The 280 in San Mateo County is designed for 80MPH, with banked turns, gentle curves, etc. The speed limit is of course 65. If most people are driving 80-85, and you enter that highway and drive 60, which is a perfectly reasonable speed according to the speed limit, you are much more likely to cause an accident than if you just drive the speed of the people around you.
It's fine to point out that many people are terrible drivers and that a given crash that happens is more dangerous at a higher speed and if everybody were to drive under 60 at all times we'd all be safer, but clearly that will never happen unless we install a totalitarian government, put governors on all cars and give prison time for disabling them, with enforcers stationed everywhere to monitor. But no democracy would vote for that, so I don't think it's worth spending much mental energy on such hypotheticals.
> encourage
I think just the opposite: My feature would encourage people like me, who drive a "fast car" and can occasionally accidentally go too fast if a road is especially uncongested, to slow down to the speed that is customary or that others are driving, by reminding me that I could get a ticket and that driving faster than everyone else is dangerous.
Relative speed is important though. Where I live it’s completely legal to drive half the speed limit, and people often drive above it. People driving a consistent speed would reduce lane changes, breakaway traffic etc. imo speed consistency is what’s valuable, not preventing upper outliers.
When I was traveling in Mexico, it drove me nuts that even though I was signed in, Google Flights switched my currency from dollars to pesos every single time I opened a new tab! I think they really don't care.
I think they rely on ip for a lot of stuff they shouldn't. Getting a local esim switches me to km until I switch back to my old one. Have no idea about Australia.
Edit: after typing this realized this isn't ip, its provider. That maybe does make sense to cue off of.
Also accepting gmaps work, if only it could preemptively cache the return trip for any trip longer than an hour, so that I'm not stuck with no service trying to remember how I got there.
I clicked the link with nothing better to do, and woa, that's a really good maps app. Like, "I haven't seen something like this before" levels of good. Reorients itself with my phone, accurately - and in real time, not after I'd already walked ten steps in that direction. With other maps, I thought that maybe my phone's compass is broken. The default but optional 3D top-down view is the most comprehensible-looking map I've seen in a long time. It barely uses disk space, and going by my short experience with it, it really is very light on the battery.
Exceptional, this is what I'm using from now on. Just hope the iOS 15 support is maintained, that's a killer app to keep perfectly good devices productive even after they're restricted from everything else :)
Happy to have made you discover something cool :-)
> 3D top-down view is the most comprehensible-looking map I've seen in a long time
The 3D top-down view with building heights is possible thanks to OSM providing this data, thanks to people improving the map with apps like Street Complete and Every Door. Make sure to check them out :-)
There's with this 3D view though: it sometimes hides streets. I usually end up disabling it at some point.
Wait, there's a setting for this? I've lived in Australia for over 16 years now but everything is still in miles instead of Kms and I have never been able to find a setting to change it (although it sounds like even if I did find itz it would be mostly useless).
Ive lived in Australia for 45 years, everything is in km.. never had to touch it for miles. However i did go to the US and it showed units in miles on my phone which made no sense.
In Gmaps, Tap your profile picture, then select "Settings" and "Distance units". Choose between "Automatic", "Kilometers", or "Miles".
I got the app wrong, it was Uber. I have my phone set to English (UK) and change the measurement system to metric. Uber doesn't respect that though, so it keeps using miles.
Ah, thank you, this was it. I had the language set to English (UK), but changed the distance setting to KM. I got the app wrong, it was Uber, and Uber doesn't respect the override, so it always uses miles. Changing it to English (Australia) and Uber switches KM.
People tend to book using the currency they're familiar with, not necessarily the local one, so remembering their preferred currency makes sense. As someone who travels a lot but likes to see flight prices in USD, this is really annoying.
Another Google Maps request: Places I‘ve labelled used to at least sometimes (if inconsistently) show up in search. Now they never shop up, even when I type the exact name of the label.
Another issue with Google Maps is it not showing Plus Codes for some locations that highlight the entire area. If you however place a pin on that location, it provides a Plus Code. Pretty stupid IMHO.
Also it is really, really hard to search for "Nearby" places. Have to do it through "Directions". Really bad UX.
Google Maps core functionality is sort of in maintenance mode, and things have been slowly bit rotting over the last 3-4 years.
Unless you want to launch some AI feature (used to be chat app for ten years and then Google got bamboozled by ChatGPT…) you’ll not find allies and your career will not progress.
While you're at it, could you also change YouTube so that captions being on/off is a per-language thing and not a "The user turned off subtitles for language X, therefore it's the only they speak, therefore let's turn on subtitles for all other languages." thing? It's really annoying that whenever I watch a video that has a language different from the previous video I watched, the first thing I have to do is to turn off the captions.
This isn't a proper solution, but if you watch with yt-dlp+mpv, you can configure default audio and sub languages in mpv (globally) and they will also work for YT videos in addition to your movies and such. Plus if you do toggle them on/off for one session it won't mess up your mpv config.
> I must have changed that back from miles once a fortnight since Google Maps launched 20 years ago. That's 500 times. Totally ridiculous for a company who core goal is profiling their users...
Similar with "privacy popups" everywhere. Similar with every bank with "remember this device" feature. I add exactly the same device on every login, until it fills entirely the limit of allowed devices.
If you are following a route in Google maps in Apple CarPlay you can’t search in the map without cancelling the navigation. So you need to use another phone or another maps app.
It would be a very useful feature as on road trips you can have a passenger setup the navigation rather than the driver and then also look around and examine places along the route.
The same here. Oftentimes Google Maps switches to most exotic language "because of reasons". Then I set it manually to my preferred language, for decades now, and it's been the same language all the time. Are they trying to say with poker face "we don't track you"? I don't think my home-grown privacy habits are that good, somehow they manage to show me an ad of "blue sneakers" someone in the household mentioned this morning.
It’s true, except expand that to all big tech companies. The only time UX is changed it’s either to make ads more effective or to “streamline” things by shoving more and more of the functionality into an endless nested chain of ••• and More menus.
Clicking through the links in his article, I came across a guy who apparently did the same thing at Apple - he introduced the "auto remove" feature for expired passes added to your wallet, then promptly quit. I had no idea that's how that feature came about, but now I'm going to send a little mental thank you to him every time I get off a plane. That shit was FRUSTRATING.
Crazy because I remember that the first few iPhone OS versions had really bad autocorrect dictionaries, especially for German. The workaround for that was to make contacts for missing words because contact names never got marked as misspelled.
I need to get a job at Apple to stop “omw”-> “On my way!” (Complete with the `!`) from reappearing in my Text Replacements every month or two, no matter how many times I delete it.
(Try typing “I’m omw to the car” or something to see how annoying this is)
Apparently the current state of the art to fix this problem is to remove it and add an "omw" → "omw" text replacement in its stead. A friend recommended this to me and I haven't had problems with it since. Yes, it's nuts, but it is what it is.
> For the longest time they haven’t shown up alongside active passes.
