23 ms·
>looks at the top X hits and then interprets it to be the correct answer. LLMs are truly reaching human-like behavior then
by Almondsetat 2y ago
>looks at the top X hits and then interprets it to be the correct answer.
LLMs are truly reaching human-like behavior then
- yoyohello13 2y agoThe longer I've been in the workforce, the more I realize most humans actually kind of suck at their jobs. LLMs being more human like is the opposite of what I want.
- _heimdall 2y agoThat could very well be because the jobs are effectively useless. By no means does that mean the people are, nor is what the income allows them to do. But most jobs sure do seem pointless.
- homarp 2y agohttps://news.ycombinator.com/item?id=36373190 https://news.ycombinator.com/item?id=36373190
- bravetraveler 2y agoThe work is mysterious and important
- pdimitar 2y agoBoth are true, separately in different situations. And sometimes both at the same time. There are a lot of useless employees out there. So, so much. And a ton of bullshit jobs as well.
- Workaccount2 2y agoIt's kinda like online games. Most people who play a game are not too great at it, a large subset is pretty good, and then it's smaller and smaller groups as the ability increases. At the top you get the people who are true pros, they write the books, the guides, they solve the hardest problems, and everyone looks up to them. But spin the wheel and get a random SWE to do some work? It's not gonna be far off from an random 1v1 lobby.
- PickledChris 2y agohttps://danluu.com/p95-skill/ https://danluu.com/p95-skill/ Continues to apply
- anal_reactor 2y ago> And for games like Overwatch, I don't think improving is a moral imperative; there's nothing wrong with having fun at 50%-ile or 10%-ile or any rank. But in every game I've played with a rating and/or league/tournament system, a lot of people get really upset and unhappy when they lose even when they haven't put much effort into improving. If that's the case, why not put a little bit of effort into improving and spend a little bit less time being upset? Interesting read, but I feel like the author could've spent just one more minute on this sentence. How good you are at given activity often doesn't matter, because you're mostly going to encounter people around your own level. What I'm saying is, unless you're at the absolute top or the absolute bottom, you're going to have similar ratio of wins to loses regardless whether you're a pro or an amateur, simply because an amateur gets paired with other amateurs, while a pro gets paired with other pros. In other words, not being the worst is often everything you need, and being the best is pretty much unreachable anyway. This can be very well extended to our discussion about SWEs. As long as you're not the worst nor the best, your skill and dedication have little correlation with your salary, job satisfaction, etc. Therefore, if you know you can't become the best, doing bare minimum not to get fired is a very sensible strategy, because beyond that point, the law of diminishing returns hits hard. This is especially important when you realize that usually in order to improve on anything (like programming), you need to use up resources that you could use for something else. In other words, every 15 minutes spent improving is 15 minutes not spent browsing TikTok, with the latter being obviously a preferable activity.
- jimbokun 2y agoUnfettered capitalism is pretty good at figuring out which is which. It’s pretty core to Elon Musk’s animating philosophy: get as many jobs as possible then see if there’s any negative impact. Not as appropriate in a government setting where the impact goes far beyond personal profit and loss.
- exe34 2y agoif a CEO can run 7 companies and still play monopoly with the ship of government, then maybe CEOs aren't really that useful.
- matwood 2y agoThe problem is defining negative impact and also timing. For example, I can stop doing backups and save time and money. There is zero negative impact right up until the point I need to use the backup, then the impact is catastrophic.
- Toenex 2y agoMaybe we already have Universal Basic Income, you just need to have a pointless job to collect it.
- ornornor 2y agoI never thought about it this way, but it does make sense.
- nyarlathotep_ 2y agoYES, this is exactly the case and why the Twitter layoffs and now the "DOGE" purge is a terrible thing (even in cases where it was totally legitimate to eliminate "waste"). "They had useless make-work jobs and sent 4 emails a week and watched TikToks the rest of the time" So? There's FAR too many people and nowhere near enough jobs for a large portion of people to do something that is both "real", and provides actual economic value. Far more important that people have some form of dignity and can pay to feed their families and live a life with some material standard. Anyone who's been in a corporate role knows there's loads of people that have a dubious utility and value--and people with "tech skills" are NOT exceptions to this rule, at all.
