Hey, I'm the author of the blog post. Thank you for submitting this. If you have any questions feel free to ask and please let me know how the writing was. It's one of my first posts so I'd like to improve
You should change the license to AGPL and 'custom, contact for payment details', and provide a link to this as why you did so.
Simply put, anything not a viral license like GPL allows parasitization by companies effectively living off FLOSS devs, with absolutely nothing to gain. Human rights under GPL were meant to apply to humans, not '3 lawyers in a trench coat' (corporations).
They can make their decisions (snubbing a dev of code they deem good enough for enterprise). And you can make comparable decisions, punishing them for the sheer hubris.
It also reaffirms that my decision of AGPL for everything is the right one. They can contact for custom terms.
> It also reaffirms that my decision of AGPL for everything is the right one. They can contact for custom terms.
Since your replies below are focusing on compensation: have you actually made a nontrivial amount of money with that model?
I would expect that should be a prerequisite to reaffirm it was the correct decision, especially if you're giving unsolicited advice to strangers about how they should license their software.
Some people want others to freely use their software and choose MIT precisely because it's more free than GPL. There's nothing wrong with just making something for free and giving it away if that's what you want. Not everybody has to be chasing money in all their activities.
The author said he was proud of this outcome and nervous at how widely his hobby project will be deployed. That sounds like the ambition of many open source authors and a win. Might never have happened with GPL.
> Some people want others to freely use their software and choose MIT precisely because it's more free than GPL.
MIT license is absolutely not 'more free' than the GPL.
In fact, MIT means you give up effective ownership and control. You lose control and contributions.
And what do you get for that loss of control? Exposure. Or, in this and many other cases similar, you get diddly shit. Some company paracitizes your code, sometimes even demands SOC questionnaires and 'do this bug NOW', and other abuse.
> Not everybody has to be chasing money in all their activities.
Talk about missing the point! This was all about money. It was about a job at the company where the code is being used in a production manner. And they didn't even bother to give an interview.
And not many of us are independently wealthy, and can do things that we want with no monetary care. And, most FLOSS devs aren't that. Instead, they're being used as unpaid stepping stones so some overvalued AI hypesquad can vibecode (or slotmachine programming) faster.
> The author said he was proud of this outcome and nervous at how widely his hobby project will be deployed. That sounds like the ambition of many open source authors and a win. Might never have happened with GPL.
That's where I hope the author relicenses as LGPL and proprietary, and doesn't give Anthropic any more free professional work.
And if it never would have happened with the GPL, gasp, they would have had to pay developers to create it.
And until I'm independently wealthy, I too will license AGPL. If you're making money on my stuff, I want a cut. Simple as that.
>MIT means you give up effective ownership and control. You lose control and contributions. And what do you get for that loss of control? Exposure. Or, in this and many other cases similar, you get diddly shit.
Isn't that what true freedom is?
You can argue that more freedom is a net burden for both the individual and society (tragedy of the commons), but that doesn't negate the aspect of it being more free to begin with.
>And not many of us are independently wealthy, and can do things that we want with no monetary care.
Indeed. But not many people contribute to any kind of OS community to begin with (regardless of the license). I would like to one day, but then the industry laid me and hundreds of thousands off in the last few years and those plans were delayed.
There definitely is a certain level of privilege in being able to provide knowledge to others on the side. Even morose if you're part of an organization that pays you to do so.
So, what is your net-worth, in that you fight for freebies to corporations? What net worth should I strive towards so I can be nonchalant and passe about money?
I'm certainly not there.
Also more curious, is the AGPL doesn't affect humans doing stuff. It affects companies when they grab, modify, and host and not share contributions. Read about anti-TIVOization. That's why the AGPL. I'm guessing you know this, and why you're attacking my viewpoints as 'missing the point'.
And yes, copyright is everywhere. And the GPL has some of the sanest terms to reuse, as long as you follow the requirement. And the GPL also further grows the ecosystem, due to virality.
But Anthropic wasn't exactly submitting code either, were they? In my world, parasites get antiparasitic drugs.
> So, what is your net-worth, in that you fight for freebies to corporations? What net worth should I strive towards so I can be nonchalant and passe about money?
I've seen people with un-stressed about money with net-worths that are orders of magnitudes below those that seem to obsess about it.
Your motivations are your motivations, if you don't like the idea of someone using your work to make money without giving you a cut, you can do you, but why is it hard to understand that other people might just not care that much about it (or, gasp, even find their work being used more rewarding than the potential monetary compensation)
> Also more curious, is the AGPL doesn't affect humans doing stuff.
It does affect humans doing stuff that isn't malicious, like if you need to solve a problem by modifying the code then now you also have to make that change public which is a hassle, I'd rather not have to track or maintain such things. I'd rather not have to think about that, and I care more about such nuisances than I care about the possibility of companies stealing it.
The scenario you are describing (discovering a problem to fix, being able to fix it, but then not sharing the fix with other people) is the exact reason why GPL has been invented: to force people to share their work, so that we can all have better software, together. Maybe the software you are using wouldn't have been that good if other people weren't forced to share their improvements. Your small effort is going to help others, and their small efforts are going to help you even more. This mindset of sharing should be natural but, as you just proved, people are lazy and so the license has to force them.
Cool. Thanks for the link, but I wasn't actually trying to steal anyone's thunder, and ... I did read the article. Just felt that it wouldn't hurt to link to it.
Also, that discussion gets pretty mean. Didn't feel like I wanted to send people there. I just wanted to give the guy a pat on the back, and bring some humor into it. Been there. Sucks.
I don't know--As a non-celebrity tech worker, it's actually kind of comforting to see a company that doesn't just automatically roll out the red carpet and grease internal wheels, just because a candidate once wrote some very popular software. It sounds like there are still companies that make you go through the same whiteboard hazing and non-deterministic hiring process as the rest of us mere mortals, regardless of how well known you are.
Actually, I don't disagree with some of the people that had issues with him, but I do have issues with folks that refuse to look at past performance, in general (I'm biased. I have a great deal of past performance, and can prove it).
It was just a kind of nasty conversation, and I didn't feel that it was appropriate to deliberately send folks there. I'm not really into the whole "Make the Internet Darker for Everyone" schtick.
Hey, great work, and just wanted to lend my voice in support! It's kind of wild how many open source devs have a story along similar lines. (Mine is the time when Mojang used my voxel engine..)
Now that this is trending on Hacker News, surely there will be a happy ending when someone from Anthropic sees this post and hires you with sincerest apologies and everyone lives happily ever after? Can we get a positive story out of this, universe?
What's the risk? The current state is "your application has been thrown into a fire and will never be seen by human eyes". How can it get worse than that? There is no downside to complaining on HN, except for getting the reputation that you really wanted to work there, which, again, isn't that negative of a thing.
I don't know. I have no comparison but it is common for crates to be released under MIT. I took over the maintainership from the original author so the license was already there. I rewrote pretty much everything so I guess I could try changing the license now but that's not something I wanna think about.
I do the work because I see it as payback for all the great open source software I use all the time.
I really like the copyleft idea, however, I think you did nothing wrong, IMHO, because if large corps like an idea, they will rather reimplement it rather than even bothering with ways to conform to AGPL or buy an alternative licence.
Particular in the age of AI, all source available code has become pretty much public domain (value is still in maintenance, etc). License have mostly become a compliance/ideology game that alienates most people. However, changing the license on the main repo, with only a minor version bump, would be a nice asshole move to get their attention past HR (won't make a difference, but if you have nothing to lose).
Copyright is but one pillar of intellectual property law.
I’d like to see an attempt by useful freedom respecting software projects to deploy patents to combat non-free reimplementations.
A GPL license that grants you rights to the backing patent as long as the software you develop with it is also released under the GPL license.
Use the library for closed source software? Copyright violation. Reimplement the software under another license? Patent violation. Create something slightly different and call it the same thing? Trademark violation.
Not sure of the rest of the world, but at least in the US, patenting “software” is a pretty murky subject legally (at least it feels that way when trying to do some basic research on it) Something that seems common among sources discussing it is that “Software Related Inventions” (eg, a computer that does XYZ) can be patentable, but software/code itself is not literally patentable. Seemingly, because we’re talking about libraries that would be pure software, not a product for sale based on it, you wouldn’t be able patent libraries like you’re talking about.
I’d provide links to some discourse of this, but honestly I think it’s better to search “can you patent software in the US” and do a brief read of various sources, because the terminology between them can seem somewhat counterfactual to eachother.
Copyright mostly protects big corps nowadays. That's because you need lawyers to enforce copyright, and if the other side has more money the battle may not be worth it.
On the other hand, Meta was found torrenting terabytes of books and for them it's a nothingburger. The rules are really meant for commoners.
> but it is common for crates to be released under MIT
Something that isn't brought up enough in the "rewrite everything in Rust" discussions is that the API guidelines explicitly recommend MIT/Apache to "maximize compatibility" (i.e., corporate friendliness, or developer and user exploitation): https://rust-lang.github.io/api-guidelines/necessities.html#...
Your project has been around for a while, but it's crazy to me that anyone still open sources anything under MIT (or similar) in the era of LLMs. Are they that confident in their job security? Are they already independently wealthy? Frankly, even a proper copyleft license is likely to just be ignored, or the code laundered through an LLM-assisted rewrite, by these companies. I prefer to just keep anything I can't sell all to myself rather than release it, at this point.
Hey, I really liked the post and especially the title. Quite surreal but also very fitting at the same time. The writing was great too. Hope you keep going. I’d love to read more.
This lands. I discovered an emergent feature in GTP40 and when I tried to post about it on the developer forum, the spam filter removed my post. I asked GPT40 to rewrite it for me. I posted the update, and got banned. There's too much 'noise'. People like Einstein and Tesla would've gone unnoticed today, as I doubt they would've become "social media influencers" just to promote their ideas.
If we are at the point where a hiring manager for a position deeply related to an open source library is not at least checking if the authors would be interested, I'm not sure.
I honestly think this is some system failure, even a Claude based one. I hope someone in the Claude Desktop team sees this and reaches out to you. Cheers!
Hey mate, I would just like to say that I wish they at least find it in their hearts to reward you for the value you have provided to them. Knowing cut throat american corps, I'm afraid the chances are nil. Even if a good amount for you is peanuts to them.
Don't know. A company can have a huge valuation on the stock market but that does not necessarily mean that they have cash to pay wages or can afford to pay a large team. If all they have are stocks they have to find somebody that buys those stocks with cash, then find a way not to run out of those money before selling more stocks. Eventually do an exit and stop worrying or become profitable.
I have always preferred permissive over copy-left, because I've historically been unable to use packages at work, which puts food in my mouth, as a developer who spends some time contributing to projects, especially those that I use at work.
This has changed everything. AGPL and GFDL from now on.
you're right about MIT vs GPL confusion. people brainwashed themselves into thinking MIT is "more open", because it's more permissive, but it lets others profit off your code without contributing back.
GPL makes them share or pay to relicense, since you own the copyright. with MIT, they don’t need to ask. MIT just benefits big corps. GPL better protects the open-source spirit, and paradoxically, the ownership of your work.
And yes, people being able to use your code for whatever they want is absolutely more open than having restrictions on how/who gets to use it.
One other model that can also work well is to dual license as GPL + commercial, so people who want to publish their work can use the GPL license but you can potentially fund the project from license sales to closed source users using the commercial licensing option. I see this a fair bit in the audio community I work within.
If you link against GPL code, your code needs to be GPL compatible. There are some IPC based workarounds, but they are too annoying and slow in most cases.
>And yes, people being able to use your code for whatever they want is absolutely more open than having restrictions on how/who gets to use it.
Yes, this is why people should use free not open , and GPL is more free when you report to the entire community otherwise you are in the famous case from a story where an USAian was claiming "Amerika is the land of the free, we are free to own slaves"
Yeah, basically MIT is "more open" in the short term, while GPL is more open on the long term. GPL, while restricting some freedoms right now, is actually enabling the remaining freedoms to be sustainable in the future. Very similar to how law enforcement works out with regards to a sustainable society, and how market restrictions work out to create a sustainable and diverse market.
I'm disappointed about your resigned, almost subservient tone. This company is profiting immensely off of your work, and they don't even give you the courtesy of a job interview?
~~Have you considered a copyleft licence like LGPL?~~ Answered in a sibling comment
> This company is profiting immensely off of your work
I wouldn’t say that’s exactly the case. Not to denigrate the author or anything, but this library is a relatively minor part of what Anthropic is doing. It’s a UI manipulation library, specifically one that simulates keyboard and mouse inputs. While something like that is certainly necessary for the project in question, it’s not anything that couldn’t be rewritten in-house without too much difficulty, especially since they’re only using a subset of the platforms supported by the library.
I’m sure that working on this project has provided the author with expertise in this area that Anthropic could benefit from, and so in that sense it’s still a shame that they wouldn’t give him an interview, but that’s really all that can be said about it.
> it’s not anything that couldn’t be rewritten in-house without too much difficulty
This is my experience, at every group I’ve been in. Extending the date a bit is much easier than involving legal for approving a new library.
The group I’m in now sunk a substantial amount of money into a lawsuit for a library that accidentally made its way in, so are now “No LGPL.” with some crazy loops and approvals required if there’s really no alternative (very rare). From their perspective, it’s cheaper and safer to rewrite than not be in compliance, unintentionally or not.
You have to think about other users as well. One person taking advantage of you doesn't mean you have to cut off all the people not taking advantage of you.
Expecting a reward from open source software is a recipe for disappointment. I have contributed code to projects by companies that say I'm a mentally-ill household object. I'm not going to change the license of my open source projects to get back at them, because the collateral damage against entities that aren't evil simply isn't worth it. (It's also somewhat unlikely that the people working on NTP servers at Facebook wrote those policies, so...)
> Through a friend of a friend, I found out that Anthropic had an open position in the team implementing the secret, unreleased feature of Claude Desktop using enigo. I wrote a cover letter and sent out my application. An automatic reply informed me that they might take some time to respond and that they only notify applicants if they made it to the next round. After a few weeks without an answer, I had assumed they chose other applicants.
I've mostly stopped applying to the big companies long time ago (10+ years) precisely because I'd never hear back regardless of the match or the credentials.
The only exception has been JaneStreet — they've contacted me almost immediately after a cold application with a small cover letter about my interests.
Yet going the referral route, it's relatively easy to get an interview almost anywhere, even Google or Apple.
In my 6 jobs in my career from college internship to startups to Big Tech, I have never gotten a job from sending an application into a site. It's always been through (somehow) tracking down a person to speak to and get a referral.
I would recommend that here. There's no reason why Anthropic couldn't use your talents! See if you can find a friend-of-a-friend who is there, and then do a phonecall with them.
They put an absurd about of effort into their recruiting, so I'm not really surprised. Pretty much every math-related content creator I listen to has advertised for them.
I doubt anyone who works there is "take what we can get" calibre. They want to attract people who casually solve college-level math puzzles for funsies. So I imagine it's the opposite and if you get hired there, you're surrounded by people who are extremely accomplished.
Yes and no, I imagine the biggest qualification for Jane Street would still be humility and not OCaml interest or expertise, and the pay probably has something to do with people's desire to apply, too.
I wasn't talking about why people would apply, I was talking about the statement that they responded fast to a cold application. Generally speaking, if you drop a ton of applications for what I assume is a lack of willingness or skill to deal with OCaml, that certainly factors in your hiring process.
Lets make a new license: If you wont hire me, use my library and make over $100m in revenue a year, you must pay a commercial license to use my software equivalent to the total cost (equity grants included) of an average principal engineer or director who manages 50+ people at your company in your highest COL metro, whichever is higher. For OSS work that isn't mostly one author, make it go to the foundation for the OSS project instead and apply the rule to principal maintainers. You could even scale it in multiples of revenue in principle engineer units of $1b per principle engineer of global revenue.
IMO I think foundational projects that every single bigtech uses like ffmpeg should get on this licence yesterday. They would start getting millions because it still would be way cheaper than making it themselves in their bloated cost structures.
I agree with the spirit of this comment, but I worry about the implementation.
See the comment of Manly read in this section. Once the threat of payment approaches, you can just switch to a free fork. A single person can't really win a trial against a big, well-funded company.
I don't really see how this is an issue, depending on the license text it's trivial to make the license apply in the same manner. As for winning, I think that's more of a US-centric view, if you sue elsewhere in the world there's plenty of courts that are happy to slap big tech.
I think Jacobsen vs Katzer [0] is the most relevant one to the discussion here, but there a number of successful cases on this front. If memory serves, BusyBox has also managed to enforce GPL in court on a number of occasions.
Suing a major corporation still seems like quite a bit of work, and what's the end goal? Is it to humble a major company, or to get paid? Because if it's the latter, it feels like there are easier ways to do so.
This is obviously a subjective opinion, but at least in my mind, the point is to defend your rights. No one else is going to come along and defend you against the corporate steamroller.
They can fork it, but can they find the maintainers? If it's just their own internal employees, then they definitely have less expertise in that codebase.
That would just open the door to commercial competitors to undercut the price by reverse engineering it (https://en.wikipedia.org/wiki/Clean-room_design). You'd be playing on the proprietary software industry's home turf and they will straight up curb stomp you.
It’s a nice idea but couldn’t a big company simply move its engineering team to a subsidiary that doesn’t get sales revenue?
(I’m not an accountant!)
Would be hilarious to bury a clause like “Modified MIT license — head of HR must publicly announce any employment application rejections of the maintainers while wearing a chicken suit).”
Because
- They can decide more easily what he works on
- They know he loves this work and is very capable of doing it
- They can own his output, a competitive advantage
- He will likely cost them ~nothing anyway
This is an incredibly short-term thinking. The reason is simple: the author is not obliged to continue while this sort of thing can be demoralizing.
I don't know about the author's approach to this matter, but if I would find out that a company is making a killing using my software and then that company would refuse to even give me an interview I'd probably stop loving doing what I do. Sure, the software is under MIT license and it was the author's choice to do so, but what's the point of doing it under such a license when you can't even count of it mattering in a resume? What's the point of providing free labor to a company with revenue in billions? If you look at the author's blogpost, the only benefit the author mentions is making the number of downloads go up and that's just pathetic.