They have always shown up for me, and the only way to delete them is from the wallet app. Note: from the app. You can't delete it from convenient screen where you access them
True enough, but considering most websites these days consider "a comfortable margin" to mean "4-6 inches", I'm delighted to see a site which actually lines things up close to the left side of my monitor. Like I said, that's how text should be.
> I added an AbortController to the debounced search function, so that it aborts any previous queries when a new one is made. This means that the search results are always relevant to what the user is currently typing.
To me one of the most annoying things an application can do is go off and do something before I'm done telling it what to do. Filters that apply themselves without an explicit indication that I'm done setting them up, or searches that are constantly re-executing as I'm typing. Wait for me to stop.
When I implemented search-as-you-type on my blog, I decided to wait for the current search suggestion request to complete before doing a new one. Seemed like a reasonable balance between responsiveness and not overloading the server.
When I implemented one I simply filtered the results that came back from previous requests to see if they matched what the user had typed in the mean time. That way the UI might get relevant results with lower latency than would otherwise be possible with no risk that a non-matching result would show up to confuse the user.
The video showing the problem in the article seems to show an avalanche of queries towards the server, I'm surprised no one cared about it, but I guess it's frontend people thinking "It's the backend/ops that have to deal with the problem!".
I wait about 250 ms before firing the request, if the user (well, me) continues typing, then the timer gets cancelled and the app waits another 250ms.
The earlier search will be on a shorter input which will find more results; imagine waiting for everything that contains "e" before you can get the one result for "europa" that you wanted. Isn't that pattern going to happen most of the time - waiting for a large amount of useless generated and transferred and disposed-of, before the search you want will run?
That's why I set it to run after the third letter and only return up to 7 suggestions. In your example, it wouldn't run until "eur", which should narrow it down quite a bit. Given my rather small blog article corpus and aggressive full text search indexing, it's very fast.
Firing queries all the time is especially annoying if your users are in another continent, and you don't have proper state management to only show results for the latest query, as opposed to the latest response.
RTT from Europe to AWS us-* can easily get to multiple seconds during peak times.
I hate this on booking websites. Especially if the filters are in a sidebar on the left and do not fit your viewport and every god damn time you change something it scrolls up, starts loading, puts filters into read only mode until it's done just so you can add the next filter...
I think a good middle ground is to wait a few hundred ms at least, for the user to stop typing, before sending off the query. Or maybe, still send the query, but don't populate results until they stop.
It is loathable and contemptible behavior. I despise it. Writing in fancy code editors is an exercise in "remember this display was very expensive calm down".
> type out 'i' and 'f'
RED ALERT! RED ALERT! UHHHHH HEY THERE CHAMP YOUR IF-THEN STATEMENT ISN'T CLOSED OUT PROPERLY!
Yeah, I know, I'm not done fucking typing yet. Give me a second.
Thankfully I can usually turn this crap to only show these after a certain delay, for many languages I write in I usually turn it off completely except for on build/run. Let the complier/shell throw me errors, the LSP can take a chill pill.
It's nothing to do with quality. It's prioritization. Companies pick things they think people want (or what the team wants to build) without testing or experimenting with users and data.
This is actually an example of the problem, not the solution. There are probably much more useful things for the team to be doing but they let one guy add the thing he wanted.
It's because the people who spend the most money on the scam cards for GTA:O didn't care. I personally stopped playing the game after getting so fed up with the loading times, I timed it for an hour and found out I was spending more time watching the loading screen than I was actually playing missions.
The article says nothing about the hiring, which is kind of the most important part of the whole escapade. Right now, it's a bit "something was bugging me, and when the company hired me, I fixed it", which, great?
> It reminds me of George Hotz’s legendary single week at Twitter in 2022, where he joined just to fix a login popup that was bothering users, then bounced.
The author remembers this, uh, event differently than I remember it...
George Hotz boldly claimed that he could "fix Twitter search" faster than those lazy Twitter devs, only to bail almost immediately. Hubris!
On the way out, he removed that login popup as a sort of consolation prize.
Well, that's a bit of my time gone (re)looking into GeoHot, patent trolls, and now comma.ai.
Comma.AI by George Hotz sounds very interesting, it's basically a $999 "comma 3x" smartphone with an OBD-II connector and a $99 wiring harness that can add an equivalent of a Tesla Autopilot to many cars manufactured in the last 10 years (even Tesla's own cars, too), for a total cost of $1098, whilst being OSS and available on GitHub, and — get this — even having ssh access to your car! Optional cloud subscription plans are $10/mo for your own SIM, or $24/mo with bundled cellular data.
Sadly, it does NOT have an equivalent of Tesla Sentry Mode yet, https://github.com/commaai/openpilot/issues/29912, which is kind of unfortunate, because Tesla's own implementation of Sentry Mode is using 250W of power — depleting the entire 80kWh battery from 80% to 30% in like 7 days (".5*80kWh over 7 days" = 238W) — openpilot would have been a nice alternative at what'd presumably be around 5W or less ("40kWh / 5W" is 333 days).
I followed the whole saga on Twitter. He shared a video of his browser saying, "I fixed it in 5 minutes," and 5 days later he was still trying to figure out why his PR was failing the build. When Twitter engineers told him to write tests, he rage quit.
In the reverse situation, I've worked at places where the IP lawyers basically made it impossible to submit PRs to open source code.
But sometimes explaining the exact inputs and the line number where you know the problem is can grease the wheels enough that you can convince someone else to write the fix for you. I didn't technically give you any code. But I did give you free QA.
I specifically attempted to get a job at Discord so I could submit a PR to make giant emojis be a toggle setting rather than automatic. I know the feeling.
(If anyone works at Discord, please me and the rest of my server are begging you)
I don't know if this helps but I've been adding a full stop next to emoji exactly for this. It doesn't fix it for ~new people but it's something for yourself?
Discord is an electon app IIRC so in theory it should be possible to make a client side mod which fixes this. Not sure if that would result in your account getting banned though...
I wonder if it's legal for corporations to have employees that they send off to get hired at other companies, do some stuff in those companies that are beneficial to their actual employer, and then leave before the probationary period ends.
IANAL, but it’s almost certainly legal, as long as all parties involved adhere to the applicable non-disclosure agreements, non-compete agreements, and intellectual property provisions of their employment contracts. Even then it’s likely to remain a civil matter in most cases.
Companies can sue each other for nearly anything, so any level of this behavior could result in a lawsuit. It wouldn’t cross the line into criminality until it involved some fraudulent deception or blatant corporate espionage. For a recent example of that, see the ongoing litigation between Rippling and Deel. (But even that egregious espionage activity remains limited to civil court, at least for now.)
"to have employees that they send off to get hired at other companies, do some stuff in those companies that are beneficial to their actual employer, and then leave before the probationary period ends."
To me that sounds like not disclosing, that they work also for another company and this certainly ain't legal on most jurisdictions.
It's probably not the law (it would be shitty when working at a 7/11 on the weekends to have tolegally disclose all your other income resources)
But basic employee contracts cover these aspects, including working in the interest of the company and IP assignments, and usually exclusivity if you're full time.