- mschuster91 2y agoWe don't just have "bullshit jobs" (which is an actual term these days), we have a "bullshit economy" as well - centered around advertising because without advertising most of the bullshit just wouldn't sell. Like, if you already got a car, you can drive it for 10-20 years easily, or more if you take well care of it. But advertising makes you think you "need" a new car every few years... because that keeps the economy alive. You buy a car and sell the old one to someone else who can't afford a new car but also wants a new one, so their old car goes off to Africa or whatever to be repaired until truly unrepairable. But other than the buyer in Africa who actually needed a new car, neither you nor the guy who bought your old car would have needed a car. And cars are a massive industry that employs many millions of people worldwide - so if you'd ban advertising for cars, suddenly the bubble would pop and you'd probably have a fifth of the size remaining, and most of it from China because the people in Africa can't afford what a brand new Western made car costs. Or Temu, Shein, Alibaba and godknowswhat other dropshipping scammers. Utter trash that gets sold there, but advertising pushes people to buy the trash, wear it two times and then toss it. A giant fucking waste of resources because our worldwide economy is based on the dung theory of infinite growth. It has worked out for the last two, three centuries - but it is starting to show its cracks, with the planet itself being barely able to support human life any more as a result of all that resource consumption, or with the economy and the public sector being blown up by "bullshit jobs". We need to drastically reform the entire way we want to live as a species, but unfortunately the changes would hurt too many rich and influential people, so the can gets kicked ever further down the road - until eventually, in a few decades, our kids are gonna be the ones inevitably screwed.
- eru 2y ago> But most jobs sure do seem pointless. Do you include the private sector? Why do corporations engage in this kind of charity? Do we need more competition?
- blackqueeriroh 2y agomost jobs are absolutely not useless. They might seem useless to you, but the work has to get done. Personally, I think that a receptionist as a building is useless, but I would be pretty pissed off if my packages kept getting stolen or I had to go get each one when it came at my place of business.
- isaacremuant 2y agoOr maybe just extremely inefficient due to the huge complexity of reality and how it hides a lot of the power dynamics and real decision making. Big entities are such that if you take it all down, you feel the side effect of output (maybe value, maybe something else) but if you take Hugh chunks, you might not feel much because they're so extremely ineffictive and value creation doesn't correspond with value received for the individuals that created it.
- throw10920 2y agoWhat's the evidence for "most"?
- h2zizzle 2y agoMany jobs are quite helpful and even necessary, if done for ~2 hours a day. They become "useless" in aggregate when they're forced to be minded by the same person for 8 hours (because of opportunity cost, effects on health well-being, etc., you end up "breaking even" or worse on QoL and net productivity). Overall economic productivity is high enough that a lot of positions could be split into 2 or 3 short shifts, at full pay - IF you don't factor in the various financial boondoggles that we've gotten ourselves wrapped up in. If you made the decision to wipe out a lot of these obligations (mostly to rich people), we could get to that kind of set-up, solvently.
- Paracompact 2y agoI imagine you're a fellow Graeberian. I feel the same way you do (and deeply so), but I don't have the confidence to give numbers, let alone such idealistic ones. How do you support your own numbers?
- adverbly 2y agoDo they suck at their jobs or do their jobs suck?
- shreezus 2y agoThis is why agentic AI will likely cause a cataclysim in white-collar labor soon. The reality is, a lot of jobs just need "OK" performers, not excellent ones, and the tipping point will be when the average AI is more useful than the average human.
- jimbokun 2y agoThe excellent performs are only one or two turns of Moore’s law away from the OK ones.
- freehorse 2y agoIf Moore's law is applicable in such a case, that is.
- agrippanux 2y agoI had a similar conversation with my CEO today - how does the incoming crop of college grads deal with the fact AI can do a lot of entry level jobs? This is especially timely for me as my son is about to enter college. So I ended up posing the question to Claude and the response was “figure out how to work with me or pick a field I can’t do” which was pretty much a flex.
- ornornor 2y agoParadoxically, the hardest jobs to automate are physical jobs it seems. A white collar worker is threatened by AI, blue collar not as much. I can totally envision AI software engineers (they’re already okay if you check their work), but as of yet there are no AI plumbers or mechanics. Maybe there won’t be, given the costs associated why producing physical machines vs software ones.
- metek 2y agoYour average white collar worker is certainly challenged, but I think the talent of neurodiverse people is going to become even more vital as average-ability people are more and more challenged. Of course, there's the saying: "A man is his own easiest dupe, because what he wishes to be true, he will generally believe to be true." and I'm neurodivergent, so it makes sense that my assumption that shit'll probably turn out okay for me is a foregone conclusion.