I am reminded of an another, similar case with a library called "FluentAssertions". This library used to be free to use by anyone until the author changed the license and started charging money for commercial use. The author did that because he spend several year maintaining the library on his own time and dime and megacorpos like Microsoft wouldn't even bother to donate despite using it extensively. What happened afterwards was that the author got shat on by everyone on the internet for daring to ask for money. In the company I work for his library has been replaced with an another free fork at a incredibly fast pace. All that free labor and the author got dropped as soon as they fell out of line.
The worst thing is that it wouldn't probably take much to make the author of the library happy. Even if they weren't interested in hiring him they could still acknowledge him, talk to him a bit to maintain good relations, throw him a nice donation as a thank you and now it would be a nice, good PR story instead of an another reminder that corporations are just looking to squeeze out value out of all of us.
Sorry but have to call b.s. here. Many of us did in fact, in this very forum, kept pointing out that dumping on GPL (and the man and his ideas behind it) was a mistake and that non-gpl oss was for chumps. And we were greyed to oblivion. Same exact story as with surveillance tech.
Hackernews is hugely responsible for many of the ailments of this field in 21st century. Own it.
> Many of us did in fact, in this very forum, kept pointing out that dumping on GPL (and the man and his ideas behind it) was a mistake and that non-gpl oss was for chumps. And we were greyed to oblivion.
I'm not entirely sure that was intentional. On Reddit, it would be called 'brigading', and basically getting your corpo-techbros to -4 and flagkill posts.
If done fast enough, you only need 5 500+ karma accounts to sink a post.
Sometimes, I'll say something unpopular, but defensible. Its interesting to see the dramatic swings those contentious posts take.
The fact that the management of this forum, who are VCs, permitted such a mechanism is part of the "own it, HN" assertion. HN has baked in something like 'peter principle' into the forum. Karma Grifters who post articles that get them over the 500 points are then ala Peter Principle granted the right to 'grey out' actual thoughtful comments.
Are you telling the management of this outfit never looked into this phenomena?
How about ageism? Mr. Paul Graham and personality cult asserting that anyone over 20 something is no longer viable for leading edge tech work?
We used to call these VCs "vulture capitalists" in the 90s. We geeks were so right about so many things in the 90s: We were right about GPL. We were right about VCs. We were right about surveillance tech. We were right about outsourcing ...
But alas, "corpo-techbros" empowered by thoughtless forum software courtesy of Paul Graham and company got into this mess.
> Are you telling the management of this outfit never looked into this phenomena?
To counter, I think that HN is being used as a testcase to shove techbro and VC ideology across all of tech. And secondly, its some of the most potent tech market research. Its a textsearch goldmine.
I believe YC knew what they were doing, and intentionally chose this course of action.
> Additionally, founders of YC companies see each other's usernames show up in orange, which — although not an explicit benefit — does allow fellow YC founders to immediately identify one another in discussions.
Even with the significant bias here, I still read it. I also read lobsters as well, which is here minus techbro insanity.
Right now it is just a hobby and there are still a number of bugs remaining. Since I don't have an income from it, I can't dedicate more time to it. Hiring me would allow me to work on fixing them full time and make the progress much faster
Hey, props to your attitude, and I wish you the best of luck.
Obviously, you've provided value to a company in a really in-demand area. It doesn't feel right to treat the contributors like this. Sadly, it seems that the companies have the power and the intent to just abuse and exploit
I don't have a solution. I am just expressing my frustration from the perceived injustice.
Correct, you cannot remove a version or the whole crate unless very specific criteria are fulfilled. You can "yank" versions. That prevents people from adding the version as a new dependency, but if you relied on it before it got yanked, your build will succeed.
I wouldn't delete old versions even if I could though. My goal is to publish a rock solid library that everyone can depend on and build awesome projects with
Why wouldn't a company want to hire the fore-most expert in a tool that is critical to the company? They are hiring for someone with that exact expertise.
A competitor could hire the OP instead, get them to work on improving the software for a few years. Giving the competitor a major head start.
Worst-case scenario, the tool they are building doesn't work out and Anthropic has a pretty good developer to put on other projects.
I can't say for Anthropic, but I've seen Google hire people working on open source projects that were aligned with the skills they were looking for. Desktop search and collaborative editing comes to mind, although I might be mis-remembering?
But he's not doing the work for free. He's doing something else for free which they use. He has domain knowledge with the library that noone else has, their can either pay someone to learn it or they can hire someone with it.
And that's exactly what these companies can abuse.
If this wasn't available for free, they would gladly pay for a programmer to create it. But if it's already free, they can use it as a starting point. Maybe they'd need to internalize/extend it. But the option of paying for the work already done is gone.
Do this for each npm dependency and you're looking at huge savings.
This is still illogical. You can hire the original maintainer and pay an incremental cost, or you can hire a random developer and pay the initial learning cost + higher incremental cost.
If every company using a library chose the former, then every hour of development would be paid for (from the perspective of the maintainer) and the cost would be spread out across all its users.
You can use what is as is. Then you can ignore all of the other issues if they don't impact your bottom line.
Don't get me wrong, I like your corporate OSS financing model. But there seems to be not enough incentive for companies to use it. Why take ownership for a small cost, when you can use an imperfect thing with no cost?
Was wondering something similar, if OP had blogged it earlier when he found claude was using it and re-posted it in HN/reddit it in a sensational way to capture eyes. Maybe through one of the forums he could have got an introduction and a job doing what he loves.
OP still has a chance now, maybe not anthropic, even other competitors can come knocking.
> Sucks, but that's the reality of hiring (and getting hired) in tech in general.
If you’re in the inside, it doesn’t suck at all, it’s so much safer.
Hiring a new person, based on a few hours of interviews, and a resume half full of exaggerations and lies, is such a ridiculous gamble. Worst part is, if you realize they’re not a good fit, it’s sometimes incredibly hard to get rid of someone, more often not an option at all.
My experience with Anthropic and OpenAI is they’re not super interested in experience and don’t take internal references very seriously. Most of who I know that was hired were fairly junior folks and they have an early weed out that’s a fairly rudimentary but very specific Python programming quiz typically administered by very junior (like 1-2 years out of school) - even when interviewing extremely senior and experienced people of some substantial success and renown. This isn’t uncommon - meta and others do this too. But the programming quiz at Anthropic is sudden death and the first round, and the people administering it are looking for a very specific implementation that if you don’t see it immediately they just Gen Z stare and don’t discuss etc. It’s one of the more amateur selection processes designed with an extreme bias against more senior folks (frankly it felt unintentional just naive). (Meta etc scale the programming weight to seniority and the administrators scale as well - asking for depth of understanding of concepts as seniority grows with the expectation experience brings more to the table than syntactic knowledge).
So getting an internal reference and being highly qualified for something they need done isn’t enough. You need to also make it past the 20 years old gate keepers and their amateur hour hiring process.
Unfortunately, the choice of license likely won't matter in the nearest future (if not already so). If a tech giant wants you open-source library, they will just point their agent to it and ask "to rewrite in the style of War and Peace". And more unscrupulous players won't even bother with a rewrite, as we've seen recently in the case of Cheatingdaddy/Pickle.
The flip side is that if they can technically pull that off then the cost of writing the library has dropped so low that an OSS maintainer probably wouldn't have to work too hard to write it anyway.
Being able to rewrite existing working code sufficient to copyright-launder it isn't the same as being able to write it from scratch, unfortunately, especially since LLMs seem to be allowed to ignore quite a bit of copyright law with complete impunity.
Imo it's totally plausible that something will be expensive & time consuming to create, even with LLMs, but still easy to fork outside current licensing restrictions with LLMs.
Rewriting it with a guarantee of not introducing any errors is still beyond current LLM capabilities, and there might be a certain correlation between that capability and the capability of writing it from scratch.
>If a tech giant wants you open-source library, they will just point their agent to it and ask "to rewrite in the style of War and Peace"
Is there any evidence of this happening? And any legal theory behind how it might have the intended effect? Training being fair use does not make AI a magical copyright-removal box.
AI is already "too big to fail" in every conceivable way. If there is a conflict between AI and the law, then the law will inevitably be rewritten, reinterpreted or repealed in the way most beneficial to the service of AI.
Tbf, I'd rather get a "we didn't review your CV" response than a template "we are continuing with other candidates :)" response. It softens the blow considerably and helps me as an applicant better keep track on which variation of the CV is working best because I can just remove this datapoint.
I had a hiring manager tell me a couple weeks ago that I probably made it further because I applied late. All the people using AI and automation to apply are hitting the apply button early, and the laggards tend to be humans.
Its a perpetum mobile. Hiring managers use automation to filter candidates, coz its too many. Candidates see they dont even pass automatic filtering, so they apply with tailored CVs x10. This means even more CVs and more filtering and more CVs and more filtering and more CVs etc etc etc
Im curious at which point ppl will understand its counter productive.
Never, because they already know it is but fixing it requires both sides to deescalate in lockstep while individuals on both sides would benefit from not descalating
Unfortunately, "we didn't review your CV" is a great way to get sued in the US if the name of the applicant is in any way potentially indicative of the applicant's gender, race, religion, or any other protected status.
And we now see a lot of proof that that is warranted: lots of people in the US are now openly racist and/or have other prejudices. It is fair to assume many people don’t get a fair chance because of this.
The first company I worked for (a smallish business) interviewed a guy who seemed normal and knowledgeable enough in the screener call. At the time, they just asked that people being in code samples to discuss in the second interview (this predated GitHub) and he brought in some obviously copy-pasted code that didn't fit together.
He then, without prompt, in the middle of a conversation mentioned that he was the second coming of Christ. The interviewers ignored the comment and continued the interview.
When he didn't get the job, he sued the company for religious discrimination. Fortunately, the interviewers could honestly say they didn't discuss or ask about his religious beliefs, and he lost. It was said he did this elsewhere as a a scam, though I never verified it.
The simple matter of fact is that it doesn't matter how neutral you are; there are enough people out there who will look for any way to perceive and benefit from a grievance that you must assume they will.
Of course you're right, there will always be people that try to abuse the system. But the bigger picture here is that more people are truly being discriminated for their color, religion, sex, etc, than there are people abusing the system. A system that improves live for many should not be removed because some people try to abuse it.
We are discussing whether or not it is safe for a company to reply to an applicant with "we didn't review your CV".
The crux of my point is that the potential for a perceived grievance is sufficient to trigger legal action. Whether bigotry was involved or not is entirely beside the point, because even if the company absolutely was 100% not discriminating, they are still vulnerable for creating a situation where they could be perceived as having done so.
In no way am I advocating for removing protections for disadvantaged groups. I'm not arguing that bigotry doesn't exist. There's no point in bringing up the fact that it does.
You know what's cheaper than getting hundreds of baseless court cases dismissed? Not replying to someone with "we ignored your application".
For good or ill -- probably the latter -- headshots are expected with CVs in most of Europe. It's a local custom.
Source: I worked in Germany and had to deal with this. (In fact, one of the ways I made my application stand out from other North Americans was to learn this ahead of time and include a headshot in my original application)
To be fair with Anthropic, they probably get unfathomably many applications for everything, on top of the cold calls/emails. They're one of the hottest companies in the world, so I'd expect tens of thousands of applicants. Media writing about $100m+ hiring deals in AI does not help, either.
Aren't they an AI company? Couldn't they sort it out? If Anthropic, of all companies, can't sort out incoming job applications, what exactly are their tools for?
The world is also full of totally delusional people who dreamed up the idea of using winzip to compress VRAM on the GPU, and now Anthropic will definitely hire them for $1M year for this genius solution, so better write up a glammed up resume and auto-send it once a day for any open position.
Is it? What I've been marketed is that AI is widely available for every task I could conceptually think of, and that I should be using it for everything.
Opportunity also is a precious resource. Redirecting resume to /dev/null is wasting it. I have a hard time believing that LLMs, with all their sophistication, aren't ideally suited for this task.
I don’t know how slighting customers who want to work for you works out in the long term. You end up getting fewer opportunities from them and their friends in the future.
Seattle is full of people who will tell you what it’s like to work for Amazon and how you don’t want to work there. I guess if you’re big enough though the money papers over a lot of sins. The smaller you are, the more people you can piss off before you run out of prospects. Anthropocene still has a long way to go before they are Facebook, who struggles because something like 50% of the people who would work for FB already have.
IDK, compute isn't really that precious. If $20/mo can get you many (?) invocations of their research agent, I feel like it could pretty easily be worth it to screen applications for jobs that pay $350K/year -- and that's just "entry-level"![1]
That said, their career page puts this at the very top of the details section:
We value direct evidence of ability: If you’ve done interesting independent research, written an insightful blog post, or made substantial contributions to open-source software, put that at the top of your resume!
This guy seems rad, but his GitHub[2] and this blog are both light details or links, which is odd considering that his LinkedIn[3] is detailed+professional. Perhaps Anthropic does have Claude screening resumes, but he didn't express the nature of the situation clearly enough for it to catch it?
Otherwise, the only other explanation I see that doesn't look terrible for Anthropic is they didn't see a need for more Rust expertise...?
Tbf the other summer recruitment practice in AI this summer is Zuck running round offering engineers with some sort of reputation $100m+ windfalls, so maybe all the OP needs to do is add "author of computer interaction library used by Anthropic" to his LinkedIn profile to acquire that garage full of Ferraris
Nah, if there was ever a time where making meaningful contributions to open source was important to land you a good paying job in a hot tech company, that died a long time ago. The people making these decisions don't care, unless you have someone inside to put your resume first it doesn't matter that you wrote all the code that makes their product even possible, the hiring manager won't care.
I might just be old but i really haven't felt like contributing to open source at all lately because i've bills to pay and kids to care for and taking time out of this just for the sake of enriching some billion dollar corp that will eat me and spit me out doesn't feel like a good investment for my time.
Sometimes i feel sad that it came to this but this is the place we're living in right now.
they should have used their AI to scan through resume... they are AI company afterall. shame they missed this guy. it shows their resume-scannign AI is useless.
Poor poor... I always felt that too many people in hr decision making underestimate the role of talent. Many awesome software products stem from teams with extraordinary talent. Average people create average software.
I wonder if it was geolocation? Anthropic is based in SF, the author seems to be based in Munich, and maybe they're not open to hiring people who aren't based in the US right now? Given the state of US visas right now, this wouldn't shock me.
My company, which is significantly smaller, hires people in multiple countries across the world. You don't need an office to hire (I am sure there so exist countries where you do, but I expect they are the minority).
Also I low how the IT hiring has a become a paradox:
Companies won't hire you if you don't have enough projects in your portfolio, but by the time you will have enough stars on your github projects they have already used you to their own goals and are "not interested".
I don't know if this is a good opinion, but I don't think it's a good idea for independent individuals to use highly permissive licenses on their open source software. Companies will just suck it up and might not contribute back. It distorts the market because if the software didn't exist, they'd have to hire people, contract it out, etc. somebody would get paid. you've saved a huge company from having to hire people to develop the software they need, which is good for them, but imo just gives the companies incentives to devalue engineers. I also think the value of somebody open sourcing their work as a means to getting a job is questionable and never really been backed up by any data.
Excellent point. If you really want to make your project open out of the goodness of your heart, then use GPL. Otherwise you are explicitly giving any business, no matter how big or wealthy, permission to use it with no expectation of giving anything back. The license says that quite clearly. It says nothing about rewarding the author.
Through a friend of a friend, I found out that Anthropic had an open position
"Coffee" with the friend of a friend would he better strategy than a cover letter in that case...more work, but better strategy.
Because logically, getting hired requires demonstrating you are "the kind of person we want to work with." Being qualified on paper is not necessarily required.
You basically need to avoid putting your name into the main queue as long as possible. Make them do it manually on their side and keep track of it. If they want you, then you’ll bypass all the crap this way.
Not too often tbh. You either get dogshit LeetCode or, more often, just a general chat about what you've done and know. There, social skills play a massive role. Make simple projects sound like state-of-the-art, present everything cool that happened as something you were directly involved in, present the 2 most obscure bugs in the project as something you fixed every other Tuesday when you get bored...
As someone that works in HR, the incompetent HR combined with using AI for ATS ( or not knowing how to use ATS at all) is one of the core problems when it comes to losing quality candidates and is to blame for this. It should be illegal to hire HR from any education other than law, psychology, management and economy background. That way the responsibility would be larger, the ROI on HR would be higher (because the retention of the candidates and the quality of the candidates). Simply paying and promoting people with any educational background in a HR role is a waste of money which also creates problem for the company and not just employees.
> It should be illegal to hire HR from any education other than law, psychology, management and economy background.
A lot of people with education in management/business do go into HR, at least in countries I know, and it does not help. People with extensive management experience would help but they will only take more senior roles.
The other qualifications open opportunities interesting and well paid careers. How would you attract those people into HR?
I am not even sure it would help if you could.
I think the suggestion in the old management book by the guy who turned around Avis that you should have an old style personnel department to do admin and advice, and managers should have more involvement might be a way forward, but I am not sure it would work given the current level of regulation (in the UK anyway - I imagine most wester countries are the same). A lot of the function of HR is to avoid legal risk (e.g. fire people according to the rules, so go through the motions of warnings etc).
I think that old departments (personnel departments) should have been just modernized in reality. To be frank, in some cases a mix of HR/Legal department is cost saving too.
What it really comes to is that a lot of people love to micromanage everything. If you hire someone that has integrity and educational background in subject, he/she will warn you if the decision you are making will have consequences in the long run. If you have someone that does not have relevant education, that simply does not happen. The managers micromanage, those people receive salaries and if they step out of the line even when they are right, they are reminded that they do not have relevant knowledge in said department (law/economy). This in turn leads to a lot of people gaining something called shallow experience which then in turns leads those people to hire someone that des not pose the risk to their position further down the line.
The problem being in this case is that there are a lot of misses that happen when the HR is organized like that; from illegal hirings, not knowing key economic factors, not having a clue about the business itself, no clue about laws and procedures and so on. Which in turn does not really protect the company because the company loses both the money and employees.