Yeah I'm aware employment contracts might stipulate it. But violating a contract isn't against the law. Worst case you could get sued (though with an employment contract, the limit of repercussions are generally just termination).
Being binding is kinda of the whole purpose of a contract. If violating it is void under the law the company should change lawyers.
To put your argument under a different angle, there are many written laws you can violate with very limited consequences if any, but they are still laws.
Contracts aren't written by the country, and enforcing them is civil matter so there's nuance, but violating an enforceable contract you provably agreed to is against the law. Whether you can get away with it is another question.
There are two types of law. Contracts are civil law. Breaking them does not break criminal law. Civil vs. criminal law has different procedures, different burdens of proof, and different potential consequences.
When it comes to contracts, no, there are no "laws", there are agreements between parties that can be enforced if taken to court, and in that sense they are binding. But breaking them does not break any law... it just breaks an agreement.
> For a breach of contract to rise to the level of criminal activity, the act must involve elements of fraud, intent to deceive, or theft. These cases go beyond simple noncompliance with contractual terms—they involve behaviors that violate state or federal laws. Some scenarios where contract breaches may involve criminal activity include:
> * Fraudulent intent: If a party enters into a contract without any intention of fulfilling the terms, this may constitute fraud. For instance, accepting payment for services without any intention of delivering.
> * Pattern of deceptive behavior: When a party repeatedly breaches contracts with the intent to defraud others or engage in fraudulent schemes, it can elevate the breach to a criminal offense. A pattern of deceptive behavior indicates a systematic intent to deceive and defraud, which may result in criminal charges.
Not really without researching(also I am european and might have assumed wrong about US), but something with conflict of interest? Especially if another company ordered you to work for someone else. If all is disclosed, probably fine, but undisclosed? Definitely would not work in europe. Breach of trust etc.
Not sure it falls foul of broader laws, but it almost certainly breaches your employment contract, which likely includes something about following the policies of your employer; that policy (in many companies you likely have to go through onboarding training and annual refreshers on it) probably includes a code of employee conduct that has specific mention of conflicts of interest.
You'd achieve more by simply telling the company that you need a certain feature added to their product. If you're an important customer for them, you could probably negotiate a price for them to prioritize the work.
I'm speaking from the perspective of company A, who needs a feature added to company B's product.
They could send their engineers to work for company B, sure, but those engineers' time is still costing money. And those engineers are completely unfamiliar with B's codebase, so they won't work as efficiently. Might as well just pay company B directly for the feature work.
I often have the thought that it would be pretty awesome to take jobs for 6 months here and there just to implement specific features I want in my favorite SW, apps, sites, etc.
* Join Logic Pro team for 8 months and add better score notation tools
* Join Apple's iOS Music app and fix the weird blip that happens at ~17 seconds on any track
* Google Maps to stop the navigation/directions from spelling out how to get from my house to El Camino Real, which I've only done about 10,000 times.
It seems obvious, but I also don't think it's optimal depending on what you are trying to do.
In some scenarios, e.g. bad mobile internet connection, you may also be happy with a slightly stale result (where you still have to ensure correct ordering of responses), depending on how the search overall is implemented.
One additional data point: Algolia used to do query cancellation in the past but stopped doing that (I think at least 5 years ago now), which you can test with the HN search. I'm not sure about their reasoning, but for them that seems to be the best overall default search experience.
Good PR, but AbortController doesn't really help with stopping the server from processing the request. I have seen so much of this type of search that just continues processing in the backend even if the client has long gone caring.
I agree. I was expecting the author would implement debouncing or throttling to reduce the number of unnecessary requests, alongside fixing the data race issue. Here's an excellent article on debouncing and on throttling, and the difference between them:
There's nobody who really has a library that's set up to feed a sequential task into and have it force a synchronous call to be async with breakpoints to check for early termination.
This seems like a problem Sorhus should have a library for, but he does not.
I've had the conversation too many times in the last couple months about how setTimeout() does absolutely nothing to fix this problem in NodeJS. Even Java had trouble with this and tried to delete the API that seemed like it should support this problem, due to undefined behavior.
There was an old legend for an Apple bug (but I can’t exactly remember what). He complained about this macOS bugs for years. Worked for Apple for a couple months, fixed the issue, then quit.
They have, but they're beyond grasp of most developers.
Tests were invented to express the "why" for the normal guy. They don't strictly prevent compilation, but a proper workflow will see them halt your process in the same way, offering the same outcome.
Granted, there are a lot of horribly written tests out there that don't tell you "why" — or, well, anything. As always, people will find a way to abuse anything you put in front of them. But when used well...
With a test, it might link up some functionality with "why" and pass, but then what happens if a business requirement just isn't a requirement anymore? The test will still pass. I'm thinking of something sillier, like a language that forces you to justify why for your code, and then regularly quizzes you if the business reasoning is still true. If anything changes, it rips out the code and breaks your site. :) So then you have to go in to fix it.
I'd also love it if this were applied to politics and laws.
Having the source lets you fix something for yourself, there are an increasing number of barriers being put up to prevent you submitting a fix upstream.
Going through this right now with part of libpng, their mailing list doesn't seem to like my email.
Using a source-based distro (previously Gentoo, now NixOS) lets me solve the problem for myself, even if my PR never gets accepted. Right now the count is at 4 patches in software I use that I submitted upstream that were (for one reason or another) never accepted.
In at least one case, I later found out that I was not the only person to submit a fix for the problem I was running into, but their discussion on the ML also went without comment 3 years earlier.
Reminds me when I got banned from Amazon for suspected fraud (had an old account, but deleted my email and number since it was in a lot of DB dumps). After I got hired, I reached out to the guy in charge of the anti-fraud team at Amazon, and got unbanned. Emails to support etc. did nothing before I reached out internally (unbanned by 1am the next day).
My biggest Amazon annoyance. I'm often looking for some product, reading Reddit and other reviews. They usually link to amazon.com.
Then it asks me to switch my profile to American/$. But then in order to order I need to switch back to Germany/€.
It's just super cumbersome. Just let me view stuff from any region without switching profiles. If I order from that region you can tell me to switch profiles. But not just for viewing it.
In the same vein. Why is there no, I want this thing, but from a German seller.
Reminds me of when someone links to a product on some (non-Amazon) website, I go there, it says "this is the US site, you should go to the [your country] site!", I click OK, and it takes me to the homepage.
With Amazon it's even worse, if you click on an US Amazon product link on an Android device:
My app is set to use Amazon in Germany. I click a link on a page and this opens the app. The app says it needs to switch to US. If I do that, I'm signed out of my German account and end up in an empty US account. I also think that I didn't even land on the product page afterwards. So I had to sign out of that US account and re-sign in into the German account to use the app normally.
So I basically just can't see the link unless I long-press and select open in new tab.
I get this from the Nothing (phone) emails, the direct links in them go to the UK site, but if I let it redirect me to the US site, it's just the homepage. Extremely annoying.
This is also infuriating for all those sites that take to their homepage when you log in instead of taking you back to the page you were viewing.
