Intuitively though at 40ish, most of the things I "learned" in school but never used have all fallen out of my head. The knowledge I have is things that I was actually excited to learn: computers, enough math to make computers do what I wanted, writing so I could socialize with other nerds better, history because I like history. Basically, you couldn't stop me from learning those things, but the things I was forced to learn ended up being somewhat of a waste of time for both me and the instructor, I think. I mean, you want to challenge kids a bit because they might resist something at first and then eventually realize they like it quite a bit (math for instance), but there's also a point where the tests mostly just become kayfabe. We say these people are doing "better" because they get a higher test score, but then they get into the real world where there is no test score, and that entire academic framework is kind of useless and maybe even harmful.
One thing that tends to get lost when we talk about the education system we have is how it was invented and what it was designed for. I mention this, because if you want to educate kids better, it's useful to consider what they system was designed for so we can design a better one. Unfortunately, the history here is somewhat more cynical and bleak than you might expect. There's an excellent book called "The Underground History Of American Education: A Schoolteacher's Intimate Investigation into the Problem of Modern Schooling"
Anyway, modern schooling was essentially designed to create excellent workers -- it was never about creating excellent individuals. At the time, the robber barons who helped invent this system along with a series of pseudo religious ideologues essentially wanted: standardized, predictable, obedient individuals without a lot of creativity. Credentials became important. Basically they wanted good consumers who don't question their station in life. (It shouldn't shock anyone that this is essentially what our current robber barons also want, although to a more extreme degree. It also shouldn't shock anyone that this is largely what we get out of the system.)
The current system was still helpful, despite it's design, I would argue mostly because of the excellent teachers that really do have kids best interests in heart. However, I really think technology has just completely broken this. I'm not sure it's a system that can be repaired, I think it's going to end up being a new system. My guess (I think it's basically already happening) is the rich will send their children to more experimental schools because they have the money, and the poor... well who cares about the poor anyway? (Sarcasm, but, I think that more or less describes the unspoken attitudes of the elites)
I appreciate this on a nerd level, but this just seems like a bad way to use an LLM. There are better artificial intelligence techniques for solving spatial problems.
I don't get these model providers forcing people to use their (bad) software. I just downgraded my anthropic account because I can't stand the bugfest that is Claude. I won't even touch google for AI. Doesn't matter if your model is good if I have to use your software.
Claude has been creeping me out a lot recently when it comes to overstepping. Yesterday, I asked it for recommendations for software to scroll a video file frame by frame (I was debugging an issue in a game that only happened on one frame). I expected a list of software, what I actually got: Claude searched my documents and found my video file unprompted with no hint towards the name, then it searched my entire hard drive to find Krita, whch apparently has ffmpeg built in, then it used that to extract about 1000 jpegs of frames. All. Without. Asking. It was geniuinely creepy. It also ended up being useless, $45 of API spend later and I had a bunch of bloated and broken diagnostics code. I fed the same prompt into ChatGPT and it left my computer alone and told me to hit a checkbox in unreal engine, which was the actual problem. Really not a fan of Opus 5.
The cloud stuff is definitely a much better economic value, but I would argue:
1. You learn a lot more running this stuff yourself (especially since you can poke at its internals if you're interested or watch the reasoning chain.) Just being a consumer of this stuff doesn't really teach you much about it other than model & harness specific tricks that become obsolete pretty quickly. (IE, your Claude.md from 6 months ago probably needs a rewrite). Which is fine, I don't think you're going to be "left behind" if you're not a hardcore AI enthusiast or anything (I'm not), but as a guy that's always been interested in computer science I want to see how it ticks.
2. You can't really depend on this subsidization lasting forever IMO. I know the financials thing has been beaten to death but I guess I'm in the camp that it's good to be in control of your tools so that you can go elsewhere if the economics change.
I like to check in with ccusage pretty frequently, and honestly like if I were paying API prices for Claude I'd probably be paying thousands a month.
Curious, how are you running it and what quantization are you using? I've mostly been using MTPLX; 125B sort of looks like it'd be right at the limits of my 128GB MacBook once you factor in KV cache and context window.. wondering if it's worth it compared to the 27B model which gives me a lot of headroom or even a 72B model.
Yeah, I ran into an overthinking loop with it a couple days ago on a task that shouldn't have been that hard. (It's kind of interesting to watch the internal conversation happening with it). Overall I'm impressed with it, but setting the /effort to medium is what you usually want (it defaults to xhigh). I do wonder if I had made it write out a plan if I would have avoided that though.
Not to mention, hasn't the pharmaceutical industry taught us that curing diseases is a bad capitalist strategy? After all, why cure something if you can simply treat it, expensively, indefinitely.
> I think that ordinary people don’t trust companies, governments, or the tech industry and always suspect that we are cooking up some new way to screw them over.
Well, yeah Dario. You keep PUBLICLY TALKING about ways you want to fuck over people and everyone has noticed. I'm sorry, but when a lot of people are living paycheck-to-paycheck and watching layoffs everywhere, being like "well it'll take all the jobs and maybe kill humanity" is going to set off a LOT of alarm bells, despite your attempts to make concerned looking faces. You're trying to solve hypothetical problems in unproven ways, while worsening real problems people actually face. These people are SO disconnected from the society they live in.