Not unlikely given the HN attention this post got… Anthropic’s small hit to its reputation prolly isn’t worth the $2xx-4xxk / yr. they should be paying this nice person.
There is some dirty secret i learned in my time as a eng. manager: Working in open source / Being the maintainer of a popular library / Blogging about software: All this things won't give you necessarily a competitive edge but can work against you. It's counterintuitive but sometimes teams are looking for a more low-profile hire.
In my experience, maintaining a very popular software library, supporting open source, and blogging have absolutely all contributed to my success, and, additionally, as someone who is now a founder seeking like-minded, highly skilled engineers, those are key signals for an attractive hire.
I can understand though, perhaps in a work environment where management is unlikely to be able to retain high skilled talent, you may want 'low-profile' workers that aren't going to have as many competitors chasing after them...
How so? Care to elaborate? I get that bloggers/educators can sometimes be not the best fit for IC roles but doing open source seems like a huge advantage.
Some companies want subservient, homogenous employees that come in, do work, and can be let go if they do not perform. That's a simple equation.
If you get in somebody who is a star, however minor, that changes the equation, changes the dynamic. Now that person can have more confidence, can have more sway in the decision making. If the company wants to let them go, then they might post a message to their followers, riling them up, creating bad PR for the company. It's no longer a simple equation.
So it all comes down to the insecurities of the company.
When parent poster says things like “low profile” it should be interpreted as cheap and doesn’t know their worth. Assume all hiring managers want the least qualified and cheapest possible employee that can still get the job done.
Not always true, but true enough to be useful and more true than hiring managers admit to themselves. I’ve been a senior involved with hiring for years because while I full don’t want to manage, I also never trust my manager to hire well. They have multiple mutually exclusive narratives they tell themselves about how they hire/manage. Not all of them are true, and sometimes not any are.
It dependents on the size of the organization a lot. However in general it's likely that the new hire is the most competent of them all, which would be an immediate risk for some of the managers (e.g being displaced)
This is kinda fair, though. People who have run their own business make for really, really, awkward employees. It takes a really skilled manager to deal with them properly
IF you have a side gig it is easy to think you won’t be 100% invested in company success. If you monetize you most likely will jump ship.
There are other risks like burn out as you may read a lot of OSS contributors have — so when someone is hit by burn out it will be across the board not that they somehow will perform at their peak at job while burned out by coding on side.
I was part of the selection committee for a position once, where we selected the more junior engineer.
The probably most simple explaination would be that for some roles you like to have someone that can be easier "shaped" into a certain role. Someone who is already successful may bring their own system of doing things. This is great if it is a good fit, but can produce frictions if it isn't.
The next thing is that if you apply to a mediocre position with overly amazing credentials, it can raise suspicions. Something must be wrong with you, maybe you got amazing credentials, but you are complicated to work with. Maybe you're looking for the mediocre job just because you think it will be a walk in the park, etc. There are legit reasons for this (e.g. "my partner moved to $TOWN for her career and I am looking for something to do here, and you seem like the best fit. I know I am technically overqualified, but I wanted to go back to coding for years now and this offers me a geeat chance to give it a go").
Of all the senior canidates we have rejected the most common issue was that they didn't offer a convincing explanation to why they chose that specific position. The worst one was talking about how it would be a relaxing position for them.
Yes, sure, in theory. But the position we were filling was one with very little supervision and oversight, for room reasons. So basically one person in a room in a different building who has to maintain a bunch of stuff in addition to build up a organizational structure from scratch.
Filling it with someone who you might have to check after not for seemed like a risky bet. Call it a gut feeling. I worked together with a guy like that, which lead to me having to save the day every other week because he forgot to organize for an event he knew about months in advance.
That's often works and is a good idea, in my personal experience. It would be so much better, however, if we had a functional and affordable health care system for independent consultants. Consultants working from outside the US may actually have an advantage in this regard, depending on where they are exactly.
This is cope and propaganda to discourage people from developing their own brand. Better for the corporation if the workers have no support structure or reputation that might lead them to quit
"Developing your own brand" is not a scalable solution. There's only ever going to be a few thousand developers who are well enough known to be called a brand.
I think this blog post was the best way to get into Anthropic, and it was well-deserved. That's the reality of hiring in tech: there are many non-technical people judging whether technical people are competent or not. Escaping that matrix through things like blog posts, cold emails, and Twitter threads can be great ways to break in and get noticed by these companies.
yeah, as this is so often the case, many times good, relevant applications are missed. I hope that this Hacker News post will get to one of the key people at Anthropic and they change their minds.
At my firms I saw this happen often. HR would review, or a junior engineer and pass on very good candidates. It wasn't until I set up a review system with A-class engineers that we started to catch the best people. A-class engineers recognize themselves far better than anyone else. But they prefer to build than review resumes.
I ended up building my own head hunting firm specifically to address the whole pipeline. That helped somewhat but head hunting is its own very odd space. Full of inefficiencies and bias.
With any AI company, there are always limits you hit. Energy, compute, optimizations, inference, team resources, money, and all the flows to make it a company. HR is usually the one that gets the fewest resources.
I think the issue is that some applications are not even reviewed. HRs can also learn the expertise of identifying strong candidates if they build up the experience and frequently talk with engineers about pros and cons of resumes.
first of all because the key point is they didn't even look at his application, and by any objective criteria he should have easily got through a "worth a human looking at it" screen. but also, hiring the developer of an open source library that you want to use internally and paying them to both integrate the library and work further on it is an excellent way to have a sustainable open source ecosystem, which both anthropic and the developer will benefit from.
I’d be curious to see the outcome of changing the license to a Fair Source License or explicitly “You are not allowed to use this software if you are Anthropic, otherwise MIT”. They could still use the current version, but for any in the future they’d be forced to fork it or be prepared to face yet another legal battle (I can imagine some lawyers already salivating at the thought).
It’s also curious the author is looking inside the app for proof their software is being used. If it’s MIT, mustn’t the license be included and available somewhere easier to verify?
I would guess like him that no human engineer ever read his application. The less they would have done in that case would be to at least thank him for his work, even if they don't plan to hire him for some reason.
Automated systems, AI screening, and incompetent HR people are the bane of modern recruiting practices.
On the electron part, it's common to (ironically) not support Linux. There are pretty annoying bugs with windows management (window will stay stuck in the background), build process are always OS specific, etc.
So often not worth the maintainance.
For generic consumer products, sure, but for dev & technical power user tools the audience is big enough that these arguments doesn't hold water. Stack Overflow's latest survey shows nearly 30% of professional devs using Ubuntu specifically (https://survey.stackoverflow.co/2025/technology#1-computer-o...) and my own metrics (building a cross-platform dev desktop app with a global dev/technical user base) show pretty similar numbers: 65% windows, 20% Mac, 15% Linux. I would expect there's a significant (comfortably above 10%) Linux user base within the claude computer use audience.
The practical reality of distributing is mildly complicated, but there's now lots of good cross-distro options, and not having to deal with code signing everything makes some parts much easier than Mac & Windows. Ignoring that many users is fair enough for a startup or first MVP, but quite surprising for a company at Anthropic's level.
Uh, a company not paying money for something they can legally use for free? There are so many MIT-licensed software libraries that everyone is using in a critical place, for profit, with zero money flowing back into the ecosystem that created them. It should surprise nobody, it has been like this for over a decade now.
"Overall I am overjoyed enigo is used in Claude Desktop and I tell everyone who listens to me about it :P. It's so cool to think that I metaphorically created the arms and legs for Claude AI, but I can't help but wonder if the rejection letter was written by a human or Claude AI. Did the very AI I helped equip with new capabilities just reject my application? On the bright side, I should now be safe from Roko's Basilisk. "
I also felt like this way that did they just AI in their interviewing process?
And I have a special love towards open source.
And I personally might be happy too that a company is using my work ,but in the name of the holy licenses, Companies are just exploiting the free nature of this and the fact that it seems like not even a human looked at the person for such job, who created a library that they are using it for free...
I was thinking of creating some code in MIT license, but I am going to create a code of AGPL except if you sponsor me on github or a special one time license which can grant you MIT.
People might say that I am not fostering the open source community, but I am not giving corporations free labour so that they can be billionaires.
I once saw someone write a software with the exact same idea (AGPL + gh sponsor me to get MIT) and the people in HN were pitchforking him, that's the harsh reality of the world. People want absolutely free labour.
I think open source needs to ask, Have we become the modern peasants in the name of our altruism?
I think we need to compare our industry with other industries. No other industry relies on free labour from random people, which comes with no support or promises.
I once told some non-techie folk about some code I wrote. It did something super simple and wasn't that big. They were all asking why I didn't sell it and thought it was crazy I would give it away for free with the BSD license. It was 900 lines of code... For us, that's nothing but for an average person they just think "I built it, I'll sell it"
The problem is that 900 lines of code is also nothing for your potential customers as well. Non-programmers have a very poor ability to judge how difficult something is and how worth paying for it is. 900 lines is probably less effort for most organizations than it is to evaluate paying for the functionality.
Out on the super, super far end of the distribution you may have things like paying for what is essentially 900-ish lines of extremely, extremely carefully vetted code for things like encryption, but that is very, very exceptional.
I've got a few open source projects on my GitHub that are in the 900 line range, and I know they're used in a few "interesting" places but I'm not crying about it because the simple truth is the commercial value of that code is simply $0. If I tried to sell it to the people using it, they would perfectly rationally just say no. I am abundantly compensated for it by all the other open source software I get to use.
Ya I also believe it this way,
Mostly I like to build stuff for my own problems or something that I just find fascinating.
I am still in high school, so I was doing some question sheet that our teachers provided and there was an answer key but it had answers of everything. Now I don't know how other people approached it but I am really impatient and so I just open up answer key side by side but it reveals every answer.
So I firstly created an AI to ocr to card generator but it was an hit or miss and so I discussed it with my friend and he said that he used to use paint and somehow in his convuluted manner basically have a slider which would reveal answer...
I found it incredible and so I just created a single index.html that can do it. (Although vibe coded), Now I can't even think of monetizing such ideas when I realize that there are creators of some really incredible stuff and long convulated stuff and even they aren't sponsored so I have always felt that the scripts that I write or projects around such ~.5-3k loc. I just don't think of monetization.
I just don't know.. I like hacking stuff, I just feel more comfortable rebuilding stuff even if its mediocre if I feel like I can change it to suit my purpose better
I think that the only other industry that is gives as much completely free stuff might be research/science related, but maybe its due to the fact that computer are computer science too and thus related to academics.
I really just love tinkering with software and just the aspect of freedom that it can provide , but sadly, I find it just hard to really make money without being a job and such stories on which we are discussing, just makes me feel like I am kinda right.
On one hand we have 100 million payouts to researchers and on the other we have this, such disparity is kinda sad I suppose.
Other comparison would be that we are equipped with tools (software) other people build completely for free. And we can improve those tools and propose them to be implemented (PR). Or just continue using our custom-modified tool (Forking).
And we often get the luxury to ask questions and receive answers (Issues) directly from the manufacturer (author, contributor).
And we need not much investment to set up our own factory... thus "materials" can be free and then we give away our product.
A friend of mine is maintainer of an open source service used (at least, at one time) by all of the major social media platforms as a load-bearing piece of their infrastructure (intentionally keeping it vague). My friend was invited to interview at one of the biggest and was rejected after having a bad whiteboard session. Of course they immediately replaced my friend's service (ha!)
Since licensing has come up a few times in this thread: I've been hearing recently that the Mozilla Public licence (or the EU Public licence) is a good middle ground between the "viral" GPL and the "do whatever" MIT - as per my understanding, if your code is MPL or EUPL, it can still be incorporated as part of software that has a different licence, but any direct changes to the MPL/EUPL licensed code itself has to be shared openly.
Does anyone here have experience with them, or knowledge about whether that description is more or less correct?
Good findings, the rest not surprising tho.. online recruitment doesn't work at all these days. most likely your app wasn't read by anybody meaningful and did not trigger right flags in the HR system to even be spotted by clueless ladies working there.
This post can give you some visibility unless somebody sees it as frustration/negativity then they won't bother either.
aside of the core topic, best way to get a job these days is unfortunately either some elite job boards that work and both sides know why... or personal relations.
All the automatic HR/recruitment platforms is illness and i'm sure that's what victimized your genuine application there.
I assume your choice to describe average HR reps as “clueless ladies” isn’t meant to suggest that you respect e.g. women software engineers on your team any less. But if the gender of the clueless HR employees isn’t relevant, why mention it? Maybe worth reflecting on whether calling them clueless ladies rhetorically emphasizes their cluelessness
The OP should have left off "ladies" as that was unnecessary -- clueless is a sufficient descriptor for many people in many roles (whether genuine behavior, strategic fiefdoming, or learned helplessness).
The distribution in other categories is fascinating, and HN doesn't format tables well.
Though "saving a click" typically refers to spammy clickbait news articles that bury the lede, which a statistical table directly relevant to the conversation does not qualify as.
I read the original comment as implying that the average HR person reviewing your resume will be a clueless woman. Not as implying that only women work there or that people of other genders working there are more clueful.
The comment is open to interpretation, and you are free to interpret it in a less charitable way. The ambiguity is absolutely something we can and should criticize the comment for
I’m not sure the percentage of companies that use software for highlighting candidates, but Anthropic almost certainly does and this [2] source says 75+% do.
So since men wrote the software that didn’t highlight the candidate, is it the clueless men that caused this?
Yes - people in HR departments are often female and often clueless, but I don't see the parent denying this. The wording of OP connected both though, which is sexist and can be considered "evil".
Funny enough, I see this whole framing as sexist itself.
Nobody would have bat an eye if he said "clueless guys" or "clueless gents", and given the prevalence of women in HR, that wording would actually have more chances of having a sexist background to it.
“guys” is gendered but is very often used to mean a general group.
>given the prevalence of women in HR, that wording would actually have more chances of having a sexist background to it.
The reason there are more women than men in HR is clearly because the men they do hire are too clueless and get fired faster. Ever have an HR department with all men? Most dysfunctional department I’ve ever interacted with! “Clueless HR men” is just redundant. The ~25% that exist are DEI hires. So it wouldn’t be sexism, it would be reality.
You‘re right, but that just reflects the structural sexism in our society while the wording by the op was intentional (I suppose. If not, I might as well be more sexist then he is).
Nopes, nothing to the ladies in general nor any other gender in general. Just a shortcut to my own negative experience with HR by example. English is my foreign language and in my country we are not that allergic to terminology. But clueless processors stay as valid… regardless of particular denomination.
After years hearing justly about bad things perpetrated by males as a class, without any concerns about generalizations, I think we are mature enough to also call for responsibility in the other side of the aisle.
Having a free pass for doing evil stuff is what gave man their bad rep, should we now for equity give women a pass to become the new slave lords?
Never mind the application rejection, you'd have hoped that in choosing to base Claude Computer Use on his library that they might have at least reached out to the developer to say thanks, preferably with some token show of appreciation like a few Anthropic shares.
This is why wealth accumulation is so terrible. People with lots of money drive science and technology. They accumulate more wealth from science and technology whilst demonstrating a complete lack of understanding of the thing that's making them the money.
Most executives and investors just throw shit at the wall to see what sticks, imo. Then move on to the next place. That's why golden handshakes exist.
I'm with Luke Smith [1] when it comes to non-copyleft licenses like MIT.
Andrew Tanenbaum of the MINIX fame was similarly surprised to find that Intel had quietly included the OS he wrote in Intel chips, making it perhaps the most widely used OS in the world. He seemed disappointed no one ever reached out to him to tell him about it [2]
To intentionally mix metaphors, I can't believe I'm about to knowingly kick this well worn hornets nest.
It seems obvious that if Tanenbaum, or any open source project used a GP license in lieu of a permissive legally familiar license like MIT or BSD, the likelihood of the project being used in a commercial product would reduce to nearly zero. Intel would have used a different OS for their management engine.
I'm glad the GPL exists and believe the world is a better place because of it, but it feels like more and more it's salad days are in the past and the world has moved on.
The ops experience reminds me of the story of the maintainer of homebrew that despite widely being used at google was not able to be hired for a job there. It's disappointing and feels unjust, and I wish it was different.
We should stop coding for free for billionaire organizations. The romantic era of Open Source is over.
The only projects with a permissive license, I am comfortable sending PRs nowadays are the kind of projects that will hardly enable a big monopolist to extract more rent from society while being covertly funded by the debasing of currency promoted by the FED via Cantillon Effect.
Over a decade ago, it was my dream job to get a job at one specific FAANG company that is widely known to use a project I've contributed to.
I'm a developer with a project they use, so, I thought, for sure someone would review my resume after applying on their website. Nope.
After being ignored for a while, even having to get a Master's degree because no offers after a Bachelor's, I finally emailed a Director, who was previously a fellow committer at the project. People under him were not hiring at the time, but a recruiter from a different group has contacted me shortly, and I've had a 2-day flyout onsite arranged for two different positions, and had offers to join either one.
I don't mean to downplay the author's skills, but I don't see how creating an input simulation library fast-tracks someone for consideration in an AI-related engineering role.
It mentioned an "open position in the team implementing the secret, unreleased feature of Claude Desktop," which doesn't specify whether the "secret" feature is AI-related or UI-related. My guess leans towards the former.
You got me, adding “using enigo” makes all the difference — I guess position is exclusively for working with this one library and that’s the position they got overloaded with applications on and couldn’t process one sent by OP.
It clearly does make a difference as to why he thought his experience was relevant to the job (i.e. what we were discussing before), and I think you agree with that hence your somewhat "selective" omission when you posted the quote.
I wasn't making a statement about whether his experience was relevant to the job. I don't know the author and don't automatically doubt his knowledge. I was simply sharing the opinion that being the author of that UI library alone does not fast-track someone for the "Software Engineer working on Claude Code" position at Anthropic.
I wasn't aware of that; it wasn't clearly specified. It only mentioned a "secret" feature, but I assumed it was AI-related rather than UI-related. Additionally, Anthropic's Claude Code position on their website states that they expect their developers to work across the stack, including both front-end and back-end.
