Ask HN: Why do HN programmers seem happy about losing their jobs to AI?
First "vibe Engineering", now "context Engineering". Support for these kinds of posts is overwhelming.
In both cases, it seems programmers are feeding generative AI models to the extent that their involvement is minimal or almost redundant, making what they do seem trivial and devaluing their own work. You seem to be writing off all HN devs as a homogenous mass and claiming that (a) they think that they will lose their jobs and (b) that that is welcome. Plenty will disagree with one or both of those. Unless you are here to start a fight, consider a more nuanced question, for a relatively broad audience. (FWIW I disagree with you on both.) > all HN devs as a homogenous mass Many of the highly upvoted posts of the last ~month have been about "vibe engineering" and "context engineering". The comments on these posts suggest an air of enthusiasm one doesn't see on posts about programming. > consider a more nuanced question As a long-time poster here I simply do not agree. > I disagree with you on both Your rationale? By using AI to automate things you once did manually, you are: - Helping a model do something better you were once able to do (and get someone to pay you to do). Such a time may arise when your input is no longer required. - Propagating the idea that such things no longer require the skill they once did. - Propagating the idea that it doesn't take you as long to do something you once did, begging the question why employers would pay you as much as they once did. - Failing to continue to learn as you once did (that numerous studies have since confirmed; people who use generative AI don't exhibit the same critical thinking as those who don't). I use high-level tools to automate things all the time, such as scripts to avoid error-prone manual tasks, and high-level languages rather than writing low-level assembly code byte-by-byte ... usually, when they the right things to do. However, given that my life now is mostly working on trying to reduce global warming, and I have fairly high standards for my output I think, I don't by default want to use tools which (a) cause large amounts of needless carbon emissions one way or another and (b) inject random lies/nonsense into their output which is hard to track down and may in fact slow me down overall. I have been listening carefully to talented people I trust who are getting coding assistant results that they are happy with (even given their TLC to fix the errors and lies), and will carefully evaluate for myself when the hype has diminished a little, as I have with other new tech, and when mostly the good stuff is still standing. But at the moment I am learning a whole new complex toolset in a different area, in building physics to reduce emissions, and cannot really do both at once properly even if I was at my 18yo sharpest and wasn't trying to have a good social life too. As I boringly repeat, I am not anti AI. I have a near 40yo AI degree and have hundreds of thousands of product units on the market featuring what is at least clever data science. I do object to overhyped LLMs from vastly rich peeps wanting to hoard even more resources while being oblivious to those who have provided the intellectial property that they have perloined, while flattening the creators' servers with badly behaved bots too! I have already nominally retired (though am working on a self-funded PhD) so the "I will lose my job" cannot apply and is a counter example to one of your universal claims. Getting tokens on screen has not been the bottleneck for me in a long time. Knowing what to build and how to evolve an application from the current state into a future state, while keeping it short term, medium term, and long term coherent, is the bottleneck. Until the AI can take the prompt: "Improve the product so that it will generate more sales, satisfy the government, and not piss off existing users", it is not going to replace me. For every task less than that, it does marginally improve my workflow. For any programmers out there who are so untalented that they barely exist as more than a human token generator, they are indeed at risk. I've been a software developer, engineer, architect - whatever you want to call it - for 40 years. Here's one simple truth that’s held up my entire career: Stay current with technology, or your career will stall (or worse). That’s it. But staying current gets harder with age: - Family responsibilities reduce time for learning. - Workload increases, leaving less time to explore. - Developer community isolation makes it harder to stay connected. - Learning new tech gets slower with age. - You get jaded, most “new” tech is just old ideas repackaged with a lot of hype. That said, you should try vibe coding. For someone like me, experienced but not always hands-on, it’s incredible how fast and well you can produce code. But here’s the catch: the more experience you have, the better your results. That’s why junior devs are falling behind. Veterans intuitively understand complexity, architecture, and their role in the process. They can “vibe” good solutions with little training - just experience. And that’s the problem. You need experience to thrive in this new paradigm. But with students leaving CS programs and companies freezing junior hiring, we’re killing the pipeline. That’s dangerous. LLMs aren’t general AI, they’re tools. And tools are only as good as the craftspeople using them. Meanwhile, if you’re still in the game, you can’t afford to get lapped. I have been in the industry for over 50 years. The one constant has been change. Perhaps in the last decade there has been a certain degree of stability, but it won't last. What you know will be useless tomorrow. You need to be constantly learning and adapting or you are in the wrong job. There is no "losing my job to AI". What I could do before I can now do x100. Now I can write 10,000 lines of highly directed code in one day. The only people losing their jobs are the ones who were not real programmers at heart and did it for money or simply as a job and they can be fired for all I care, I will do their work too. Who is ever writing 10k lines of code in one day? I've been writing code professionally for over a decade and love it. Generative AI has rather (un)ironically taken from me much of what I used to love about it. Edit: With a username like that, I think you're a bot. Sounds like you should move on and make room for your replacement then. And usernames reveal bots? You should expand on that as a business except you now have one false positive for a 0% success rate so far. Your account is clearly a new one, created so you could so argumentatively reply to this post. Your responses don't make you sound so nearly as clever as you think. Why would anyone not be happy? There's jolly great fun to be had all around. - The Ph.Ds and other experts building LLMs are ecstatic because they're suddenly getting paid a fortune and having an entire industry paying attention to their every little utterance. - The venture capitalists suddenly have a whole new technology to make their investment bets on and there are hordes of founders eager to vie for funding. - The grifters and hustlers have new sexy buzzwords to attach to their dubious products to try to rip off the credulous. - The fad chasers and résumé driven developers have a whole new silver bullet to pursue. - The influencers, know-it-alls, pseudo-intellectuals, and net kooks have something new to pontificate on and debate each other about. - The sort of programmer getting benefit from LLMs and arguing that education is no longer necessary get to fantasize about finally being free of gatekeeping and more skilled software engineers looking down on them. I'm probably right on the edge of violating some kind of guideline so I shall stop there and say only that the chattering on HN should not be mistaken for what's going on in the real world. HN is mostly a finance forum. There is a veneer of "tech startup" discussion around it, but you shouldn't interpret the "hacker" to mean much besides hacking finance. Time has proved that notion right, a lot of the characteristic discourse of HN stems from economics more than hard science or code preferences. I'd take the average comment with a grain of salt. There are talented programmers that use this site, but most users are probably not a prolific Carmack-esque SWE. Time has proved that notion right Which timeline do you have in mind? I look at HN pretty often and it's almost entirely a forum of not-finance full of posts and comments on not-finance topics. A simpler and more plausible theory is that HN has many nerds whose views you disagree with. As someone who has always considered themselves to have a hacker mindset, I have always noticed that there is a lack of critical thinking by HN commenters which is replaced by advanced regurgitated knowledge. Akin to sounding smart but actually dumb. Sure but that's still 'I don't like these nerds'. It doesn't make HN a 'finance forum'. /g/ talks about more technology than HN does. You might not like those nerds either, but they aren't muzzled by VCs. What is /g/? Who is being muzzled? HN is a finance forum but then VC muzzle it? This is pretty hard to follow, sorry. You know what? Don't even bother, I think this discussion has said enough for itself at this point. Yes. You said it was a finance forum and that seems to have been inaccurate.