It doesn't work that way. A link is to an ASIN (unique product in Amazon). The same product might have a different ASIN in a different region. The product description, title etc might vary.
https://www.amazon.de/dp/1408855658 And https://www.amazon.com/dp/1408855658 Are both the same item. Same with https://www.amazon.de/dp/B086WQWSNT/ And https://www.amazon.com/dp/B086WQWSNT/
Seems like changes .com to .de works with every item i can find. Some just won’t be deliverable in a different geo.
Clearly there exists some higher-level product concept with a mapping of country->ASIN for that product. How hard is it to lookup the correct ASIN?
Maybe you end up on the home page when there is no ASIN for your country? There should be a nicer message telling you that’s what happened. But if it always dumps to the home page, that’s just dumb — Amazon could easily do the lookup.
There's no such mapping, you could do a text search and if the item name matches then maybe US ASIN 2785334 should correspond to UK ASIN 3894948, or maybe some third party just named something ambiguously. They're essentially GUIDs and essentially separate businesses.
We all recognize this as an anti-pattern that has directly led to a poor customer experience, right?
Sorry this is just not true. ASINs can be the same cross marketplace. There are lower level identifiers available if needed
To be precise, sometimes there are matching ASINs and other identifiers cross-marketplace, but often they are unique.
> The same product might have a different ASIN in a different region
But why?
> In the same vein. Why is there no, I want this thing, but from a German seller.
eBay used to let you filter results by region, but apparently that ruined some kind of metric and the option is gone. When buying games or books or what have you I only want sellers in the EU. I don’t care that it’s available in the US or UK, shipping and duties are going to be 4x what the product is worth.
This https://i.imgur.com/fTRrVnP.png is not gone, look it's right there.
It’s gone for me, and there are a lot of frustrated support threads from other Europeans so I’m not alone.
I managed to find the option on the advanced search on ebay.ie, but I still get page after page of American items as the result…
It was gone for years, it's useful that they've recently restored the option.
It doesn't work in all cases, but I can often replace the .com with .ca. It doesn't guarantee availability, but can at least check if the same listing is available on the Canadian store.
If it works in some cases it's a simple lookup if the same item exists then show that. And if not you can show similar products in your region.
its a pain, and the option to switch it back is hidden inside the accounts section, which is hard to find and even take 3-4 clicks.
Interesting. I still have a bricked phone from my onboarding at Google, and no internal people cared either. There's a tool I could have used to fix it, but it's accompanied by a message saying that if you use it without permission you'll be fired.
> accompanied by a message saying that if you use it without permission you'll be fired
Probably why none of the internal people cared either. They didn't want to be the person on the line in case it was determined that the usage wasn't valid.
I'm curious how you bricked it beyond repair though. Most devices have a way to enter a recovery/flash mode where you can upload your own firmware from the bootloader. And if you haven't unlocked the bootloader then I don't get how you could have bricked it unless there's an Android bug... which would have probably triggered a more serious look.
Most likely, the tool to upload the firmware from the bootloader required some sort of hardware signing key, which was present within the tool that he wasn’t supposed to use without permission.
It used to be, if you left a provider without paying for the device on time, they would remote-brick it, which would burn a fuse so it wouldn’t work anymore. I used this to get out of paying for the device once (Verizon), since I couldn’t use it anymore, why would I pay for it?
A company the size of Google will for sure have it enrolled in MDM that prevents that.
> There's a tool I could have used to fix it, but it's accompanied by a message saying that if you use it without permission you'll be fired.
Sometimes I name certain APIs/function names/whatever with a "do_not_use_or_you_will_be_fired" suffix. Generally for hacks I don't want people to copy-pasta. I can't actually fire anyone, but it gets peoples attention (especially more junior folks).
React has a bunch of those: https://github.com/facebook/react/blob/ddf8bc3fbac7aefbf557e...
Of course, it doesn't stop people from asking if they can use it: https://github.com/reactjs/react.dev/issues/3896
FRP? Plenty of ways to bypass that, even without any official tools.
Yep. I had no luck getting my Facebook account unblocked until 9 months later when by chance I did some work for them. Instantly unblocked.
This is the level of epic I aspire to in life
Except they're working at Amazon now.
Next level epic is hand your resignation letter right after you get unbanned. "My job here is done."
And then banned again the next day.
tbh you could probably easily have enough gripes with Epic to do this too... but then you'd have to move to Wisconsin.
At least their headquarters are cool
Yes, but now they have free bananas
seems like they could turn this into a lucrative side hustle "super premium secret support" embrace the technofascist feudalism!
That's actually how a lot of laptop and phone schematics and related repair information gets leaked.
I'd love for this to happen for me. I lost my Amazon account because they miswrote my phone number - having a single leading digit invalid for international use. So couldn't use text MFA and my phone with the OTP app had died the previous day.
If Google Maps would like to hire me so the km/miles switch can remember I only ever want to see distances in km, my contact details are in my HN profile.
I must have changed that back from miles once a fortnight since Google Maps launched 20 years ago. That's 500 times. Totally ridiculous for a company who core goal is profiling their users...
> That's 500 times. Totally ridiculous for a company who core goal is profiling their users
Seven interviews later and 1 PR later: Fails in A/B due to declining user engagement
It's funny cuz is true. Except it'd probably be one long design doc with 10 rounds of review, 15 CLs (PRs) and months of rollouts later ... fails A/B due to declining user engagement.
Nailed it in one. Or (similarly) never makes the priority/cut-off list because "what metric does it move?"
Well, the setting is for kilometers, so the metric is metric.
While you’re at it, can you find and punch the guy who thinks it’s a good idea to zoom the map to “actual size, 1cm = 1cm” mode for your entire trip?
I assume he’s also the one that taught it to spitefully let you drive off the side of the screen if you ever zoom out manually so that you can see more road on the phone than you can in real life. (With a “recenter” button that will zoom you all the way back in).
Satnavs had this all figured out in 2005.
Also punch the guy that allows Google Maps navigation to flip back to a route that's been specifically rejected.
Earlier this week Google prompted me with "your route may be affected by tsunami warning". Indeed, so I chose the longer, inland route rather than the coast roads.
15 minutes later I realise it's rerouted me "due to traffic conditions" -- obviously the coast road isn't as busy!
(This has happened many times before, but this was the first time I had a safety reason not to take the faster route.)
While you’re there can you add a ‘how much I value my time’ input field for tolls? Google suggests I spend $20 through 3 tolls to save a single minute. Constantly.
Edit: and while you’re there, move the ‘speed camera ahead, is it still there?’ Dialog. IT COVERS THE DAMN SPEED LIMIT ICON.
And also while you're there, if no car ever in the history of your app goes down the road at the speed limit ever it's a good indication you'll never be able to ever do it at that speed. e.g. small narrow single lane country roads which are only theoretically 60mph roads.
Imagine the data they must have on the speeds people actually drive on every mile of every road, they’d easily be able to warn you not that you’re “over the speed limit” as in driving 70 on the freeway, but more usefully, if you’re in the top X percentile of speeds usually or even currently driven, which actually is a decent measure of unsafe was and would also be a great predictor of likelihood to get a ticket.