Anthropic probably gets tens of thousands of applications. They seem to have filled their queue before even reviewing this particular candidate. Unfortunate but just reality.
Always always always try to get into direct contact with the actual hiring manager. Blog author had a friend of a friend let them know a relevant role was open. The correct move is NOT to blindly apply. It’s to ask for an intro to the engineering manager responsible for the role.
I'm very much starting to re-consider open source. It mainly seems to be a way for already incredibly wealthy companies to get things for free, or to strategically release things to crush their competitors.
Maybe we ought to go back to paying for proprietary software. A lot of people used to make money that way, ie by selling their own desktop app.
reminds be of the time creator of Homebrew was rejected by Google in coding rounds. but this is even worse, they would not even interview this guy. shame on Anthropic... (or is it Misanthropic?)
In my opinion, lots of open source was developed as a sort of portfolio to get hired. From 2019 onward, my impression is that your open source projects (regardless of how much they are used) matters less and less and it’s about HR mysteriously picking you up in their process than anything else. I think, now, your open source portfolio matters exactly nothing in the decision to get hired.
I remember back in 2014-2019, it was hard and competitive to contribute to open source projects as they were tightly guarded. There are many projects that I use now in package.json that are looking for a maintainer. A complete 180 flip.
My guess is that real free open source will disappear in a few years and what will remain are open source projects monetized by some business somehow.
It’s a sad reality but that’s what the current people at the top have decided today.
Why is no one talking about how they had an indirect contact at Anthropic but didn’t use that connection? Your chance of getting hired is way higher with a referral.
Everyone is suggesting that AI rejected this candidate but that brings up two points:
1) Is the hiring AI so incompetent that it did not realize it had a "S-tier pull" in the process and should have immediately prioritized the find?
2) Was the candidate's submission so bad that a reviewing AI couldn't even tell the massive relevance he had to their work?
I suppose, alternatively, Anthropic could just not really care about Claude Desktop enough to hire a specialist for one part of the stack. Perhaps they're looking for much more "full stack AI" who can do a lot. They have 350-400 total engineers, is that enough to hire a specialist for Claude Desktop?
I guess my question is: Did the AI fail, did the candidate fail, or did the AI work well and we just don't know the criteria it was succeeding in using.
The next version should have a feature where the first thing it types into any text box is "Anthropic, I wrote this library! Please look at my CV!" and then deletes it.
> Unfortunately they thanked me for my application but said the team doesn't have the capacity to review additional applications.
Okay, they were just busy doing work and didn't have any time to look at applications so they shuttered the JD and auto-rejected anyone in the pipeline. Seems reasonable
Another reminder that if you write software under an MIT licence or similar then you're just working for companies like Anthropic for free.
Use GPL or AGPL. It's the best thing we have.
Remember that companies like Microsoft spend billions on PR and their goal is to make you think what's good for them is good for you. This is rarely the case.
I feel this is really a blog post about two indirect topics; one that has not been addressed for many years now, and another that is not new, but has been getting seemingly ever more acute recently.
The first is the issue of permissive licenses like the MIT license, that seems likely far beyond an appropriate license structure for today’s world and environment, I would even argue inappropriate since the .com bubble. Software and creating has changed a lot since the 1980s to such a degree that I don’t think even the originator and early supporters of permissive licenses would be supportive of…peoples work being used in critical ways to build two and three digit billion dollar corporations without any kind of reward or compensation. It’s an odd kind of peak dystopian hybrid of communism and capitalism, sacrifice of the self for the benefit of the very few.
I think it is at least time to discuss archiving things like the permissive MIT license (assuming it even makes any kind of difference at this stage) that are from not only a different developmental stage, different environment, but even a totally different country, society, nation, and world even.
The second theme of this blog post seems to be the absolute seizure of the… what should we call it?…resource allocation of people? I cannot recall right now, but I feel like this is the second blog post themed around someone core to some function of some big tech company being rejected by said tech company; and that’s in the backdrop of the cacophony of people dealing with all kinds of dystopian insanity in the employment/job market from fake/scam jobs, AI interviews, etc. The system seems to be totally breaking down to some degree, even if it is still limping along, as is evident by the massively downward revised job creation numbers over several quarters now. How do you “revise” jobs numbers from 139,000 to 19,000? Ignoring any political partisanship, “revising” an estimate downward by 86% is not just an “whoopsie”, it’s evidence that thins are broken, regardless of why or even how. They’re clearly broken.
I have approaching 0% confidence with anything related to Congress actually doing its job since it has effectively abdicated its cute role that provides it legitimacy, but discussing both of these topics in public can have a chance at forcing the muppets in Congress to address the issues, even if only for narcissistic and selfish reasons of being (re)elected to enrich themselves after they’ve gone back on their lies to get elected. And no, neither team is the better team; it’s all a con-job.
> I feel bad about my tweet, I don’t feel it was fair, and it fed the current era of outragism-driven-reading that is the modern Internet, and thus went viral, and for that I am truly sorry.
Building popular software doesn't mean you're a good programmer, especially since at that point Google was looking heavily at CS concepts and he admittedly wasn't good at that.
It's also possible he would have been hired if he applied for L-1. A lot of people get an ego check applying to Google where they're a senior staff engineer or a CTO at a small company and get an L5 offer.
True but surprisingly grinding leetcode puzzles also doesn't mean you're a good programmer. In fact, in my decades-long now programming career, I've had to take many more decisions of the homebrew kind (e.g. how the thing is going to work, how the API is going to look like, will the users love or hate that feature, etc.) than the leetcode kind. And now I am thinking the former is even more important. If you get the leetcode part wrong, worst thing your code would be slow. Not a good thing but also not a complete disaster - you can come back and optimize later. If you screw up the design and interface part, nobody would be using it - or worse, they'd be using it in ways it wasn't supposed to be used - and then it doesn't matter how fast it is.
it kind of happened, he went through seven interviews. from the same post:
> But ultimately, should Google have hired me? Yes, absolutely yes. I am often a dick, I am often difficult, I often don’t know computer science, but. BUT. I make really good things, maybe they aren't perfect, but people really like them. Surely, surely Google could have used that.
I don't think most people who behave in this manner have enough self-reflection to write something like that. They would rather write that they are opinionated, principled or decisive or some other bs.
Exactly this. No amount of cred, smarts, and genius that ends with "and I'm a bit of dick" will save you from my automatic red-line veto when hiring. I'm far from alone in this.
None of those people ever applied for a job at Big Corp where one of the most important aspects is to be able to work well with other people and tactfully navigate the social structure of the company.
Nobody is saying it hinders their prospects in general. They're just saying that "being a dick" is incompatible with a specific kind of job: one that requires collaborative and cooperative work with other people and navigating the social hierarchy of a company.
In my experience, it's the nice people that get fired and the assholes that get promoted. It's not exactly a secret that silicon valley is full of arrogant assholes.
I mean, he's also the same guy who apparently thought "Unix ideas that have worked for literally decades, nah fuck that. I know better".
It took over a decade before the project made some improvement on how the default install path is handled.
To my knowledge it still has absolutely atrocious dependency resolution relative to things like DPKG.
Not hiring this guy is honestly like a fancy restaurant not hiring the guy who comes up with the new McDonalds obesity burger special menu. What he created is popular, it's not good.
Google is not a fancy restaurant. Five-guys private consultancy is a fancy restaurant. Google is the McDonalds of all McDonaldses, it makes software that is used by everybody, whether they want it or not, and you can't turn a corner without hitting something they control.
Homebrew is “just” a package manager, not the core of part of Google. They could rip it out overnight and won’t even notice it. And Google gave him a fair shot.
This guy got rejected by some automated system without even interview.
> Homebrew is “just” a package manager, not the core of part of Google.
Yes, this. Sometimes I wonder if those coming up with the Homebrew example have any experience whatsoever with software development. I mean, sure the project is popular and surely doesn't hurt on a resume. But does it showcase any level of technical expertise or mastery? No, I'm afraid not. I would bet that the majority of software engineers would be able to put together an equivalent system in a week or so. Think about it, and pay attention to what are the system's usecases. It's hardly rocket science.
>does it showcase any level of technical expertise or mastery? No, I'm afraid not.
But it does show that he can develop and ship a popular product, something outside the capability of so many “great engineers”. Good luck generating any revenue on the backs of smart engineers that have no stomach for understanding the nuance of development over and above writing and checking in code.
It says nothing about fitting office culture of Google, or generating successful projects, for that matter. If he’s so good, how come there’s nothing else besides brew?
Not inaccurate; it's perfectly understandable to a native English speaker, the nuance is subtle, "this guy", "him/that guy", but it is clear and commonly used language.
"I just spoke to a guy about X, his opinion was different to the guy I spoke to about it last week. This guy said Y, but that guy insisted it was Z."
I was thinking about this the other day. I think it might just be a thing of Google looking for a different thing than what made his open source project famous.
Without no knowledge of the details further than mxcl's tweet; probably any performance issues even on simple code, get infinitely multiplied when running at Google's scale, slogging the thing, on Google's dime. From what I've seen of him, mxcl is good at designing a really approachable product, and on running an open source project. But homebrew is really slow, even on the latests Macs, even for basic cases.
To me it seems then that he'd be more fit for a product owner/manager position than an engineering one, and that could be the root of his not-hiring.
This story feels after all the years still awkward. Many people at Google don't have anything that impressive on their resume like being the creator of homebrew. Commentary like "Google looks only for computer scientists, so you need to have studied CS" is so out of touch that i sometimes questions if these people ever worked in a big corporation. There are thousands of different roles, many multiple times suited for that guy. I suspect the people who vetoed didn't like that guy for some other reasons.
Everyone should. The only way to balance corporate power is collective action by individuals, and sharing information is a requirement for that. Corporations can't get away with quite as much brazen sociopathy if their actions are transparent and reported without - or a different - spin.
> Just change the license. The company probably won't notice and keep pulling the new changes. Now you have a legal case.
I think the world already grew tired of rug pull tactics. If you want your reputation to go down the crapper with a lame attempt to shake down an end user, go right ahead.
Hey, I'm the author of the blog post. Thank you for submitting this. If you have any questions feel free to ask and please let me know how the writing was. It's one of my first posts so I'd like to improve
You should change the license to AGPL and 'custom, contact for payment details', and provide a link to this as why you did so.
Simply put, anything not a viral license like GPL allows parasitization by companies effectively living off FLOSS devs, with absolutely nothing to gain. Human rights under GPL were meant to apply to humans, not '3 lawyers in a trench coat' (corporations).
They can make their decisions (snubbing a dev of code they deem good enough for enterprise). And you can make comparable decisions, punishing them for the sheer hubris.
It also reaffirms that my decision of AGPL for everything is the right one. They can contact for custom terms.
> It also reaffirms that my decision of AGPL for everything is the right one. They can contact for custom terms.
Since your replies below are focusing on compensation: have you actually made a nontrivial amount of money with that model?
I would expect that should be a prerequisite to reaffirm it was the correct decision, especially if you're giving unsolicited advice to strangers about how they should license their software.
Some people want others to freely use their software and choose MIT precisely because it's more free than GPL. There's nothing wrong with just making something for free and giving it away if that's what you want. Not everybody has to be chasing money in all their activities.
The author said he was proud of this outcome and nervous at how widely his hobby project will be deployed. That sounds like the ambition of many open source authors and a win. Might never have happened with GPL.
> Some people want others to freely use their software and choose MIT precisely because it's more free than GPL.
MIT license is absolutely not 'more free' than the GPL.
In fact, MIT means you give up effective ownership and control. You lose control and contributions.
And what do you get for that loss of control? Exposure. Or, in this and many other cases similar, you get diddly shit. Some company paracitizes your code, sometimes even demands SOC questionnaires and 'do this bug NOW', and other abuse.
> Not everybody has to be chasing money in all their activities.
Talk about missing the point! This was all about money. It was about a job at the company where the code is being used in a production manner. And they didn't even bother to give an interview.
And not many of us are independently wealthy, and can do things that we want with no monetary care. And, most FLOSS devs aren't that. Instead, they're being used as unpaid stepping stones so some overvalued AI hypesquad can vibecode (or slotmachine programming) faster.
> The author said he was proud of this outcome and nervous at how widely his hobby project will be deployed. That sounds like the ambition of many open source authors and a win. Might never have happened with GPL.
That's where I hope the author relicenses as LGPL and proprietary, and doesn't give Anthropic any more free professional work.
And if it never would have happened with the GPL, gasp, they would have had to pay developers to create it.
And until I'm independently wealthy, I too will license AGPL. If you're making money on my stuff, I want a cut. Simple as that.
>MIT means you give up effective ownership and control. You lose control and contributions. And what do you get for that loss of control? Exposure. Or, in this and many other cases similar, you get diddly shit.
Isn't that what true freedom is?
You can argue that more freedom is a net burden for both the individual and society (tragedy of the commons), but that doesn't negate the aspect of it being more free to begin with.
>And not many of us are independently wealthy, and can do things that we want with no monetary care.
Indeed. But not many people contribute to any kind of OS community to begin with (regardless of the license). I would like to one day, but then the industry laid me and hundreds of thousands off in the last few years and those plans were delayed.
There definitely is a certain level of privilege in being able to provide knowledge to others on the side. Even morose if you're part of an organization that pays you to do so.
> And until I'm independently wealthy, I too will license AGPL. If you're making money on my stuff, I want a cut. Simple as that.
A. So much for "Not everybody has to be chasing money..." as missing the point
B. What hubris to claim that just because you wrote something it is now "yours" in any meaningful way. The copyright lobby has infected everywhere.
So, what is your net-worth, in that you fight for freebies to corporations? What net worth should I strive towards so I can be nonchalant and passe about money?
I'm certainly not there.
Also more curious, is the AGPL doesn't affect humans doing stuff. It affects companies when they grab, modify, and host and not share contributions. Read about anti-TIVOization. That's why the AGPL. I'm guessing you know this, and why you're attacking my viewpoints as 'missing the point'.
And yes, copyright is everywhere. And the GPL has some of the sanest terms to reuse, as long as you follow the requirement. And the GPL also further grows the ecosystem, due to virality.
But Anthropic wasn't exactly submitting code either, were they? In my world, parasites get antiparasitic drugs.
You're mixing events. Tivoization resulted in GPLv3, while AGPL emerged as a response to SaaS.
> So, what is your net-worth, in that you fight for freebies to corporations? What net worth should I strive towards so I can be nonchalant and passe about money?
I've seen people with un-stressed about money with net-worths that are orders of magnitudes below those that seem to obsess about it.
Your motivations are your motivations, if you don't like the idea of someone using your work to make money without giving you a cut, you can do you, but why is it hard to understand that other people might just not care that much about it (or, gasp, even find their work being used more rewarding than the potential monetary compensation)
> Also more curious, is the AGPL doesn't affect humans doing stuff.
It does affect humans doing stuff that isn't malicious, like if you need to solve a problem by modifying the code then now you also have to make that change public which is a hassle, I'd rather not have to track or maintain such things. I'd rather not have to think about that, and I care more about such nuisances than I care about the possibility of companies stealing it.
The scenario you are describing (discovering a problem to fix, being able to fix it, but then not sharing the fix with other people) is the exact reason why GPL has been invented: to force people to share their work, so that we can all have better software, together. Maybe the software you are using wouldn't have been that good if other people weren't forced to share their improvements. Your small effort is going to help others, and their small efforts are going to help you even more. This mindset of sharing should be natural but, as you just proved, people are lazy and so the license has to force them.
Do you feel like Claptrap did?[0].
In all seriousness, good work. Sorry about the rejection, but it reminds me of the story about the Homebrew guy getting rejected by Google[1].
[0] https://youtu.be/hDzWw5rfefQ
[1] https://x.com/mxcl/status/608682016205344768
Also Microsoft and AppGet:
https://news.ycombinator.com/item?id=23331287
I have fond memories of playing Claptrap in Borderlands Presequel. None of my friends do though, his vaulthunter.EXE ability made few friends.
As discussed at length on this page at https://news.ycombinator.com/item?id=44808807 .
Cool. Thanks for the link, but I wasn't actually trying to steal anyone's thunder, and ... I did read the article. Just felt that it wouldn't hurt to link to it.
Also, that discussion gets pretty mean. Didn't feel like I wanted to send people there. I just wanted to give the guy a pat on the back, and bring some humor into it. Been there. Sucks.
I don't know--As a non-celebrity tech worker, it's actually kind of comforting to see a company that doesn't just automatically roll out the red carpet and grease internal wheels, just because a candidate once wrote some very popular software. It sounds like there are still companies that make you go through the same whiteboard hazing and non-deterministic hiring process as the rest of us mere mortals, regardless of how well known you are.
Actually, I don't disagree with some of the people that had issues with him, but I do have issues with folks that refuse to look at past performance, in general (I'm biased. I have a great deal of past performance, and can prove it).
It was just a kind of nasty conversation, and I didn't feel that it was appropriate to deliberately send folks there. I'm not really into the whole "Make the Internet Darker for Everyone" schtick.
Hey, great work, and just wanted to lend my voice in support! It's kind of wild how many open source devs have a story along similar lines. (Mine is the time when Mojang used my voxel engine..)
Now that this is trending on Hacker News, surely there will be a happy ending when someone from Anthropic sees this post and hires you with sincerest apologies and everyone lives happily ever after? Can we get a positive story out of this, universe?
It might have been the motivation behind the post in the first place, though not without risk.
What's the risk? The current state is "your application has been thrown into a fire and will never be seen by human eyes". How can it get worse than that? There is no downside to complaining on HN, except for getting the reputation that you really wanted to work there, which, again, isn't that negative of a thing.
Do you think that making your product AGPL would being you more money/recognition/jobs for your effort?
I don't know. I have no comparison but it is common for crates to be released under MIT. I took over the maintainership from the original author so the license was already there. I rewrote pretty much everything so I guess I could try changing the license now but that's not something I wanna think about.
I do the work because I see it as payback for all the great open source software I use all the time.