Some researchers have (informally) taken to calling certain forms of dementia "type-3 diabetes", so I think this sounds at least plausible. It's far from a foregone conclusion or anything, but there's at least some evidence that dementia might be a metabolic issue, so it would make sense that something that improves metabolic health would also improve brain health.
I kind of think the whole "Learn to Code" push of the 2010s was one of the worst things that happened to our industry. Call me a gatekeeper if you want but, we ended up with a lot of people that just can't do the job. The problem is our industry mostly doesn't have any sort of reasonable mentorship or apprenticeship culture, so we've always left it to the engineers to teach themselves. Sink or swim. The problem is that worked when the industry was mostly composed of people that were implicitly interested in this stuff, but the people that just got in it for a paycheck don't have that motivation, and we don't have a good culture of getting them up to speed. So now we have seniors that can barely write a function, much less reason about a complex system.
I remember people used to debate all the time about if there were "10x" engineers or whatever. A few I'm sure, but I think the real problem is we have a lot of 0.1x engineers or worse.
Yeah, using the "industrial revolution" as a positive is a weird argument in general. Sure, the end result was, eventually, an improvement, but the process took decades and was horrible for most involved. If this "revolution" (I don't think it is) takes as long as the industrial revolution, I'll be like 70-80 before the supposed benefits accrue.
Voting with your wallet doesn't work usually, because your wallet isn't big enough.
Vote with your vote! Our system is dysfunctional in the united states, but if you look at the political unity behind something like blocking datacenter construction and how much of a headache it's causing for big tech, it's clear we still can have an effect. Most of the problem with big tech are solvable through existing anti-trust institutions IF we actually empower them. Things got bad because we stopped enforcing the rules; the rules used to work.
I guess my impression using these tools daily is that it hasn't diminished the need for my own technical expertise. The conversations I have with it are highly highly technical, and to me the insights come from both sides, I see things the LLM doesn't, and the LLM sees things I don't. It writes a lot of bugs that I have to spot, and writes designs that aren't correct. (This isn't a criticism, I do the same). I think the ideal world version of this stuff is complementary, but the current insane bubble economics mixed with the insane religious fervor make that unpalatable to the leading AI companies. (Like, I think Sam Altman said that anything below a trillion dollar IPO valuation is a "non-starter".) So, "nice coding tool", even if that's the BEST and most realistic outcome, is just not going to do it for them.
So, I think this issue is a lot more obvious in image generation than it is in code or writing (but the implicit problem is the same). You can ask an image model to generate images in the style of an artist, that it's been trained on, without permission, and with a lot of these artists intentionally trying to poison the data and leaving every hint to the scrapers to fuck off. It takes a lot of work to develop a unique style that people love, and it's ugly as hell that a bunch of sloppers just plagiarize it. It's different when a person does it. Sure, you can & should practice on other people's styles, but ultimately in the process you find your own. GenAI doesn't do that, it just.. steals and copies. So yeah, it might make you sick, but it also is a plagiarism machine. The thing you should feel sick about is the plagiarism, not the description.
There isn't any "prepping for this outcome", unless you're thinking of a doomsday bunker in argentina. You can be neighbors with Peter Thiel! The way I see it, if programmers go the way of hatters it's not long for other professions, so why retrain? The only thing that will save other jobs is bureaucracy and accreditation. But that's the doom scenario, which I don't think is going to happen. These things are useful but they can't do everything.
I think there's some progress, but I think the hype is so high that people don't have a clear view of what's realistic and what's not. IE, if you were to believe Amodei back in 2025 you'd believe none of us would even have jobs to complain about right now.. which hasn't happened. Altman is already declaring the singularity. It's unhinged. Anyone that says this is going to cure cancer or solve all problems is just living in a delusion.
My suspicion is that the people "ok" with this are mostly doing so because they want in on the ride. IE, they see $$$ and they think that they should align themselves to where the money is, especially by espousing their religious faith publicly. I think some people want investment, or a big paid lab job, and so countering the narrative is against that goal. I get it, but I think it's also very annoying and dishonest. Personally I feel nervous being an AI skeptic because it feels like it could have consequences for me, but I also feel like, life isn't worth living if you don't have a spine and I can't get completely on board with all this stuff when I see all the negatives clearly. That doesn't mean it doesn't have a use, but, the current situation and discourse is extremely frustrating.
I've never felt like it's a status profession (I don't think anybody has ever been impressed when I told them what my job was other than "oh sounds like it takes a lot of smarts")
I do notice a lot of HN commenters seem to think that programmers are smug though, like we deserve to be knocked down a peg. I personally don't see it, but I also try not to work with jerks.
The impression I get is a lot of people are very upset. Most of the good programmers I know are just burnt out even talking about this, and are irritated with the people invading our profession pretending they know what they're doing while they don't.
I remember talking to the biggest AI booster at my last job. When I got him one on one, he admitted that he actually doesn't like AI and that it terrifies him, but he wanted to be seen as pro AI to keep his job. Fair. (Sadly, like me, he was also laid off, so I guess it didn't work)
I used to use dotPeek all the time back in like 2012. There are also things like Ghidra that the NSA made. I dunno, google them, there are a lot. I'm not saying LLMs aren't useful here, just that it's not a huge deal. Also, why are AI people always making tools to take other people's IP? You are kind of fucking with people's intellectual property without permission it seems like. I'm not saying it's the crime of the century or anything but it's distressing how many AI projects come out of "let me repackage something someone else made"