I saw this video recently where the author set up a camera to record sections of highways and measure the speeds of drivers, and make cool graphs out of it.
https://youtu.be/TYTaNsnBjcw
With the data you're mentioning, it's probably just as easy to build an accident predictor model as well.
Or sell the data to third parties instead because that wouldn’t bring profit.
I share your intuition that your likelihood of getting a ticket is related to the speed of other vehicles. Presumably, police choose to prioritise ticketing the worst offenders when there are too many offenders to handle.
But I don't share your intuition that safety is also relative in that way. If you're driving dangerously (too fast, or while drunk), you're driving dangerously, even if everyone else is driving dangerously too. If you're in a country where nobody wears a seatbelt, it's still prudent to wear a seatbelt, just as much as in a country where that is the norm. I don't think Google Maps should encourage people to drive as dangerously as everyone else. Quite the opposite!
"dangerously" and "fast" are not synonyms (while drunkess obviously is dangerous regardless). The 280 in San Mateo County is designed for 80MPH, with banked turns, gentle curves, etc. The speed limit is of course 65. If most people are driving 80-85, and you enter that highway and drive 60, which is a perfectly reasonable speed according to the speed limit, you are much more likely to cause an accident than if you just drive the speed of the people around you.
It's fine to point out that many people are terrible drivers and that a given crash that happens is more dangerous at a higher speed and if everybody were to drive under 60 at all times we'd all be safer, but clearly that will never happen unless we install a totalitarian government, put governors on all cars and give prison time for disabling them, with enforcers stationed everywhere to monitor. But no democracy would vote for that, so I don't think it's worth spending much mental energy on such hypotheticals.
> encourage
I think just the opposite: My feature would encourage people like me, who drive a "fast car" and can occasionally accidentally go too fast if a road is especially uncongested, to slow down to the speed that is customary or that others are driving, by reminding me that I could get a ticket and that driving faster than everyone else is dangerous.
Relative speed is important though. Where I live it’s completely legal to drive half the speed limit, and people often drive above it. People driving a consistent speed would reduce lane changes, breakaway traffic etc. imo speed consistency is what’s valuable, not preventing upper outliers.
When I was traveling in Mexico, it drove me nuts that even though I was signed in, Google Flights switched my currency from dollars to pesos every single time I opened a new tab! I think they really don't care.
You know what's even more annoying? Google Maps app on the iPhone uses the local currency for hotels and it doesn't let you change it at all.
Oh, "local" as defined by your IP too, so enjoy your VPNs.
The only solution is using the website instead, it has a currency dropdown.
This is super frustrating when you travel a lot. You can change it back to your preferred currency, but it doesn't really stick.
I think they rely on ip for a lot of stuff they shouldn't. Getting a local esim switches me to km until I switch back to my old one. Have no idea about Australia.
Edit: after typing this realized this isn't ip, its provider. That maybe does make sense to cue off of.
You know what’s funny. The browser sends an Accept-Language header that they completely ignore.
When I worked at Google; 10y ago!; I used the internal googler feedback form to open an issue for this bug. No replies.
Every year I fill up the feedback form on Google map to complain about this bug. Some years, I even did it twice. For good measure.
This bug is shameful at this point.
Also accepting gmaps work, if only it could preemptively cache the return trip for any trip longer than an hour, so that I'm not stuck with no service trying to remember how I got there.
Suggestion to use an offline app that will save you from such surprises.
https://www.comaps.app/
I clicked the link with nothing better to do, and woa, that's a really good maps app. Like, "I haven't seen something like this before" levels of good. Reorients itself with my phone, accurately - and in real time, not after I'd already walked ten steps in that direction. With other maps, I thought that maybe my phone's compass is broken. The default but optional 3D top-down view is the most comprehensible-looking map I've seen in a long time. It barely uses disk space, and going by my short experience with it, it really is very light on the battery.
Exceptional, this is what I'm using from now on. Just hope the iOS 15 support is maintained, that's a killer app to keep perfectly good devices productive even after they're restricted from everything else :)
Happy to have made you discover something cool :-)
> 3D top-down view is the most comprehensible-looking map I've seen in a long time
The 3D top-down view with building heights is possible thanks to OSM providing this data, thanks to people improving the map with apps like Street Complete and Every Door. Make sure to check them out :-)
There's with this 3D view though: it sometimes hides streets. I usually end up disabling it at some point.
Woah. What a gem you have provided us here.
Download the map of the area in Google maps.
Wait, there's a setting for this? I've lived in Australia for over 16 years now but everything is still in miles instead of Kms and I have never been able to find a setting to change it (although it sounds like even if I did find itz it would be mostly useless).
Ive lived in Australia for 45 years, everything is in km.. never had to touch it for miles. However i did go to the US and it showed units in miles on my phone which made no sense.
In Gmaps, Tap your profile picture, then select "Settings" and "Distance units". Choose between "Automatic", "Kilometers", or "Miles".
Pick the units you want.
I got the app wrong, it was Uber. I have my phone set to English (UK) and change the measurement system to metric. Uber doesn't respect that though, so it keeps using miles.
That's in the mobile app. In the web browser, there's no such setting. Or at least, none that I can find.
Bottom right, left click on the scale.. |----------|
It "seems" persistent for me.
It could be using your phone/browser language settings; try English (Australia) rather than English (United States).
Ah, thank you, this was it. I had the language set to English (UK), but changed the distance setting to KM. I got the app wrong, it was Uber, and Uber doesn't respect the override, so it always uses miles. Changing it to English (Australia) and Uber switches KM.
Also the currency in Google Flights... It always defaults based on your IP geolocation.
Defaulting to local currency is good since people usually book in local currency, putting switch button at the end is really bad though.
People tend to book using the currency they're familiar with, not necessarily the local one, so remembering their preferred currency makes sense. As someone who travels a lot but likes to see flight prices in USD, this is really annoying.
Another Google Maps request: Places I‘ve labelled used to at least sometimes (if inconsistently) show up in search. Now they never shop up, even when I type the exact name of the label.
Another issue with Google Maps is it not showing Plus Codes for some locations that highlight the entire area. If you however place a pin on that location, it provides a Plus Code. Pretty stupid IMHO.
Also it is really, really hard to search for "Nearby" places. Have to do it through "Directions". Really bad UX.
You can put "... near <location>" at the end of your query to get nearby places. "... near me" also works
Google Maps core functionality is sort of in maintenance mode, and things have been slowly bit rotting over the last 3-4 years.
Unless you want to launch some AI feature (used to be chat app for ten years and then Google got bamboozled by ChatGPT…) you’ll not find allies and your career will not progress.
I could take a quick look, though I’m not so sure I’d even find the code.
Your comment does not constitute a bug report, however. At a minimum:
- Are you logged in or out?
- What browser?
- What country is your profile set to?
- What country are you sending requests from?