I really like the copyleft idea, however, I think you did nothing wrong, IMHO, because if large corps like an idea, they will rather reimplement it rather than even bothering with ways to conform to AGPL or buy an alternative licence. Particular in the age of AI, all source available code has become pretty much public domain (value is still in maintenance, etc). License have mostly become a compliance/ideology game that alienates most people. However, changing the license on the main repo, with only a minor version bump, would be a nice asshole move to get their attention past HR (won't make a difference, but if you have nothing to lose).
Copyright is but one pillar of intellectual property law.
I’d like to see an attempt by useful freedom respecting software projects to deploy patents to combat non-free reimplementations.
A GPL license that grants you rights to the backing patent as long as the software you develop with it is also released under the GPL license.
Use the library for closed source software? Copyright violation. Reimplement the software under another license? Patent violation. Create something slightly different and call it the same thing? Trademark violation.
Not sure of the rest of the world, but at least in the US, patenting “software” is a pretty murky subject legally (at least it feels that way when trying to do some basic research on it) Something that seems common among sources discussing it is that “Software Related Inventions” (eg, a computer that does XYZ) can be patentable, but software/code itself is not literally patentable. Seemingly, because we’re talking about libraries that would be pure software, not a product for sale based on it, you wouldn’t be able patent libraries like you’re talking about.
I’d provide links to some discourse of this, but honestly I think it’s better to search “can you patent software in the US” and do a brief read of various sources, because the terminology between them can seem somewhat counterfactual to eachother.
Copyright mostly protects big corps nowadays. That's because you need lawyers to enforce copyright, and if the other side has more money the battle may not be worth it.
On the other hand, Meta was found torrenting terabytes of books and for them it's a nothingburger. The rules are really meant for commoners.
> but it is common for crates to be released under MIT
Something that isn't brought up enough in the "rewrite everything in Rust" discussions is that the API guidelines explicitly recommend MIT/Apache to "maximize compatibility" (i.e., corporate friendliness, or developer and user exploitation): https://rust-lang.github.io/api-guidelines/necessities.html#...
Your project has been around for a while, but it's crazy to me that anyone still open sources anything under MIT (or similar) in the era of LLMs. Are they that confident in their job security? Are they already independently wealthy? Frankly, even a proper copyleft license is likely to just be ignored, or the code laundered through an LLM-assisted rewrite, by these companies. I prefer to just keep anything I can't sell all to myself rather than release it, at this point.
Hey, I really liked the post and especially the title. Quite surreal but also very fitting at the same time. The writing was great too. Hope you keep going. I’d love to read more.
Since you're asking: I took a pause mid-reading and told myself: "Woah, I like their writing style."
Wow, what a great compliment. Thank you :-)
Thank you for your service
Fun write up, lovely irony (if your work did actually help AI auto-reject you).
If I was you, I would probably feel similar "you used my project, you probably want to hire me!"
But there's a logical fallacy there.
Your creation being useful to a person or company ≠ you being a fit to work with/for them full time.
Still, you deserved human eyes on the question from their side.
It was a fun and easy read
This lands. I discovered an emergent feature in GTP40 and when I tried to post about it on the developer forum, the spam filter removed my post. I asked GPT40 to rewrite it for me. I posted the update, and got banned. There's too much 'noise'. People like Einstein and Tesla would've gone unnoticed today, as I doubt they would've become "social media influencers" just to promote their ideas.
edison was pretty good at self promo
Absolutely. He had a rare combination of skills and personality traits. He was good at sourcing unorthodox talent too -- he employed immigrant Tesla!
I wonder if it's useful for you to put a few subtle "hire me"s on your repo, mailing list, ect?
If we are at the point where a hiring manager for a position deeply related to an open source library is not at least checking if the authors would be interested, I'm not sure.
I honestly think this is some system failure, even a Claude based one. I hope someone in the Claude Desktop team sees this and reaches out to you. Cheers!
I hope that was just an auto-reply rejection, that it got caught in the HR bureaucracy, and some human developer sees this and re-considers.
Hey mate, I would just like to say that I wish they at least find it in their hearts to reward you for the value you have provided to them. Knowing cut throat american corps, I'm afraid the chances are nil. Even if a good amount for you is peanuts to them.
Which is why my position is GPL > MIT..
They could literally give him 100k, 1mil or even 10mil which would still be a rounding error in their books.
Don't know. A company can have a huge valuation on the stock market but that does not necessarily mean that they have cash to pay wages or can afford to pay a large team. If all they have are stocks they have to find somebody that buys those stocks with cash, then find a way not to run out of those money before selling more stocks. Eventually do an exit and stop worrying or become profitable.
they are raising another billion dollar round
Are they even profitable?
Profitable schmofitable! But seriously, that is orthogonal to whether those figures are rounding errors at anthropic's financial level.
I have always preferred permissive over copy-left, because I've historically been unable to use packages at work, which puts food in my mouth, as a developer who spends some time contributing to projects, especially those that I use at work.
This has changed everything. AGPL and GFDL from now on.
you're right about MIT vs GPL confusion. people brainwashed themselves into thinking MIT is "more open", because it's more permissive, but it lets others profit off your code without contributing back.
GPL makes them share or pay to relicense, since you own the copyright. with MIT, they don’t need to ask. MIT just benefits big corps. GPL better protects the open-source spirit, and paradoxically, the ownership of your work.
And yes, people being able to use your code for whatever they want is absolutely more open than having restrictions on how/who gets to use it.
One other model that can also work well is to dual license as GPL + commercial, so people who want to publish their work can use the GPL license but you can potentially fund the project from license sales to closed source users using the commercial licensing option. I see this a fair bit in the audio community I work within.
Why would it be unfeasible to just share the code parts that are GPL?
If you link against GPL code, your code needs to be GPL compatible. There are some IPC based workarounds, but they are too annoying and slow in most cases.
LGPL exists too.
Sure, but the thread was initially about GPL.
>And yes, people being able to use your code for whatever they want is absolutely more open than having restrictions on how/who gets to use it.
Yes, this is why people should use free not open , and GPL is more free when you report to the entire community otherwise you are in the famous case from a story where an USAian was claiming "Amerika is the land of the free, we are free to own slaves"
Yeah, basically MIT is "more open" in the short term, while GPL is more open on the long term. GPL, while restricting some freedoms right now, is actually enabling the remaining freedoms to be sustainable in the future. Very similar to how law enforcement works out with regards to a sustainable society, and how market restrictions work out to create a sustainable and diverse market.
Can you please send me your resume:
Andrew@gambit.us
Are you the head of AI at a military contractor? This is probably information you should disclose when asking people to send you their resumes.
I'm disappointed about your resigned, almost subservient tone. This company is profiting immensely off of your work, and they don't even give you the courtesy of a job interview?
~~Have you considered a copyleft licence like LGPL?~~ Answered in a sibling comment
> This company is profiting immensely off of your work
I wouldn’t say that’s exactly the case. Not to denigrate the author or anything, but this library is a relatively minor part of what Anthropic is doing. It’s a UI manipulation library, specifically one that simulates keyboard and mouse inputs. While something like that is certainly necessary for the project in question, it’s not anything that couldn’t be rewritten in-house without too much difficulty, especially since they’re only using a subset of the platforms supported by the library.
I’m sure that working on this project has provided the author with expertise in this area that Anthropic could benefit from, and so in that sense it’s still a shame that they wouldn’t give him an interview, but that’s really all that can be said about it.
Also worth noting that NONE of the AI companies are profiting at all, let alone “immensely”.
Google is, but not from AI.
> it’s not anything that couldn’t be rewritten in-house without too much difficulty
This is my experience, at every group I’ve been in. Extending the date a bit is much easier than involving legal for approving a new library.
The group I’m in now sunk a substantial amount of money into a lawsuit for a library that accidentally made its way in, so are now “No LGPL.” with some crazy loops and approvals required if there’s really no alternative (very rare). From their perspective, it’s cheaper and safer to rewrite than not be in compliance, unintentionally or not.
You have to think about other users as well. One person taking advantage of you doesn't mean you have to cut off all the people not taking advantage of you.
Expecting a reward from open source software is a recipe for disappointment. I have contributed code to projects by companies that say I'm a mentally-ill household object. I'm not going to change the license of my open source projects to get back at them, because the collateral damage against entities that aren't evil simply isn't worth it. (It's also somewhat unlikely that the people working on NTP servers at Facebook wrote those policies, so...)
> Through a friend of a friend, I found out that Anthropic had an open position in the team implementing the secret, unreleased feature of Claude Desktop using enigo. I wrote a cover letter and sent out my application. An automatic reply informed me that they might take some time to respond and that they only notify applicants if they made it to the next round. After a few weeks without an answer, I had assumed they chose other applicants.
I've mostly stopped applying to the big companies long time ago (10+ years) precisely because I'd never hear back regardless of the match or the credentials.
The only exception has been JaneStreet — they've contacted me almost immediately after a cold application with a small cover letter about my interests.
Yet going the referral route, it's relatively easy to get an interview almost anywhere, even Google or Apple.
In my 6 jobs in my career from college internship to startups to Big Tech, I have never gotten a job from sending an application into a site. It's always been through (somehow) tracking down a person to speak to and get a referral.
I would recommend that here. There's no reason why Anthropic couldn't use your talents! See if you can find a friend-of-a-friend who is there, and then do a phonecall with them.
This is the right approach. Connect with a few hiring managers.
>The only exception has been JaneStreet
Huh. I guess if you decide to make OCaml your company's primary programming language, you have to take what you can when it comes to devs.
They put an absurd about of effort into their recruiting, so I'm not really surprised. Pretty much every math-related content creator I listen to has advertised for them.
I doubt anyone who works there is "take what we can get" calibre. They want to attract people who casually solve college-level math puzzles for funsies. So I imagine it's the opposite and if you get hired there, you're surrounded by people who are extremely accomplished.
Yes and no, I imagine the biggest qualification for Jane Street would still be humility and not OCaml interest or expertise, and the pay probably has something to do with people's desire to apply, too.
I wasn't talking about why people would apply, I was talking about the statement that they responded fast to a cold application. Generally speaking, if you drop a ton of applications for what I assume is a lack of willingness or skill to deal with OCaml, that certainly factors in your hiring process.
Not accurate. They pay very well, and their interview process is supposedly quite challenging.
A thing to consider, though, is ethical: they seem to have been involved in market manipulation. [0]
[0] https://www.bbc.com/news/articles/c5y0zgrevl1o
I'll say it: why would they pay him if he's already doing the work for free from their PoV?
Oh, they ignored him. I am not sure if that puts the company in a better light.
Lets make a new license: If you wont hire me, use my library and make over $100m in revenue a year, you must pay a commercial license to use my software equivalent to the total cost (equity grants included) of an average principal engineer or director who manages 50+ people at your company in your highest COL metro, whichever is higher. For OSS work that isn't mostly one author, make it go to the foundation for the OSS project instead and apply the rule to principal maintainers. You could even scale it in multiples of revenue in principle engineer units of $1b per principle engineer of global revenue.
IMO I think foundational projects that every single bigtech uses like ffmpeg should get on this licence yesterday. They would start getting millions because it still would be way cheaper than making it themselves in their bloated cost structures.
I agree with the spirit of this comment, but I worry about the implementation.
See the comment of Manly read in this section. Once the threat of payment approaches, you can just switch to a free fork. A single person can't really win a trial against a big, well-funded company.
I don't really see how this is an issue, depending on the license text it's trivial to make the license apply in the same manner. As for winning, I think that's more of a US-centric view, if you sue elsewhere in the world there's plenty of courts that are happy to slap big tech.
Genuine question that might sound trollish: do you please have examples of such cases?
I think Jacobsen vs Katzer [0] is the most relevant one to the discussion here, but there a number of successful cases on this front. If memory serves, BusyBox has also managed to enforce GPL in court on a number of occasions.
[0] https://jolt.law.harvard.edu/digest/jacobsen-v-katzer
Suing a major corporation still seems like quite a bit of work, and what's the end goal? Is it to humble a major company, or to get paid? Because if it's the latter, it feels like there are easier ways to do so.
This is obviously a subjective opinion, but at least in my mind, the point is to defend your rights. No one else is going to come along and defend you against the corporate steamroller.
They can fork it, but can they find the maintainers? If it's just their own internal employees, then they definitely have less expertise in that codebase.
Might as well hire the actual expert.
If the license were copyleft forking would not be a solution.
That would just open the door to commercial competitors to undercut the price by reverse engineering it (https://en.wikipedia.org/wiki/Clean-room_design). You'd be playing on the proprietary software industry's home turf and they will straight up curb stomp you.
It’s a nice idea but couldn’t a big company simply move its engineering team to a subsidiary that doesn’t get sales revenue?
(I’m not an accountant!)
Would be hilarious to bury a clause like “Modified MIT license — head of HR must publicly announce any employment application rejections of the maintainers while wearing a chicken suit).”
Couldn't they just get Opus to rewrite the lib?
The model probably has the lib in it tbh.
Because - They can decide more easily what he works on - They know he loves this work and is very capable of doing it - They can own his output, a competitive advantage - He will likely cost them ~nothing anyway
While that sounds rational, I worry that the same reasoning is not applied in the HR department.
But that might be just my frustration from experiences.
To continue the devil's advocate: why bother with all of this, if the company doesn't have to and the OSS version is enough anyway?
This is an incredibly short-term thinking. The reason is simple: the author is not obliged to continue while this sort of thing can be demoralizing.
I don't know about the author's approach to this matter, but if I would find out that a company is making a killing using my software and then that company would refuse to even give me an interview I'd probably stop loving doing what I do. Sure, the software is under MIT license and it was the author's choice to do so, but what's the point of doing it under such a license when you can't even count of it mattering in a resume? What's the point of providing free labor to a company with revenue in billions? If you look at the author's blogpost, the only benefit the author mentions is making the number of downloads go up and that's just pathetic.
I am reminded of an another, similar case with a library called "FluentAssertions". This library used to be free to use by anyone until the author changed the license and started charging money for commercial use. The author did that because he spend several year maintaining the library on his own time and dime and megacorpos like Microsoft wouldn't even bother to donate despite using it extensively. What happened afterwards was that the author got shat on by everyone on the internet for daring to ask for money. In the company I work for his library has been replaced with an another free fork at a incredibly fast pace. All that free labor and the author got dropped as soon as they fell out of line.
The worst thing is that it wouldn't probably take much to make the author of the library happy. Even if they weren't interested in hiring him they could still acknowledge him, talk to him a bit to maintain good relations, throw him a nice donation as a thank you and now it would be a nice, good PR story instead of an another reminder that corporations are just looking to squeeze out value out of all of us.
Is it short term? Seems like the MS and the others got exactly what they wanted.
Exploiting the passion for free work is a trade that will keep happening as long as there are passionate inexperienced people.
Sorry but have to call b.s. here. Many of us did in fact, in this very forum, kept pointing out that dumping on GPL (and the man and his ideas behind it) was a mistake and that non-gpl oss was for chumps. And we were greyed to oblivion. Same exact story as with surveillance tech.
Hackernews is hugely responsible for many of the ailments of this field in 21st century. Own it.
I don't think we're in disagreement. And it wasn't me greying people out. But yes, I see what you mean by the "spirit of HN".
> Many of us did in fact, in this very forum, kept pointing out that dumping on GPL (and the man and his ideas behind it) was a mistake and that non-gpl oss was for chumps. And we were greyed to oblivion.
I'm not entirely sure that was intentional. On Reddit, it would be called 'brigading', and basically getting your corpo-techbros to -4 and flagkill posts.
If done fast enough, you only need 5 500+ karma accounts to sink a post.
Sometimes, I'll say something unpopular, but defensible. Its interesting to see the dramatic swings those contentious posts take.
The fact that the management of this forum, who are VCs, permitted such a mechanism is part of the "own it, HN" assertion. HN has baked in something like 'peter principle' into the forum. Karma Grifters who post articles that get them over the 500 points are then ala Peter Principle granted the right to 'grey out' actual thoughtful comments.
Are you telling the management of this outfit never looked into this phenomena?
How about ageism? Mr. Paul Graham and personality cult asserting that anyone over 20 something is no longer viable for leading edge tech work?
We used to call these VCs "vulture capitalists" in the 90s. We geeks were so right about so many things in the 90s: We were right about GPL. We were right about VCs. We were right about surveillance tech. We were right about outsourcing ...
But alas, "corpo-techbros" empowered by thoughtless forum software courtesy of Paul Graham and company got into this mess.
> Are you telling the management of this outfit never looked into this phenomena?
To counter, I think that HN is being used as a testcase to shove techbro and VC ideology across all of tech. And secondly, its some of the most potent tech market research. Its a textsearch goldmine.
I believe YC knew what they were doing, and intentionally chose this course of action.
I'm guessing you're not in the VC or founder club. I only found about that https://github.com/minimaxir/hacker-news-undocumented
> Additionally, founders of YC companies see each other's usernames show up in orange, which — although not an explicit benefit — does allow fellow YC founders to immediately identify one another in discussions.
Even with the significant bias here, I still read it. I also read lobsters as well, which is here minus techbro insanity.
> I believe YC knew what they were doing, and intentionally chose this course of action.
Seems plausible. They certainly are not dummies. They just have a different 'value system'. So, yeah.
(Thanks for lobsters tip. til.)
They definitely know now, even if it wasn’t the original intent.
I guess the author can learn their lesson and not use a permissive software license which lets behemoth corporations do exactly this.
It's very sad, but the resigned and almost subservient tone of the author does not lead me to believe a lesson has been learned.
They probably already have someone working on top of it but it's just closed source because of the license.
But they are hiring someone else to work on it.
> I found out that Anthropic had an open position in the team implementing the secret, unreleased feature of Claude Desktop using enigo
Right now it is just a hobby and there are still a number of bugs remaining. Since I don't have an income from it, I can't dedicate more time to it. Hiring me would allow me to work on fixing them full time and make the progress much faster
Hey, props to your attitude, and I wish you the best of luck.
Obviously, you've provided value to a company in a really in-demand area. It doesn't feel right to treat the contributors like this. Sadly, it seems that the companies have the power and the intent to just abuse and exploit
I don't have a solution. I am just expressing my frustration from the perceived injustice.
Do you think they already have hired someone to work on it but are just not releasing the source code?
I don't think so. They use an outdated version straight from crates.io (at least in the publicly available version of Claude Desktop).
Seems like it's time to remove links to outdated versions. Replace them with your resume?
I believe crates doesn't really allow that, partially so that people can't easily sabotage the supply chain like that :)
Correct, you cannot remove a version or the whole crate unless very specific criteria are fulfilled. You can "yank" versions. That prevents people from adding the version as a new dependency, but if you relied on it before it got yanked, your build will succeed.
I wouldn't delete old versions even if I could though. My goal is to publish a rock solid library that everyone can depend on and build awesome projects with
Why wouldn't a company want to hire the fore-most expert in a tool that is critical to the company? They are hiring for someone with that exact expertise.
A competitor could hire the OP instead, get them to work on improving the software for a few years. Giving the competitor a major head start.
Worst-case scenario, the tool they are building doesn't work out and Anthropic has a pretty good developer to put on other projects.
Implementing the features they would want to prioritize. Just like most companies hiring OSS maintainers.
That is also a good point, but I worry that the power has shifted. I worry that companies might get away with no compensation for such efforts.
Only through trickery. E.g. "you might get a job if you work for free for us". In other news see many tech job ads these days :)
I can't say for Anthropic, but I've seen Google hire people working on open source projects that were aligned with the skills they were looking for. Desktop search and collaborative editing comes to mind, although I might be mis-remembering?
That's like having someone say gives you free tractor blueprint but you can hire its inventor to come and put it together, or some other engineer.
An FOSS project is rarely production ready that is really free as in beer considering TCO. Especially for a tech company.
But he's not doing the work for free. He's doing something else for free which they use. He has domain knowledge with the library that noone else has, their can either pay someone to learn it or they can hire someone with it.
And that's exactly what these companies can abuse.
If this wasn't available for free, they would gladly pay for a programmer to create it. But if it's already free, they can use it as a starting point. Maybe they'd need to internalize/extend it. But the option of paying for the work already done is gone.
Do this for each npm dependency and you're looking at huge savings.
This is still illogical. You can hire the original maintainer and pay an incremental cost, or you can hire a random developer and pay the initial learning cost + higher incremental cost.
If every company using a library chose the former, then every hour of development would be paid for (from the perspective of the maintainer) and the cost would be spread out across all its users.
A counterproposal:
You can use what is as is. Then you can ignore all of the other issues if they don't impact your bottom line.
Don't get me wrong, I like your corporate OSS financing model. But there seems to be not enough incentive for companies to use it. Why take ownership for a small cost, when you can use an imperfect thing with no cost?
The author should have just asked the friend of a friend for a warm intro instead of trying to go through the main gate.
Sucks, but that's the reality of hiring (and getting hired) in tech in general.
Was wondering something similar, if OP had blogged it earlier when he found claude was using it and re-posted it in HN/reddit it in a sensational way to capture eyes. Maybe through one of the forums he could have got an introduction and a job doing what he loves.
OP still has a chance now, maybe not anthropic, even other competitors can come knocking.
> Sucks, but that's the reality of hiring (and getting hired) in tech in general.
If you’re in the inside, it doesn’t suck at all, it’s so much safer.
Hiring a new person, based on a few hours of interviews, and a resume half full of exaggerations and lies, is such a ridiculous gamble. Worst part is, if you realize they’re not a good fit, it’s sometimes incredibly hard to get rid of someone, more often not an option at all.
My experience with Anthropic and OpenAI is they’re not super interested in experience and don’t take internal references very seriously. Most of who I know that was hired were fairly junior folks and they have an early weed out that’s a fairly rudimentary but very specific Python programming quiz typically administered by very junior (like 1-2 years out of school) - even when interviewing extremely senior and experienced people of some substantial success and renown. This isn’t uncommon - meta and others do this too. But the programming quiz at Anthropic is sudden death and the first round, and the people administering it are looking for a very specific implementation that if you don’t see it immediately they just Gen Z stare and don’t discuss etc. It’s one of the more amateur selection processes designed with an extreme bias against more senior folks (frankly it felt unintentional just naive). (Meta etc scale the programming weight to seniority and the administrators scale as well - asking for depth of understanding of concepts as seniority grows with the expectation experience brings more to the table than syntactic knowledge).
So getting an internal reference and being highly qualified for something they need done isn’t enough. You need to also make it past the 20 years old gate keepers and their amateur hour hiring process.
Yikes this sounds awful.
Unfortunately, the choice of license likely won't matter in the nearest future (if not already so). If a tech giant wants you open-source library, they will just point their agent to it and ask "to rewrite in the style of War and Peace". And more unscrupulous players won't even bother with a rewrite, as we've seen recently in the case of Cheatingdaddy/Pickle.
what has been your experience with asking AI to create a complex project from beginning to end?
The flip side is that if they can technically pull that off then the cost of writing the library has dropped so low that an OSS maintainer probably wouldn't have to work too hard to write it anyway.
Being able to rewrite existing working code sufficient to copyright-launder it isn't the same as being able to write it from scratch, unfortunately, especially since LLMs seem to be allowed to ignore quite a bit of copyright law with complete impunity.
Imo it's totally plausible that something will be expensive & time consuming to create, even with LLMs, but still easy to fork outside current licensing restrictions with LLMs.
Rewriting it with a guarantee of not introducing any errors is still beyond current LLM capabilities, and there might be a certain correlation between that capability and the capability of writing it from scratch.
>If a tech giant wants you open-source library, they will just point their agent to it and ask "to rewrite in the style of War and Peace"
Is there any evidence of this happening? And any legal theory behind how it might have the intended effect? Training being fair use does not make AI a magical copyright-removal box.
AI is already "too big to fail" in every conceivable way. If there is a conflict between AI and the law, then the law will inevitably be rewritten, reinterpreted or repealed in the way most beneficial to the service of AI.
Unfortunately, this seems on par with recruitment practices in the summer of 2025.
I can almost guarantee that they didn't even read that application / cover letter and auto-magically rejected it.
"the team doesn't have the capacity to review additional applications"
Zero effort. They probably didn't even realize the relevance of that specific application for that role. Unbelievable, I swear!
Tbf, I'd rather get a "we didn't review your CV" response than a template "we are continuing with other candidates :)" response. It softens the blow considerably and helps me as an applicant better keep track on which variation of the CV is working best because I can just remove this datapoint.
> Tbf, I'd rather get a "we didn't review your CV" response than a template "we are continuing with other candidates :)" response.
Idk it sounds plausible that OP might just have been late to the party, and applied when the recruitment process was at the final stages.
I had a hiring manager tell me a couple weeks ago that I probably made it further because I applied late. All the people using AI and automation to apply are hitting the apply button early, and the laggards tend to be humans.
Its a perpetum mobile. Hiring managers use automation to filter candidates, coz its too many. Candidates see they dont even pass automatic filtering, so they apply with tailored CVs x10. This means even more CVs and more filtering and more CVs and more filtering and more CVs etc etc etc
Im curious at which point ppl will understand its counter productive.
Never, because they already know it is but fixing it requires both sides to deescalate in lockstep while individuals on both sides would benefit from not descalating
Unfortunately, "we didn't review your CV" is a great way to get sued in the US if the name of the applicant is in any way potentially indicative of the applicant's gender, race, religion, or any other protected status.
And we now see a lot of proof that that is warranted: lots of people in the US are now openly racist and/or have other prejudices. It is fair to assume many people don’t get a fair chance because of this.
The first company I worked for (a smallish business) interviewed a guy who seemed normal and knowledgeable enough in the screener call. At the time, they just asked that people being in code samples to discuss in the second interview (this predated GitHub) and he brought in some obviously copy-pasted code that didn't fit together.
He then, without prompt, in the middle of a conversation mentioned that he was the second coming of Christ. The interviewers ignored the comment and continued the interview.
When he didn't get the job, he sued the company for religious discrimination. Fortunately, the interviewers could honestly say they didn't discuss or ask about his religious beliefs, and he lost. It was said he did this elsewhere as a a scam, though I never verified it.
The simple matter of fact is that it doesn't matter how neutral you are; there are enough people out there who will look for any way to perceive and benefit from a grievance that you must assume they will.
Of course you're right, there will always be people that try to abuse the system. But the bigger picture here is that more people are truly being discriminated for their color, religion, sex, etc, than there are people abusing the system. A system that improves live for many should not be removed because some people try to abuse it.
We are discussing whether or not it is safe for a company to reply to an applicant with "we didn't review your CV".
The crux of my point is that the potential for a perceived grievance is sufficient to trigger legal action. Whether bigotry was involved or not is entirely beside the point, because even if the company absolutely was 100% not discriminating, they are still vulnerable for creating a situation where they could be perceived as having done so.
In no way am I advocating for removing protections for disadvantaged groups. I'm not arguing that bigotry doesn't exist. There's no point in bringing up the fact that it does.
You know what's cheaper than getting hundreds of baseless court cases dismissed? Not replying to someone with "we ignored your application".
I remember applying for a German company for something and they wanted me to submit a professional photo. For an IT role...
For good or ill -- probably the latter -- headshots are expected with CVs in most of Europe. It's a local custom.
Source: I worked in Germany and had to deal with this. (In fact, one of the ways I made my application stand out from other North Americans was to learn this ahead of time and include a headshot in my original application)
Just a datapoint: headshot are not required in Italy.
Germany is weird. But then again, that's not news.
Yes, euphemism is one of the worst diseases of our time. This is gaslighting with fewer steps. It's almost always easier to lie.
To be fair with Anthropic, they probably get unfathomably many applications for everything, on top of the cold calls/emails. They're one of the hottest companies in the world, so I'd expect tens of thousands of applicants. Media writing about $100m+ hiring deals in AI does not help, either.
Aren't they an AI company? Couldn't they sort it out? If Anthropic, of all companies, can't sort out incoming job applications, what exactly are their tools for?
> what exactly are their tools for?
Obviously, for writing and sending said job applications.
Compute is a precious resource.
The world is also full of totally delusional people who dreamed up the idea of using winzip to compress VRAM on the GPU, and now Anthropic will definitely hire them for $1M year for this genius solution, so better write up a glammed up resume and auto-send it once a day for any open position.
Is it? What I've been marketed is that AI is widely available for every task I could conceptually think of, and that I should be using it for everything.
Opportunity also is a precious resource. Redirecting resume to /dev/null is wasting it. I have a hard time believing that LLMs, with all their sophistication, aren't ideally suited for this task.
I don’t know how slighting customers who want to work for you works out in the long term. You end up getting fewer opportunities from them and their friends in the future.
Seattle is full of people who will tell you what it’s like to work for Amazon and how you don’t want to work there. I guess if you’re big enough though the money papers over a lot of sins. The smaller you are, the more people you can piss off before you run out of prospects. Anthropocene still has a long way to go before they are Facebook, who struggles because something like 50% of the people who would work for FB already have.
IDK, compute isn't really that precious. If $20/mo can get you many (?) invocations of their research agent, I feel like it could pretty easily be worth it to screen applications for jobs that pay $350K/year -- and that's just "entry-level"![1]
That said, their career page puts this at the very top of the details section:
This guy seems rad, but his GitHub[2] and this blog are both light details or links, which is odd considering that his LinkedIn[3] is detailed+professional. Perhaps Anthropic does have Claude screening resumes, but he didn't express the nature of the situation clearly enough for it to catch it?Otherwise, the only other explanation I see that doesn't look terrible for Anthropic is they didn't see a need for more Rust expertise...?
[1] https://www.levels.fyi/companies/anthropic/salaries/software...
[2] https://github.com/pentamassiv
[3] https://www.linkedin.com/in/robingrell/
Tbf the other summer recruitment practice in AI this summer is Zuck running round offering engineers with some sort of reputation $100m+ windfalls, so maybe all the OP needs to do is add "author of computer interaction library used by Anthropic" to his LinkedIn profile to acquire that garage full of Ferraris
'used' sounds weak, "Build technology powering Anthropic Claude Desktop".
Nah, if there was ever a time where making meaningful contributions to open source was important to land you a good paying job in a hot tech company, that died a long time ago. The people making these decisions don't care, unless you have someone inside to put your resume first it doesn't matter that you wrote all the code that makes their product even possible, the hiring manager won't care.
I might just be old but i really haven't felt like contributing to open source at all lately because i've bills to pay and kids to care for and taking time out of this just for the sake of enriching some billion dollar corp that will eat me and spit me out doesn't feel like a good investment for my time.
Sometimes i feel sad that it came to this but this is the place we're living in right now.
they should have used their AI to scan through resume... they are AI company afterall. shame they missed this guy. it shows their resume-scannign AI is useless.
There's an ongoing lawsuit[1] pertaining to AI-driven job applicant filtering.
[1]: https://www.cdflaborlaw.com/blog/federal-court-grants-prelim...
Poor poor... I always felt that too many people in hr decision making underestimate the role of talent. Many awesome software products stem from teams with extraordinary talent. Average people create average software.
It's been like this a few years.
It's bit more AI now and bit less boilerplate rejections.
They should have used Claude Code for reviews
I wonder if it was geolocation? Anthropic is based in SF, the author seems to be based in Munich, and maybe they're not open to hiring people who aren't based in the US right now? Given the state of US visas right now, this wouldn't shock me.
My company, which is significantly smaller, hires people in multiple countries across the world. You don't need an office to hire (I am sure there so exist countries where you do, but I expect they are the minority).
London too.
After Brexit that's still quite a hassle.
Also I low how the IT hiring has a become a paradox: Companies won't hire you if you don't have enough projects in your portfolio, but by the time you will have enough stars on your github projects they have already used you to their own goals and are "not interested".
I think you are making up scenarios in your head to try to rationalize away why you have a bad time at job interviews.
HR hands typed this post
exactly. there is a reason companies are paying through their noses for some people
I don't think is true. I had more success removing my portfolio and letting my work history speak for itself.
I don't know if this is a good opinion, but I don't think it's a good idea for independent individuals to use highly permissive licenses on their open source software. Companies will just suck it up and might not contribute back. It distorts the market because if the software didn't exist, they'd have to hire people, contract it out, etc. somebody would get paid. you've saved a huge company from having to hire people to develop the software they need, which is good for them, but imo just gives the companies incentives to devalue engineers. I also think the value of somebody open sourcing their work as a means to getting a job is questionable and never really been backed up by any data.
Excellent point. If you really want to make your project open out of the goodness of your heart, then use GPL. Otherwise you are explicitly giving any business, no matter how big or wealthy, permission to use it with no expectation of giving anything back. The license says that quite clearly. It says nothing about rewarding the author.
Through a friend of a friend, I found out that Anthropic had an open position
"Coffee" with the friend of a friend would he better strategy than a cover letter in that case...more work, but better strategy.
Because logically, getting hired requires demonstrating you are "the kind of person we want to work with." Being qualified on paper is not necessarily required.
You basically need to avoid putting your name into the main queue as long as possible. Make them do it manually on their side and keep track of it. If they want you, then you’ll bypass all the crap this way.
More fundamentally, companies only hire people they know.
“Coffee” is a way to meet “a company.”
A letter is not.
Even a letter of introduction.
Ah, it seems their AI powered cover letter review system isn't up to scratch.
Being able to get hired at a company is often unrelated to being able to generate viable products.
If you want to get hired don't focus on skills to build useful things. Focus on psychology and charisma.
or just lie
Yeah, no. You have to pass tech interviews and generally know your shit
The tech interview still doesn’t fall into parent posters categories of “able to generate viable products” or “skills to build useful things” either.
Not too often tbh. You either get dogshit LeetCode or, more often, just a general chat about what you've done and know. There, social skills play a massive role. Make simple projects sound like state-of-the-art, present everything cool that happened as something you were directly involved in, present the 2 most obscure bugs in the project as something you fixed every other Tuesday when you get bored...
Social skills are not a substitute for technical skills, but they are still important. You have to work with people.
> Yeah, no. You have to pass tech interviews and generally know your shit
You don't understand. You need to meet the hard skills bar, but there are far more bars than the hard skills one.
I think a hefty share of people here fail to understand the fact that there is way more to hiring a candidate than leetcode.
As someone that works in HR, the incompetent HR combined with using AI for ATS ( or not knowing how to use ATS at all) is one of the core problems when it comes to losing quality candidates and is to blame for this. It should be illegal to hire HR from any education other than law, psychology, management and economy background. That way the responsibility would be larger, the ROI on HR would be higher (because the retention of the candidates and the quality of the candidates). Simply paying and promoting people with any educational background in a HR role is a waste of money which also creates problem for the company and not just employees.
> It should be illegal to hire HR from any education other than law, psychology, management and economy background.
A lot of people with education in management/business do go into HR, at least in countries I know, and it does not help. People with extensive management experience would help but they will only take more senior roles.
The other qualifications open opportunities interesting and well paid careers. How would you attract those people into HR?
I am not even sure it would help if you could.
I think the suggestion in the old management book by the guy who turned around Avis that you should have an old style personnel department to do admin and advice, and managers should have more involvement might be a way forward, but I am not sure it would work given the current level of regulation (in the UK anyway - I imagine most wester countries are the same). A lot of the function of HR is to avoid legal risk (e.g. fire people according to the rules, so go through the motions of warnings etc).
I think that old departments (personnel departments) should have been just modernized in reality. To be frank, in some cases a mix of HR/Legal department is cost saving too.
What it really comes to is that a lot of people love to micromanage everything. If you hire someone that has integrity and educational background in subject, he/she will warn you if the decision you are making will have consequences in the long run. If you have someone that does not have relevant education, that simply does not happen. The managers micromanage, those people receive salaries and if they step out of the line even when they are right, they are reminded that they do not have relevant knowledge in said department (law/economy). This in turn leads to a lot of people gaining something called shallow experience which then in turns leads those people to hire someone that des not pose the risk to their position further down the line.
The problem being in this case is that there are a lot of misses that happen when the HR is organized like that; from illegal hirings, not knowing key economic factors, not having a clue about the business itself, no clue about laws and procedures and so on. Which in turn does not really protect the company because the company loses both the money and employees.