While you're at it, could you also change YouTube so that captions being on/off is a per-language thing and not a "The user turned off subtitles for language X, therefore it's the only they speak, therefore let's turn on subtitles for all other languages." thing? It's really annoying that whenever I watch a video that has a language different from the previous video I watched, the first thing I have to do is to turn off the captions.
This isn't a proper solution, but if you watch with yt-dlp+mpv, you can configure default audio and sub languages in mpv (globally) and they will also work for YT videos in addition to your movies and such. Plus if you do toggle them on/off for one session it won't mess up your mpv config.
> I must have changed that back from miles once a fortnight since Google Maps launched 20 years ago. That's 500 times. Totally ridiculous for a company who core goal is profiling their users...
Similar with "privacy popups" everywhere. Similar with every bank with "remember this device" feature. I add exactly the same device on every login, until it fills entirely the limit of allowed devices.
If you can get anyone at Google to care about that you'll be miles ahead of anyone else.
Anyway, how many metric hours are in a fortnight?
How about allowing you to use Google maps while navigating without two phones or opening Apple Maps as well?
What do you mean?
If you are following a route in Google maps in Apple CarPlay you can’t search in the map without cancelling the navigation. So you need to use another phone or another maps app.
Ohhh
Yeah it makes sense I haven't encountered that since I don't have CarPlay
It would be a very useful feature as on road trips you can have a passenger setup the navigation rather than the driver and then also look around and examine places along the route.
This, but with 24 hour clock.
And language. I’m so tired of seeing the map in Portuguese because I went to Portugal 5 years ago.
The same here. Oftentimes Google Maps switches to most exotic language "because of reasons". Then I set it manually to my preferred language, for decades now, and it's been the same language all the time. Are they trying to say with poker face "we don't track you"? I don't think my home-grown privacy habits are that good, somehow they manage to show me an ad of "blue sneakers" someone in the household mentioned this morning.
thinking google would ever care about improving ux ever again is hilarious
It’s true, except expand that to all big tech companies. The only time UX is changed it’s either to make ads more effective or to “streamline” things by shoving more and more of the functionality into an endless nested chain of ••• and More menus.
It's hilarious to see the old joke actually playing out in real life. Kudos!
On a meta note; would you consider adding a left margin to your site? Reading from the very edge of my screen feels somewhat strange.
Clicking through the links in his article, I came across a guy who apparently did the same thing at Apple - he introduced the "auto remove" feature for expired passes added to your wallet, then promptly quit. I had no idea that's how that feature came about, but now I'm going to send a little mental thank you to him every time I get off a plane. That shit was FRUSTRATING.
That reminds me, I need to apply for a job on whichever team hasn’t added a toggle to remove contact names from the autocorrect dictionary…
Crazy because I remember that the first few iPhone OS versions had really bad autocorrect dictionaries, especially for German. The workaround for that was to make contacts for missing words because contact names never got marked as misspelled.
Another good Apple "employee" story: https://www.pacifict.com/story/
https://www.youtube.com/watch?v=GMyg5ohTsVY
Oh wow. Guess I need to get a job at Apple just to add a `Mark all as read` button to voicemails.
I need to get a job at Apple to stop “omw”-> “On my way!” (Complete with the `!`) from reappearing in my Text Replacements every month or two, no matter how many times I delete it.
(Try typing “I’m omw to the car” or something to see how annoying this is)
Apparently the current state of the art to fix this problem is to remove it and add an "omw" → "omw" text replacement in its stead. A friend recommended this to me and I haven't had problems with it since. Yes, it's nuts, but it is what it is.
And the ability to undo deleting voicemails. And record voicemails client side using AI transcription to deliver thnesss
Meant to say “to deliver the same experience as Visual Voicemail” but accidentally posted it from my pocket before I could finish writing it out.
> he introduced the "auto remove" feature for expired passes added to your wallet, then promptly quit
This still didn't work reliably, unfortunately. I still have expired passes, tickets etc. in my wallet
Personally I don’t see why you would want to delete these expired passes. For the longest time they haven’t shown up alongside active passes.
> For the longest time they haven’t shown up alongside active passes.
They have always shown up for me, and the only way to delete them is from the wallet app. Note: from the app. You can't delete it from convenient screen where you access them
Well, now you know the drill!
BRB polishing my CV
Hotz said this, but I couldn't find any actual evidence so didn't include it.
Maybe you should get hired by OP and fix it yourself ;)
Site is actually open source lol - https://github.com/skeptrunedev/personal-site
>Get hired by github > force push the pr > get fired > profit
This is dark. Get hired at GitHub and you can even fix bugs in private repos! This is the real move.
I am not a fan of sites that waste screen real estate.
> On a meta note; would you consider adding a left margin to your site? Reading from the very edge of my screen feels somewhat strange.
What!? I love the fact that it's left-aligned. That's the way text should be!
Alignment and margin are different concepts. You can be left-aligned and still have a comfortable margin.
True enough, but considering most websites these days consider "a comfortable margin" to mean "4-6 inches", I'm delighted to see a site which actually lines things up close to the left side of my monitor. Like I said, that's how text should be.
Unless you have a wide screen display and have to physically move your head to read the text.
Comrade, we have invented tiling window managers. You can simply move the window
E: I still like margins though. A visual delineation from the window border is nice.
> I added an AbortController to the debounced search function, so that it aborts any previous queries when a new one is made. This means that the search results are always relevant to what the user is currently typing.
To me one of the most annoying things an application can do is go off and do something before I'm done telling it what to do. Filters that apply themselves without an explicit indication that I'm done setting them up, or searches that are constantly re-executing as I'm typing. Wait for me to stop.
When I implemented search-as-you-type on my blog, I decided to wait for the current search suggestion request to complete before doing a new one. Seemed like a reasonable balance between responsiveness and not overloading the server.
When I implemented one I simply filtered the results that came back from previous requests to see if they matched what the user had typed in the mean time. That way the UI might get relevant results with lower latency than would otherwise be possible with no risk that a non-matching result would show up to confuse the user.
The video showing the problem in the article seems to show an avalanche of queries towards the server, I'm surprised no one cared about it, but I guess it's frontend people thinking "It's the backend/ops that have to deal with the problem!".
I wait about 250 ms before firing the request, if the user (well, me) continues typing, then the timer gets cancelled and the app waits another 250ms.
The earlier search will be on a shorter input which will find more results; imagine waiting for everything that contains "e" before you can get the one result for "europa" that you wanted. Isn't that pattern going to happen most of the time - waiting for a large amount of useless generated and transferred and disposed-of, before the search you want will run?
That's why I set it to run after the third letter and only return up to 7 suggestions. In your example, it wouldn't run until "eur", which should narrow it down quite a bit. Given my rather small blog article corpus and aggressive full text search indexing, it's very fast.
Firing queries all the time is especially annoying if your users are in another continent, and you don't have proper state management to only show results for the latest query, as opposed to the latest response.
RTT from Europe to AWS us-* can easily get to multiple seconds during peak times.
I hate this on booking websites. Especially if the filters are in a sidebar on the left and do not fit your viewport and every god damn time you change something it scrolls up, starts loading, puts filters into read only mode until it's done just so you can add the next filter...
Grafana log search does this if change the currently-applied log filter, and then they charge you for search volume!
I've had to adjust my UX usage so that I don't get billed for every character I type, rather than the string I'm looking for.
I think a good middle ground is to wait a few hundred ms at least, for the user to stop typing, before sending off the query. Or maybe, still send the query, but don't populate results until they stop.
It is loathable and contemptible behavior. I despise it. Writing in fancy code editors is an exercise in "remember this display was very expensive calm down".
> type out 'i' and 'f'
RED ALERT! RED ALERT! UHHHHH HEY THERE CHAMP YOUR IF-THEN STATEMENT ISN'T CLOSED OUT PROPERLY!
Yeah, I know, I'm not done fucking typing yet. Give me a second.
Thankfully I can usually turn this crap to only show these after a certain delay, for many languages I write in I usually turn it off completely except for on build/run. Let the complier/shell throw me errors, the LSP can take a chill pill.
Datadog if you see this
The software quality is so low that if a bug bothers you, it's easier to get hired to fix it than for the company to fix it! Wow.
It reminds me of the programmer who mitigated the GTA 5 loading time problem. If even with a lot of money of GTA 5 the quality doesn't improve...
It's nothing to do with quality. It's prioritization. Companies pick things they think people want (or what the team wants to build) without testing or experimenting with users and data.
This is actually an example of the problem, not the solution. There are probably much more useful things for the team to be doing but they let one guy add the thing he wanted.
You'd think search on a docs site would be important :D
It's because the people who spend the most money on the scam cards for GTA:O didn't care. I personally stopped playing the game after getting so fed up with the loading times, I timed it for an hour and found out I was spending more time watching the loading screen than I was actually playing missions.
The article says nothing about the hiring, which is kind of the most important part of the whole escapade. Right now, it's a bit "something was bugging me, and when the company hired me, I fixed it", which, great?
I think his company is acquired by the currently he's working in, so he's acquihired.
exactly
Congrats!
> It reminds me of George Hotz’s legendary single week at Twitter in 2022, where he joined just to fix a login popup that was bothering users, then bounced.
The author remembers this, uh, event differently than I remember it... George Hotz boldly claimed that he could "fix Twitter search" faster than those lazy Twitter devs, only to bail almost immediately. Hubris!
On the way out, he removed that login popup as a sort of consolation prize.
Well, that's a bit of my time gone (re)looking into GeoHot, patent trolls, and now comma.ai.
Comma.AI by George Hotz sounds very interesting, it's basically a $999 "comma 3x" smartphone with an OBD-II connector and a $99 wiring harness that can add an equivalent of a Tesla Autopilot to many cars manufactured in the last 10 years (even Tesla's own cars, too), for a total cost of $1098, whilst being OSS and available on GitHub, and — get this — even having ssh access to your car! Optional cloud subscription plans are $10/mo for your own SIM, or $24/mo with bundled cellular data.
Sadly, it does NOT have an equivalent of Tesla Sentry Mode yet, https://github.com/commaai/openpilot/issues/29912, which is kind of unfortunate, because Tesla's own implementation of Sentry Mode is using 250W of power — depleting the entire 80kWh battery from 80% to 30% in like 7 days (".5*80kWh over 7 days" = 238W) — openpilot would have been a nice alternative at what'd presumably be around 5W or less ("40kWh / 5W" is 333 days).
Yeah, what? He seemingly joined Twitter, did fuck-all and quietly bounced. Embarrassing and completely self-inflicted.
Updated the post to tell that story more accurately. Simultaneously took down the damn blog because Github pages has some freak bug, but oh well.
Time to join GitHub
TRUE haha
Please ask them to add IPv6 support if you do.
context: https://github.com/orgs/community/discussions/10539
And then he was trying to pitch rewriting it from scratch to elon
I followed the whole saga on Twitter. He shared a video of his browser saying, "I fixed it in 5 minutes," and 5 days later he was still trying to figure out why his PR was failing the build. When Twitter engineers told him to write tests, he rage quit.
It was embarrassing to watch.
In the reverse situation, I've worked at places where the IP lawyers basically made it impossible to submit PRs to open source code.
But sometimes explaining the exact inputs and the line number where you know the problem is can grease the wheels enough that you can convince someone else to write the fix for you. I didn't technically give you any code. But I did give you free QA.
I specifically attempted to get a job at Discord so I could submit a PR to make giant emojis be a toggle setting rather than automatic. I know the feeling.
(If anyone works at Discord, please me and the rest of my server are begging you)
I don't know if this helps but I've been adding a full stop next to emoji exactly for this. It doesn't fix it for ~new people but it's something for yourself?
This is what I’ve been forced to do for years and I’m sick of it lol
Discord is an electon app IIRC so in theory it should be possible to make a client side mod which fixes this. Not sure if that would result in your account getting banned though...
What are you referring to? You can disable the automatic conversion of :) to an emoji, for example.
I think it refers to only sending an emoji, it's shown as a large icon.
This PR is quite the PR move.
I wonder if it's legal for corporations to have employees that they send off to get hired at other companies, do some stuff in those companies that are beneficial to their actual employer, and then leave before the probationary period ends.
IANAL, but it’s almost certainly legal, as long as all parties involved adhere to the applicable non-disclosure agreements, non-compete agreements, and intellectual property provisions of their employment contracts. Even then it’s likely to remain a civil matter in most cases.
Companies can sue each other for nearly anything, so any level of this behavior could result in a lawsuit. It wouldn’t cross the line into criminality until it involved some fraudulent deception or blatant corporate espionage. For a recent example of that, see the ongoing litigation between Rippling and Deel. (But even that egregious espionage activity remains limited to civil court, at least for now.)
"to have employees that they send off to get hired at other companies, do some stuff in those companies that are beneficial to their actual employer, and then leave before the probationary period ends."
To me that sounds like not disclosing, that they work also for another company and this certainly ain't legal on most jurisdictions.
Can you cite the relevant law? I've never heard of it being illegal in the US to not tell your job that you have another job.
It's probably not the law (it would be shitty when working at a 7/11 on the weekends to have tolegally disclose all your other income resources)
But basic employee contracts cover these aspects, including working in the interest of the company and IP assignments, and usually exclusivity if you're full time.
These issues are old as time.
Yeah I'm aware employment contracts might stipulate it. But violating a contract isn't against the law. Worst case you could get sued (though with an employment contract, the limit of repercussions are generally just termination).
> violating a contract isn't against the law
Being binding is kinda of the whole purpose of a contract. If violating it is void under the law the company should change lawyers.
To put your argument under a different angle, there are many written laws you can violate with very limited consequences if any, but they are still laws.
Contracts aren't written by the country, and enforcing them is civil matter so there's nuance, but violating an enforceable contract you provably agreed to is against the law. Whether you can get away with it is another question.
There are two types of law. Contracts are civil law. Breaking them does not break criminal law. Civil vs. criminal law has different procedures, different burdens of proof, and different potential consequences.