At least Max Howell (Homebrew) got an interview before getting rejected:
https://x.com/mxcl/status/608682016205344768?lang=en
In all seriousness though, the situation sucks. But there's still upside. Someone might reach out.
Not unlikely given the HN attention this post got… Anthropic’s small hit to its reputation prolly isn’t worth the $2xx-4xxk / yr. they should be paying this nice person.
There is some dirty secret i learned in my time as a eng. manager: Working in open source / Being the maintainer of a popular library / Blogging about software: All this things won't give you necessarily a competitive edge but can work against you. It's counterintuitive but sometimes teams are looking for a more low-profile hire.
In my experience, maintaining a very popular software library, supporting open source, and blogging have absolutely all contributed to my success, and, additionally, as someone who is now a founder seeking like-minded, highly skilled engineers, those are key signals for an attractive hire.
I can understand though, perhaps in a work environment where management is unlikely to be able to retain high skilled talent, you may want 'low-profile' workers that aren't going to have as many competitors chasing after them...
How so? Care to elaborate? I get that bloggers/educators can sometimes be not the best fit for IC roles but doing open source seems like a huge advantage.
Some companies want subservient, homogenous employees that come in, do work, and can be let go if they do not perform. That's a simple equation.
If you get in somebody who is a star, however minor, that changes the equation, changes the dynamic. Now that person can have more confidence, can have more sway in the decision making. If the company wants to let them go, then they might post a message to their followers, riling them up, creating bad PR for the company. It's no longer a simple equation.
So it all comes down to the insecurities of the company.
> Care to elaborate?
When parent poster says things like “low profile” it should be interpreted as cheap and doesn’t know their worth. Assume all hiring managers want the least qualified and cheapest possible employee that can still get the job done.
Not always true, but true enough to be useful and more true than hiring managers admit to themselves. I’ve been a senior involved with hiring for years because while I full don’t want to manage, I also never trust my manager to hire well. They have multiple mutually exclusive narratives they tell themselves about how they hire/manage. Not all of them are true, and sometimes not any are.
> huge advantage.
It dependents on the size of the organization a lot. However in general it's likely that the new hire is the most competent of them all, which would be an immediate risk for some of the managers (e.g being displaced)
It might be similar to how employers dislike hiring entrepreneurs. People who already have a career bigger than their job
This is kinda fair, though. People who have run their own business make for really, really, awkward employees. It takes a really skilled manager to deal with them properly
IF you have a side gig it is easy to think you won’t be 100% invested in company success. If you monetize you most likely will jump ship.
There are other risks like burn out as you may read a lot of OSS contributors have — so when someone is hit by burn out it will be across the board not that they somehow will perform at their peak at job while burned out by coding on side.
I was part of the selection committee for a position once, where we selected the more junior engineer.
The probably most simple explaination would be that for some roles you like to have someone that can be easier "shaped" into a certain role. Someone who is already successful may bring their own system of doing things. This is great if it is a good fit, but can produce frictions if it isn't.
The next thing is that if you apply to a mediocre position with overly amazing credentials, it can raise suspicions. Something must be wrong with you, maybe you got amazing credentials, but you are complicated to work with. Maybe you're looking for the mediocre job just because you think it will be a walk in the park, etc. There are legit reasons for this (e.g. "my partner moved to $TOWN for her career and I am looking for something to do here, and you seem like the best fit. I know I am technically overqualified, but I wanted to go back to coding for years now and this offers me a geeat chance to give it a go").
Of all the senior canidates we have rejected the most common issue was that they didn't offer a convincing explanation to why they chose that specific position. The worst one was talking about how it would be a relaxing position for them.
>Maybe you're looking for the mediocre job just because you think it will be a walk in the park
>The worst one was talking about how it would be a relaxing position for them
What's wrong with that? Can't you compensate being lazy with being efficient?
Yes, sure, in theory. But the position we were filling was one with very little supervision and oversight, for room reasons. So basically one person in a room in a different building who has to maintain a bunch of stuff in addition to build up a organizational structure from scratch.
Filling it with someone who you might have to check after not for seemed like a risky bet. Call it a gut feeling. I worked together with a guy like that, which lead to me having to save the day every other week because he forgot to organize for an event he knew about months in advance.
I agree with the other comments on this thread, but I have a question of my own, why not work as consultant at that point and not as team member?
That's often works and is a good idea, in my personal experience. It would be so much better, however, if we had a functional and affordable health care system for independent consultants. Consultants working from outside the US may actually have an advantage in this regard, depending on where they are exactly.
No equity.
This is cope and propaganda to discourage people from developing their own brand. Better for the corporation if the workers have no support structure or reputation that might lead them to quit
"Developing your own brand" is not a scalable solution. There's only ever going to be a few thousand developers who are well enough known to be called a brand.
> This is cope and propaganda to discourage people from developing their own brand.
This is such a US-centric cliche that it even reads as a parody. No, the man isn't keeping you down.
it doesn't mean one should not do it - but it's not an immediate benefit
> Better for the corporation if the workers have no support structure or reputation that might lead them to quit
That's exactly right.
> This is cope and propaganda to discourage people from developing their own brand.
Not really "cope and propaganda" when it's true, is it?
Weak managers and teams dont want to hire the person who actually delivers something that works.
The new person could show how unproductive they are.
I think this blog post was the best way to get into Anthropic, and it was well-deserved. That's the reality of hiring in tech: there are many non-technical people judging whether technical people are competent or not. Escaping that matrix through things like blog posts, cold emails, and Twitter threads can be great ways to break in and get noticed by these companies.
HR _hates_ hiring anyone, they just want H1-Bs.
> Unfortunately they thanked me for my application but said the team doesn't have the capacity to review additional applications.
It seems like they didn't even look at his application.
Even if someone from HR who screens for first rounds read it, they probably wouldn't understand why this might be an interesting candidate.
yeah, as this is so often the case, many times good, relevant applications are missed. I hope that this Hacker News post will get to one of the key people at Anthropic and they change their minds.
At my firms I saw this happen often. HR would review, or a junior engineer and pass on very good candidates. It wasn't until I set up a review system with A-class engineers that we started to catch the best people. A-class engineers recognize themselves far better than anyone else. But they prefer to build than review resumes.
I ended up building my own head hunting firm specifically to address the whole pipeline. That helped somewhat but head hunting is its own very odd space. Full of inefficiencies and bias.
With any AI company, there are always limits you hit. Energy, compute, optimizations, inference, team resources, money, and all the flows to make it a company. HR is usually the one that gets the fewest resources.
I think the issue is that some applications are not even reviewed. HRs can also learn the expertise of identifying strong candidates if they build up the experience and frequently talk with engineers about pros and cons of resumes.
> I hope that this Hacker News post will get to one of the key people at Anthropic and they change their minds.
Honest question: what leads you to believe they should change their mind?
first of all because the key point is they didn't even look at his application, and by any objective criteria he should have easily got through a "worth a human looking at it" screen. but also, hiring the developer of an open source library that you want to use internally and paying them to both integrate the library and work further on it is an excellent way to have a sustainable open source ecosystem, which both anthropic and the developer will benefit from.
Maybe Anthropic uses Workday for its HR, which is being sued for possible systematic discrimination by AI. (See links below)
https://www.insidetechlaw.com/blog/2025/06/workday-ai-lawsui...
https://www.forbes.com/sites/janicegassam/2025/06/23/what-th...
I hate workday.
Nor their tech. For shame! Where's the enablement?! I was promised productivity.
Yes, it sounds like a non-story to be honest.
It's a personal blog
Sure, but on HN it is a story.
The "it rejected me" in the headline should have been "it didn't notice me".
I’d be curious to see the outcome of changing the license to a Fair Source License or explicitly “You are not allowed to use this software if you are Anthropic, otherwise MIT”. They could still use the current version, but for any in the future they’d be forced to fork it or be prepared to face yet another legal battle (I can imagine some lawyers already salivating at the thought).
It’s also curious the author is looking inside the app for proof their software is being used. If it’s MIT, mustn’t the license be included and available somewhere easier to verify?
I would guess like him that no human engineer ever read his application. The less they would have done in that case would be to at least thank him for his work, even if they don't plan to hire him for some reason.
Automated systems, AI screening, and incompetent HR people are the bane of modern recruiting practices.
Perhaps, but just like vibe-coding being good-enough for some purposes they think vibe-hiring will get good-enough candidates?
I guess at least they're dogfooding it?
On the electron part, it's common to (ironically) not support Linux. There are pretty annoying bugs with windows management (window will stay stuck in the background), build process are always OS specific, etc. So often not worth the maintainance.
For generic consumer products, sure, but for dev & technical power user tools the audience is big enough that these arguments doesn't hold water. Stack Overflow's latest survey shows nearly 30% of professional devs using Ubuntu specifically (https://survey.stackoverflow.co/2025/technology#1-computer-o...) and my own metrics (building a cross-platform dev desktop app with a global dev/technical user base) show pretty similar numbers: 65% windows, 20% Mac, 15% Linux. I would expect there's a significant (comfortably above 10%) Linux user base within the claude computer use audience.
The practical reality of distributing is mildly complicated, but there's now lots of good cross-distro options, and not having to deal with code signing everything makes some parts much easier than Mac & Windows. Ignoring that many users is fair enough for a startup or first MVP, but quite surprising for a company at Anthropic's level.
In the very least contract the developer for a little bit? Aren’t these Ai companies swimming in capital? Something almost dystopian about this
Uh, a company not paying money for something they can legally use for free? There are so many MIT-licensed software libraries that everyone is using in a critical place, for profit, with zero money flowing back into the ecosystem that created them. It should surprise nobody, it has been like this for over a decade now.
I think that everyone should read this blog post
"Overall I am overjoyed enigo is used in Claude Desktop and I tell everyone who listens to me about it :P. It's so cool to think that I metaphorically created the arms and legs for Claude AI, but I can't help but wonder if the rejection letter was written by a human or Claude AI. Did the very AI I helped equip with new capabilities just reject my application? On the bright side, I should now be safe from Roko's Basilisk. "
I also felt like this way that did they just AI in their interviewing process?
And I have a special love towards open source.
And I personally might be happy too that a company is using my work ,but in the name of the holy licenses, Companies are just exploiting the free nature of this and the fact that it seems like not even a human looked at the person for such job, who created a library that they are using it for free...
I was thinking of creating some code in MIT license, but I am going to create a code of AGPL except if you sponsor me on github or a special one time license which can grant you MIT.
People might say that I am not fostering the open source community, but I am not giving corporations free labour so that they can be billionaires.
I once saw someone write a software with the exact same idea (AGPL + gh sponsor me to get MIT) and the people in HN were pitchforking him, that's the harsh reality of the world. People want absolutely free labour.
I think open source needs to ask, Have we become the modern peasants in the name of our altruism?
I think we need to compare our industry with other industries. No other industry relies on free labour from random people, which comes with no support or promises.
I once told some non-techie folk about some code I wrote. It did something super simple and wasn't that big. They were all asking why I didn't sell it and thought it was crazy I would give it away for free with the BSD license. It was 900 lines of code... For us, that's nothing but for an average person they just think "I built it, I'll sell it"
The problem is that 900 lines of code is also nothing for your potential customers as well. Non-programmers have a very poor ability to judge how difficult something is and how worth paying for it is. 900 lines is probably less effort for most organizations than it is to evaluate paying for the functionality.
Out on the super, super far end of the distribution you may have things like paying for what is essentially 900-ish lines of extremely, extremely carefully vetted code for things like encryption, but that is very, very exceptional.
I've got a few open source projects on my GitHub that are in the 900 line range, and I know they're used in a few "interesting" places but I'm not crying about it because the simple truth is the commercial value of that code is simply $0. If I tried to sell it to the people using it, they would perfectly rationally just say no. I am abundantly compensated for it by all the other open source software I get to use.
Ya I also believe it this way, Mostly I like to build stuff for my own problems or something that I just find fascinating.
I am still in high school, so I was doing some question sheet that our teachers provided and there was an answer key but it had answers of everything. Now I don't know how other people approached it but I am really impatient and so I just open up answer key side by side but it reveals every answer.
So I firstly created an AI to ocr to card generator but it was an hit or miss and so I discussed it with my friend and he said that he used to use paint and somehow in his convuluted manner basically have a slider which would reveal answer...
I found it incredible and so I just created a single index.html that can do it. (Although vibe coded), Now I can't even think of monetizing such ideas when I realize that there are creators of some really incredible stuff and long convulated stuff and even they aren't sponsored so I have always felt that the scripts that I write or projects around such ~.5-3k loc. I just don't think of monetization.
I just don't know.. I like hacking stuff, I just feel more comfortable rebuilding stuff even if its mediocre if I feel like I can change it to suit my purpose better
I think that the only other industry that is gives as much completely free stuff might be research/science related, but maybe its due to the fact that computer are computer science too and thus related to academics.
I really just love tinkering with software and just the aspect of freedom that it can provide , but sadly, I find it just hard to really make money without being a job and such stories on which we are discussing, just makes me feel like I am kinda right.
On one hand we have 100 million payouts to researchers and on the other we have this, such disparity is kinda sad I suppose.
Other comparison would be that we are equipped with tools (software) other people build completely for free. And we can improve those tools and propose them to be implemented (PR). Or just continue using our custom-modified tool (Forking).
And we often get the luxury to ask questions and receive answers (Issues) directly from the manufacturer (author, contributor).
And we need not much investment to set up our own factory... thus "materials" can be free and then we give away our product.
Come Anthropic, give this guy a fair shot. At least interview him in person or something.
A friend of mine is maintainer of an open source service used (at least, at one time) by all of the major social media platforms as a load-bearing piece of their infrastructure (intentionally keeping it vague). My friend was invited to interview at one of the biggest and was rejected after having a bad whiteboard session. Of course they immediately replaced my friend's service (ha!)
Since licensing has come up a few times in this thread: I've been hearing recently that the Mozilla Public licence (or the EU Public licence) is a good middle ground between the "viral" GPL and the "do whatever" MIT - as per my understanding, if your code is MPL or EUPL, it can still be incorporated as part of software that has a different licence, but any direct changes to the MPL/EUPL licensed code itself has to be shared openly.
Does anyone here have experience with them, or knowledge about whether that description is more or less correct?
Anthropic throw this guy a consultancy on demand job, or at least a bit of money. He's made your business rich!
Good findings, the rest not surprising tho.. online recruitment doesn't work at all these days. most likely your app wasn't read by anybody meaningful and did not trigger right flags in the HR system to even be spotted by clueless ladies working there.
This post can give you some visibility unless somebody sees it as frustration/negativity then they won't bother either.
aside of the core topic, best way to get a job these days is unfortunately either some elite job boards that work and both sides know why... or personal relations.
All the automatic HR/recruitment platforms is illness and i'm sure that's what victimized your genuine application there.
I assume your choice to describe average HR reps as “clueless ladies” isn’t meant to suggest that you respect e.g. women software engineers on your team any less. But if the gender of the clueless HR employees isn’t relevant, why mention it? Maybe worth reflecting on whether calling them clueless ladies rhetorically emphasizes their cluelessness
The OP should have left off "ladies" as that was unnecessary -- clueless is a sufficient descriptor for many people in many roles (whether genuine behavior, strategic fiefdoming, or learned helplessness).
Here is the US BLS breakout of demographics by occupation category: https://www.bls.gov/cps/cpsaat11.htm
To save everyone the click: HR managers: 76% female, HR workers: 75% female, HR assistants: 84% female.
The distribution in other categories is fascinating, and HN doesn't format tables well.
Though "saving a click" typically refers to spammy clickbait news articles that bury the lede, which a statistical table directly relevant to the conversation does not qualify as.
I found it useful at least.
That doesn't make the remaining 24% more clueful. Identifying a department by the gender of its workers seems pretty suspect.
I read the original comment as implying that the average HR person reviewing your resume will be a clueless woman. Not as implying that only women work there or that people of other genders working there are more clueful.
The comment is open to interpretation, and you are free to interpret it in a less charitable way. The ambiguity is absolutely something we can and should criticize the comment for
I never implied any of that nonsense that I’ve perhaps triggered nor I want to be responsible for other people’s interpretations outside core meaning
Neat fact: statistical independence means that two factors are orthogonal.
My prior, expressed in my earlier comment, is that cluelessness and gender are orthogonal.
Reality denial and picking on people who state simple truth is evil.
~80% of software developers are male. [1]
I’m not sure the percentage of companies that use software for highlighting candidates, but Anthropic almost certainly does and this [2] source says 75+% do.
So since men wrote the software that didn’t highlight the candidate, is it the clueless men that caused this?
[1] https://www.zippia.com/software-developer-jobs/demographics/
[2] https://www.theverge.com/2021/9/6/22659225/automated-hiring-...
Yes - people in HR departments are often female and often clueless, but I don't see the parent denying this. The wording of OP connected both though, which is sexist and can be considered "evil".
Funny enough, I see this whole framing as sexist itself.
Nobody would have bat an eye if he said "clueless guys" or "clueless gents", and given the prevalence of women in HR, that wording would actually have more chances of having a sexist background to it.
“guys” is gendered but is very often used to mean a general group.
>given the prevalence of women in HR, that wording would actually have more chances of having a sexist background to it.
The reason there are more women than men in HR is clearly because the men they do hire are too clueless and get fired faster. Ever have an HR department with all men? Most dysfunctional department I’ve ever interacted with! “Clueless HR men” is just redundant. The ~25% that exist are DEI hires. So it wouldn’t be sexism, it would be reality.
You‘re right, but that just reflects the structural sexism in our society while the wording by the op was intentional (I suppose. If not, I might as well be more sexist then he is).
Nopes, nothing to the ladies in general nor any other gender in general. Just a shortcut to my own negative experience with HR by example. English is my foreign language and in my country we are not that allergic to terminology. But clueless processors stay as valid… regardless of particular denomination.
After years hearing justly about bad things perpetrated by males as a class, without any concerns about generalizations, I think we are mature enough to also call for responsibility in the other side of the aisle.
Having a free pass for doing evil stuff is what gave man their bad rep, should we now for equity give women a pass to become the new slave lords?