When it comes to contracts, no, there are no "laws", there are agreements between parties that can be enforced if taken to court, and in that sense they are binding. But breaking them does not break any law... it just breaks an agreement.
It's not that simple.
https://www.parzfirm.com/blog/when-does-breach-of-contract-b...
> When Does a Breach Become Criminal?
> For a breach of contract to rise to the level of criminal activity, the act must involve elements of fraud, intent to deceive, or theft. These cases go beyond simple noncompliance with contractual terms—they involve behaviors that violate state or federal laws. Some scenarios where contract breaches may involve criminal activity include:
> * Fraudulent intent: If a party enters into a contract without any intention of fulfilling the terms, this may constitute fraud. For instance, accepting payment for services without any intention of delivering.
> * Pattern of deceptive behavior: When a party repeatedly breaches contracts with the intent to defraud others or engage in fraudulent schemes, it can elevate the breach to a criminal offense. A pattern of deceptive behavior indicates a systematic intent to deceive and defraud, which may result in criminal charges.
> they involve behaviors that violate state or federal laws.
> elevate the breach to a criminal offense.
Sounds like it is that simple. If you break a criminal law, then it breaks the law. Otherwise, not.
The goalpost is moving.
2 posts before:
> But violating a contract isn't against the law.
Now:
> Contracts are civil law. Breaking them does not break criminal law.
and what if you don't work there or have a salary but happen to own some equity?
Not really without researching(also I am european and might have assumed wrong about US), but something with conflict of interest? Especially if another company ordered you to work for someone else. If all is disclosed, probably fine, but undisclosed? Definitely would not work in europe. Breach of trust etc.
Not sure it falls foul of broader laws, but it almost certainly breaches your employment contract, which likely includes something about following the policies of your employer; that policy (in many companies you likely have to go through onboarding training and annual refreshers on it) probably includes a code of employee conduct that has specific mention of conflicts of interest.
You'd achieve more by simply telling the company that you need a certain feature added to their product. If you're an important customer for them, you could probably negotiate a price for them to prioritize the work.
I think we'd probably better off with the previous idea: just work for them for a period.
I'm speaking from the perspective of company A, who needs a feature added to company B's product.
They could send their engineers to work for company B, sure, but those engineers' time is still costing money. And those engineers are completely unfamiliar with B's codebase, so they won't work as efficiently. Might as well just pay company B directly for the feature work.
Sometimes Company A is better equipped to fix things in Company B's product than they are themselves.
I often have the thought that it would be pretty awesome to take jobs for 6 months here and there just to implement specific features I want in my favorite SW, apps, sites, etc.
* Join Logic Pro team for 8 months and add better score notation tools
* Join Apple's iOS Music app and fix the weird blip that happens at ~17 seconds on any track
* Google Maps to stop the navigation/directions from spelling out how to get from my house to El Camino Real, which I've only done about 10,000 times.
* ...
Howdy neighbor.
The cancellation in the denouncing seems … sort of obvious.
It seems obvious, but I also don't think it's optimal depending on what you are trying to do.
In some scenarios, e.g. bad mobile internet connection, you may also be happy with a slightly stale result (where you still have to ensure correct ordering of responses), depending on how the search overall is implemented.
One additional data point: Algolia used to do query cancellation in the past but stopped doing that (I think at least 5 years ago now), which you can test with the HN search. I'm not sure about their reasoning, but for them that seems to be the best overall default search experience.
yes, i was very annoyed
Good PR, but AbortController doesn't really help with stopping the server from processing the request. I have seen so much of this type of search that just continues processing in the backend even if the client has long gone caring.
I agree. I was expecting the author would implement debouncing or throttling to reduce the number of unnecessary requests, alongside fixing the data race issue. Here's an excellent article on debouncing and on throttling, and the difference between them:
https://css-tricks.com/debouncing-throttling-explained-examp...
There's nobody who really has a library that's set up to feed a sequential task into and have it force a synchronous call to be async with breakpoints to check for early termination.
This seems like a problem Sorhus should have a library for, but he does not.
I've had the conversation too many times in the last couple months about how setTimeout() does absolutely nothing to fix this problem in NodeJS. Even Java had trouble with this and tried to delete the API that seemed like it should support this problem, due to undefined behavior.
The sever should cancel the request handler when the client drops. Otherwise you're just opening up to accidental DOS.
There was an old legend for an Apple bug (but I can’t exactly remember what). He complained about this macOS bugs for years. Worked for Apple for a couple months, fixed the issue, then quit.
could you please remember it exactly, because I am very interested into this legend hahahha) thanks
Who hosted the backend? If you did, you could demand a fix, as the bad frontend code caused increased server load and wasted resources.
https://archive.is/ntSHm
Problem solved, so... time to move on?
I thought about it, but nah. Really enjoying the new job so far
So that’s the only hope to get MRU tab switching in Chrime - get hired at Google?
Congrats on joining Mintlify, friend. Trieve is dead, Long live Trieve!
That's one way to do it.
This link is a 404 for me
Fixed! Damn Github pages
So you are the guy from that meme
very funny watching serial apple ass kissers getting tricked into listing all the shitty things in this thread
FYI this autoplays full screen video when I visit on iOS + Firefox.
Edit: then switches into dark mode after a lag of a few seconds
autoplay my fault and will fix
dark mode idk, that is a very tiny piece of JS which should run near instantaneously
A debounce fix. They were really asleep at the wheel.
wasn't there an xkcd like this once?
Code is always the best documentation and the best thing about opensource.'
Code will tell you what but not the why. It also doesn’t always tell you the intent.
They should invent a programming language that only compiles if the why is still true.
They have, but they're beyond grasp of most developers.
Tests were invented to express the "why" for the normal guy. They don't strictly prevent compilation, but a proper workflow will see them halt your process in the same way, offering the same outcome.
Granted, there are a lot of horribly written tests out there that don't tell you "why" — or, well, anything. As always, people will find a way to abuse anything you put in front of them. But when used well...
With a test, it might link up some functionality with "why" and pass, but then what happens if a business requirement just isn't a requirement anymore? The test will still pass. I'm thinking of something sillier, like a language that forces you to justify why for your code, and then regularly quizzes you if the business reasoning is still true. If anything changes, it rips out the code and breaks your site. :) So then you have to go in to fix it.
I'd also love it if this were applied to politics and laws.
It wouldn't be too hard to add such logic to your tests. If it proves useful, someone will no doubt turn it into a language feature.
Good commit logs or comments may tell you why
What about function names, class names and variable names?
Helluva wish.
Having the source lets you fix something for yourself, there are an increasing number of barriers being put up to prevent you submitting a fix upstream.
Going through this right now with part of libpng, their mailing list doesn't seem to like my email.
Using a source-based distro (previously Gentoo, now NixOS) lets me solve the problem for myself, even if my PR never gets accepted. Right now the count is at 4 patches in software I use that I submitted upstream that were (for one reason or another) never accepted.
In at least one case, I later found out that I was not the only person to submit a fix for the problem I was running into, but their discussion on the ML also went without comment 3 years earlier.