>spotted by clueless ladies working there
Why the need for the sexist addition of ladies? People of all sexes and genders in HR can be clueless.
> clueless ladies
Come on...
Never mind the application rejection, you'd have hoped that in choosing to base Claude Computer Use on his library that they might have at least reached out to the developer to say thanks, preferably with some token show of appreciation like a few Anthropic shares.
Doesn't look like he was rejected, rather not considered at all.
This is why wealth accumulation is so terrible. People with lots of money drive science and technology. They accumulate more wealth from science and technology whilst demonstrating a complete lack of understanding of the thing that's making them the money.
Most executives and investors just throw shit at the wall to see what sticks, imo. Then move on to the next place. That's why golden handshakes exist.
I'm with Luke Smith [1] when it comes to non-copyleft licenses like MIT.
Andrew Tanenbaum of the MINIX fame was similarly surprised to find that Intel had quietly included the OS he wrote in Intel chips, making it perhaps the most widely used OS in the world. He seemed disappointed no one ever reached out to him to tell him about it [2]
[1]: https://lukesmith.xyz/articles/why-i-use-the-gpl-and-not-cuc...
[2]: https://www.cs.vu.nl/~ast/intel/
To intentionally mix metaphors, I can't believe I'm about to knowingly kick this well worn hornets nest.
It seems obvious that if Tanenbaum, or any open source project used a GP license in lieu of a permissive legally familiar license like MIT or BSD, the likelihood of the project being used in a commercial product would reduce to nearly zero. Intel would have used a different OS for their management engine.
I'm glad the GPL exists and believe the world is a better place because of it, but it feels like more and more it's salad days are in the past and the world has moved on.
The ops experience reminds me of the story of the maintainer of homebrew that despite widely being used at google was not able to be hired for a job there. It's disappointing and feels unjust, and I wish it was different.
Unless there was no MIT version.
We should stop coding for free for billionaire organizations. The romantic era of Open Source is over.
The only projects with a permissive license, I am comfortable sending PRs nowadays are the kind of projects that will hardly enable a big monopolist to extract more rent from society while being covertly funded by the debasing of currency promoted by the FED via Cantillon Effect.
People should have realized this long ago. They're working for free for mega corporations. I refuse to contribute to this.
Over a decade ago, it was my dream job to get a job at one specific FAANG company that is widely known to use a project I've contributed to.
I'm a developer with a project they use, so, I thought, for sure someone would review my resume after applying on their website. Nope.
After being ignored for a while, even having to get a Master's degree because no offers after a Bachelor's, I finally emailed a Director, who was previously a fellow committer at the project. People under him were not hiring at the time, but a recruiter from a different group has contacted me shortly, and I've had a 2-day flyout onsite arranged for two different positions, and had offers to join either one.
I don't mean to downplay the author's skills, but I don't see how creating an input simulation library fast-tracks someone for consideration in an AI-related engineering role.
Didn't he say it was for the team integrating his input library into claude desktop? Seems pretty relevant experience.
It mentioned an "open position in the team implementing the secret, unreleased feature of Claude Desktop," which doesn't specify whether the "secret" feature is AI-related or UI-related. My guess leans towards the former.
To give the full quote, it says:
> I found out that Anthropic had an open position in the team implementing the secret, unreleased feature of Claude Desktop using enigo.
where enigo is his input library. It's quite interesting that you chose to end your quote a few words before the end of the sentence.
You got me, adding “using enigo” makes all the difference — I guess position is exclusively for working with this one library and that’s the position they got overloaded with applications on and couldn’t process one sent by OP.
It clearly does make a difference as to why he thought his experience was relevant to the job (i.e. what we were discussing before), and I think you agree with that hence your somewhat "selective" omission when you posted the quote.
I wasn't making a statement about whether his experience was relevant to the job. I don't know the author and don't automatically doubt his knowledge. I was simply sharing the opinion that being the author of that UI library alone does not fast-track someone for the "Software Engineer working on Claude Code" position at Anthropic.
The role I applied to was not really AI related
I wasn't aware of that; it wasn't clearly specified. It only mentioned a "secret" feature, but I assumed it was AI-related rather than UI-related. Additionally, Anthropic's Claude Code position on their website states that they expect their developers to work across the stack, including both front-end and back-end.
Is there an license that requires payment for usage for corporations above a certain size?
Wait for the Meta offer, it could be a few millions.
Ha-ha, they also trained their LLMs on your code and maybe will even train on that blog post :)
Anthropic probably gets tens of thousands of applications. They seem to have filled their queue before even reviewing this particular candidate. Unfortunate but just reality.
Always always always try to get into direct contact with the actual hiring manager. Blog author had a friend of a friend let them know a relevant role was open. The correct move is NOT to blindly apply. It’s to ask for an intro to the engineering manager responsible for the role.
I'm very much starting to re-consider open source. It mainly seems to be a way for already incredibly wealthy companies to get things for free, or to strategically release things to crush their competitors.
Maybe we ought to go back to paying for proprietary software. A lot of people used to make money that way, ie by selling their own desktop app.
Anthropic also rejected me for a job... that I never even applied for...
This sort of silliness is what you get when you run crucial business processes using AI instead of humans.
reminds be of the time creator of Homebrew was rejected by Google in coding rounds. but this is even worse, they would not even interview this guy. shame on Anthropic... (or is it Misanthropic?)
In my opinion, lots of open source was developed as a sort of portfolio to get hired. From 2019 onward, my impression is that your open source projects (regardless of how much they are used) matters less and less and it’s about HR mysteriously picking you up in their process than anything else. I think, now, your open source portfolio matters exactly nothing in the decision to get hired.
I remember back in 2014-2019, it was hard and competitive to contribute to open source projects as they were tightly guarded. There are many projects that I use now in package.json that are looking for a maintainer. A complete 180 flip.
My guess is that real free open source will disappear in a few years and what will remain are open source projects monetized by some business somehow.
It’s a sad reality but that’s what the current people at the top have decided today.
Hire OP, anthropic
I hope Meta already contacted him.
AI companies try not to be evil challenge (impossible)
Give this man a job, Anthropic!
Why is no one talking about how they had an indirect contact at Anthropic but didn’t use that connection? Your chance of getting hired is way higher with a referral.
Dude they did not reject they did not even SEE you because they likely have 10k application per week.
Just ask your friend for an intro.
He already works for them without pay in a way. Why would they hire him?
To close the source.
To drive the development.
To prioritize some bug fixes.
bro probably didn't even go to Stanford or another "top tier CS program" (yes people literally post job ads with that requirement) smh
Can a license be modified? What happens in that case? Let’s say I want a Ferrari.
Everyone is suggesting that AI rejected this candidate but that brings up two points:
1) Is the hiring AI so incompetent that it did not realize it had a "S-tier pull" in the process and should have immediately prioritized the find?
2) Was the candidate's submission so bad that a reviewing AI couldn't even tell the massive relevance he had to their work?
I suppose, alternatively, Anthropic could just not really care about Claude Desktop enough to hire a specialist for one part of the stack. Perhaps they're looking for much more "full stack AI" who can do a lot. They have 350-400 total engineers, is that enough to hire a specialist for Claude Desktop?
I guess my question is: Did the AI fail, did the candidate fail, or did the AI work well and we just don't know the criteria it was succeeding in using.
The next version should have a feature where the first thing it types into any text box is "Anthropic, I wrote this library! Please look at my CV!" and then deletes it.
> Unfortunately they thanked me for my application but said the team doesn't have the capacity to review additional applications.
Okay, they were just busy doing work and didn't have any time to look at applications so they shuttered the JD and auto-rejected anyone in the pipeline. Seems reasonable
Another reminder that if you write software under an MIT licence or similar then you're just working for companies like Anthropic for free.
Use GPL or AGPL. It's the best thing we have.
Remember that companies like Microsoft spend billions on PR and their goal is to make you think what's good for them is good for you. This is rarely the case.
I feel this is really a blog post about two indirect topics; one that has not been addressed for many years now, and another that is not new, but has been getting seemingly ever more acute recently.
The first is the issue of permissive licenses like the MIT license, that seems likely far beyond an appropriate license structure for today’s world and environment, I would even argue inappropriate since the .com bubble. Software and creating has changed a lot since the 1980s to such a degree that I don’t think even the originator and early supporters of permissive licenses would be supportive of…peoples work being used in critical ways to build two and three digit billion dollar corporations without any kind of reward or compensation. It’s an odd kind of peak dystopian hybrid of communism and capitalism, sacrifice of the self for the benefit of the very few.
I think it is at least time to discuss archiving things like the permissive MIT license (assuming it even makes any kind of difference at this stage) that are from not only a different developmental stage, different environment, but even a totally different country, society, nation, and world even.
The second theme of this blog post seems to be the absolute seizure of the… what should we call it?…resource allocation of people? I cannot recall right now, but I feel like this is the second blog post themed around someone core to some function of some big tech company being rejected by said tech company; and that’s in the backdrop of the cacophony of people dealing with all kinds of dystopian insanity in the employment/job market from fake/scam jobs, AI interviews, etc. The system seems to be totally breaking down to some degree, even if it is still limping along, as is evident by the massively downward revised job creation numbers over several quarters now. How do you “revise” jobs numbers from 139,000 to 19,000? Ignoring any political partisanship, “revising” an estimate downward by 86% is not just an “whoopsie”, it’s evidence that thins are broken, regardless of why or even how. They’re clearly broken.
I have approaching 0% confidence with anything related to Congress actually doing its job since it has effectively abdicated its cute role that provides it legitimacy, but discussing both of these topics in public can have a chance at forcing the muppets in Congress to address the issues, even if only for narcissistic and selfish reasons of being (re)elected to enrich themselves after they’ve gone back on their lies to get elected. And no, neither team is the better team; it’s all a con-job.
Reminds me of the guy who created Homebrew being rejected by Google for failing some silly Leetcode puzzle.
Which is not something that happened, even according to Max Howell himself: https://www.quora.com/Whats-the-logic-behind-Google-rejectin...
> I feel bad about my tweet, I don’t feel it was fair, and it fed the current era of outragism-driven-reading that is the modern Internet, and thus went viral, and for that I am truly sorry.
Building popular software doesn't mean you're a good programmer, especially since at that point Google was looking heavily at CS concepts and he admittedly wasn't good at that.
It's also possible he would have been hired if he applied for L-1. A lot of people get an ego check applying to Google where they're a senior staff engineer or a CTO at a small company and get an L5 offer.
True but surprisingly grinding leetcode puzzles also doesn't mean you're a good programmer. In fact, in my decades-long now programming career, I've had to take many more decisions of the homebrew kind (e.g. how the thing is going to work, how the API is going to look like, will the users love or hate that feature, etc.) than the leetcode kind. And now I am thinking the former is even more important. If you get the leetcode part wrong, worst thing your code would be slow. Not a good thing but also not a complete disaster - you can come back and optimize later. If you screw up the design and interface part, nobody would be using it - or worse, they'd be using it in ways it wasn't supposed to be used - and then it doesn't matter how fast it is.
> True but surprisingly grinding leetcode puzzles also doesn't mean you're a good programmer.
I don't think anybody with a modicum of experience finds this surprising at all.
Are people who don't work for Google supposed to understand what these levels mean?
it kind of happened, he went through seven interviews. from the same post:
> But ultimately, should Google have hired me? Yes, absolutely yes. I am often a dick, I am often difficult, I often don’t know computer science, but. BUT. I make really good things, maybe they aren't perfect, but people really like them. Surely, surely Google could have used that.
> I am often a dick
It make things really nice and easy when someone tells me enough about themselves in just a few words to make me not want to work with them.
Maybe that's why he didn't get hired? His dickishness came through in the interviews?
I don't think most people who behave in this manner have enough self-reflection to write something like that. They would rather write that they are opinionated, principled or decisive or some other bs.
> I make really good things, maybe they aren't perfect, but people really like them. Surely, surely Google could have used that.
This line could apply to millions of people around the globe.
The #1 way to not get hired is to be a dick. The brilliant asshole is the most toxic person you can have on your team. Don’t be an asshole.
Exactly this. No amount of cred, smarts, and genius that ends with "and I'm a bit of dick" will save you from my automatic red-line veto when hiring. I'm far from alone in this.
> will save you from my automatic red-line veto when hiring
You're literally a power tripping dick hiding behind "I'm not letting other dicks in" facade.
You've got it backwards friend
umm, linus torvalds, richard stallman, elon musk to name a few
None of those people ever applied for a job at Big Corp where one of the most important aspects is to be able to work well with other people and tactfully navigate the social structure of the company.
being a dick has not hindered there prospects. id argue dickheads are more prevalent in tech due to the prevalence of autism
Nobody is saying it hinders their prospects in general. They're just saying that "being a dick" is incompatible with a specific kind of job: one that requires collaborative and cooperative work with other people and navigating the social hierarchy of a company.
In my experience, it's the nice people that get fired and the assholes that get promoted. It's not exactly a secret that silicon valley is full of arrogant assholes.
Founders of a lot of companies also tend to be dicks. But seems to do alright. Seems to be a double standard there
I mean, he's also the same guy who apparently thought "Unix ideas that have worked for literally decades, nah fuck that. I know better".
It took over a decade before the project made some improvement on how the default install path is handled.
To my knowledge it still has absolutely atrocious dependency resolution relative to things like DPKG.
Not hiring this guy is honestly like a fancy restaurant not hiring the guy who comes up with the new McDonalds obesity burger special menu. What he created is popular, it's not good.
Google is not a fancy restaurant. Five-guys private consultancy is a fancy restaurant. Google is the McDonalds of all McDonaldses, it makes software that is used by everybody, whether they want it or not, and you can't turn a corner without hitting something they control.
Or the FastAPI creator not having enough years of experience with FastAPI according to a job posting.
That is the most absolutely absurd wild damn story that I have written.
Care to provide links...
How can interviewers be such stupid, the fastapi creator had the MOST experience with it, he created it..
https://x.com/tiangolo/status/1281946592459853830
Edit: note that I wrote "according to a job posting". It's not the same as the situation in the parent comment.
https://x.com/tiangolo/status/1281946592459853830?lang=en
Homebrew is “just” a package manager, not the core of part of Google. They could rip it out overnight and won’t even notice it. And Google gave him a fair shot.
This guy got rejected by some automated system without even interview.
> Homebrew is “just” a package manager, not the core of part of Google.
Yes, this. Sometimes I wonder if those coming up with the Homebrew example have any experience whatsoever with software development. I mean, sure the project is popular and surely doesn't hurt on a resume. But does it showcase any level of technical expertise or mastery? No, I'm afraid not. I would bet that the majority of software engineers would be able to put together an equivalent system in a week or so. Think about it, and pay attention to what are the system's usecases. It's hardly rocket science.
>does it showcase any level of technical expertise or mastery? No, I'm afraid not.
But it does show that he can develop and ship a popular product, something outside the capability of so many “great engineers”. Good luck generating any revenue on the backs of smart engineers that have no stomach for understanding the nuance of development over and above writing and checking in code.
It says nothing about fitting office culture of Google, or generating successful projects, for that matter. If he’s so good, how come there’s nothing else besides brew?
So, was he rejected by an automated system or did he go thru seven interviews, as other commenters say?
You might have missed the subtle wording;
_That_ guy (Howell) got several rounds of interviews, _this_ guy (OP) got rejected by an automated system.
Not the "subtle" but simply, inaccurate wording then
Not inaccurate; it's perfectly understandable to a native English speaker, the nuance is subtle, "this guy", "him/that guy", but it is clear and commonly used language.
"I just spoke to a guy about X, his opinion was different to the guy I spoke to about it last week. This guy said Y, but that guy insisted it was Z."
GP is talking about two different people.
"Him" is the creator of Homebrew. Seven interviews at Google.
"This guy" is the creator of enigo (discussed in this thread). Automatic rejection by Anthropic.
(Edit: upon page reload i saw the quicker answer.)
Does 7 interviews not seem excessive? Got my current job with 1.
Silicon Valley lives in lalaland.
7 is a bit excessive but it might be including team matching which is a bit more informal and less obnoxious than the leetcode style interviews
I was thinking about this the other day. I think it might just be a thing of Google looking for a different thing than what made his open source project famous.
Without no knowledge of the details further than mxcl's tweet; probably any performance issues even on simple code, get infinitely multiplied when running at Google's scale, slogging the thing, on Google's dime. From what I've seen of him, mxcl is good at designing a really approachable product, and on running an open source project. But homebrew is really slow, even on the latests Macs, even for basic cases.
To me it seems then that he'd be more fit for a product owner/manager position than an engineering one, and that could be the root of his not-hiring.
This guy is so full of himself, no wonder he didnt get hired. Just read the homebrew github issues / forum and you will see what i mean…
This story feels after all the years still awkward. Many people at Google don't have anything that impressive on their resume like being the creator of homebrew. Commentary like "Google looks only for computer scientists, so you need to have studied CS" is so out of touch that i sometimes questions if these people ever worked in a big corporation. There are thousands of different roles, many multiple times suited for that guy. I suspect the people who vetoed didn't like that guy for some other reasons.
Iirc the homebrew guy did at least get an interview
Dev: I wrote a part of your software that you are bragging about. Can I have a job?
Antrhropic: tl;dr kthxbye
It's inherently risky to blog about your professional relationships under your own name and this is a weirdly small hill to die on.
He was very courteous, no deaths on a hill to be found
Publishing anything about it, regardless of content, is already a hill.
I like that people blog about these experiences and enjoy the insights, but I think it's never good for the authors..
Everyone should. The only way to balance corporate power is collective action by individuals, and sharing information is a requirement for that. Corporations can't get away with quite as much brazen sociopathy if their actions are transparent and reported without - or a different - spin.
Just change the license. The company probably won't notice and keep pulling the new changes. Now you have a legal case.
> Just change the license. The company probably won't notice and keep pulling the new changes. Now you have a legal case.
I think the world already grew tired of rug pull tactics. If you want your reputation to go down the crapper with a lame attempt to shake down an end user, go right ahead.
I wonder if he writes cover letters to every company that uses his library.
Why not? This is second easiest way to get a great paying job, second only to nepotism.
He can probably use Claude to write is for him.