- wickedsight 2y agoUgh... I've been in IT for over a decade now and many of the vacancies I see, I don't consider myself/my CV good enough. Then I work with the people who get hired for these jobs and see how low they set the bar, even though their CV might tick all the boxes.
- CooCooCaCha 2y agoPeople suck at intellectual tasks but for stuff like locomotion and basic planning we humans are geniuses compared to machines. There isn't a robot today that could get in a car, drive the the grocery store, pick stuff off the shelf, buy it, and bring it back home. That's so easy it's automatic for us.
- bobxmax 2y agoIt's quite odd that people think of hallucinations as a dealbreaker for LLMs. Have they ever even met a human being?
- _carbyau_ 2y agoAnd when I find a human hallucinating at the job I absolutely need them to do, I avoid them where possible too! But honestly, LLMs are here to stay. I don't like them for zero verification + high trust requirements. IE when the answer HAS to be correct. But generating viewpoints and ideas, and even code are great uses - for further discussion and work. A good rubber duck. Or like a fellow work colleague that has some funny ideas but is generally helpful.
- aprilthird2021 2y agoThe problem is that human beings are far more likely to know what they don't know. And we build a lot of our trusting work environments around that feature. An LLM cannot know what it doesn't know by definition.
- theshackleford 2y ago> The problem is that human beings are far more likely to know what they don't know. I’ve spent a career dealing with the complete opposite. People with egos who can just not bare to admit when they don’t know and will instead just dribble absolute shit just as confidently as an LLM does until you challenge them enough that they just decide to pretend the conversation never happened. It’s why I, someone fairly mediocre have been able to excel because despite not being the smartest person in the room, I can at least sniff bullshit.
- aprilthird2021 2y agoYeah sure, some people do this. But average humans understand the limit of their knowledge. LLMs cannot do that. You can find the right person for a space where this knowledge of limitations is necessary. Can't find an LLM which does that
- nemo44x 2y agoEver heard the saying “good help is hard to find.”? It’s not bullshit, it really is.
- metek 2y agoI try to apply my layman's understanding of whatever law of thermodynamics states that a minimum of <x> percent of a reaction's energy is lost as waste heat; whatever you try to do in life, <x> percent of your effort is going to be spent dealing with people who are utterly incompetent. I try to apply it to myself as well; there's certainly many things I'm utterly helpless with and I want to account for the extra effort required in order to carry out a given task despite those shortcomings.
- matt_heimer 2y agoThe book Artificial Intelligence: A Modern Approach starts by talking about schools of thought on how to gauge if an agent is intelligent. I think it was mimicking human behavior vs behaving rationally which I thought was funny.
- lutusp 2y ago> The longer I've been in the workforce, the more I realize most humans actually kind of suck at their jobs. And if they don't suck at their job, they get promoted until they do: https://en.wikipedia.org/wiki/Peter_principle https://en.wikipedia.org/wiki/Peter_principle .
- dstroot 2y agohttps://danstroot.imgix.net/assets/blog/img/comic2.png?auto=format%2Cenhance&fm=avif%2Cwebp&fit=max&w=3840&q=75 https://danstroot.imgix.net/assets/blog/img/comic2.png?auto=...
- pizza 2y agoIt would probably be really great for web searching llms to let you calibrate how they should look for info by letting you do a small demonstration of how you would pick options yourself, then storing that preference feedback in your profile’s system prompt somehow.
- rendaw 2y agoHere though they're not replacing a random person, they're replacing _you_ (doing the search yourself). _You_ wouldn't look at the top X hits then assume it's the correct answer.
- dartos 2y agoSplitting hairs, but LLMs themselves don’t search. LLMs themselves don’t choose the top X. That’s all regular flows written by humans run via tool calls after the intent of your message has been funneled into one of a few pre-defined intents.
- eddd-ddde 2y agoHow do you know? You could 100% create the tool to search and chose results, go through links, read more pages, etc.
- dartos 2y agoI know because that’s how these systems are built. I’ve built systems like it. If it was something brand new, Anthropic would be bragging hard about it. > You could 100% create the tool to search and chose results, go through links, read more pages, etc. That’s exactly what I’m saying. _YOU_ could build a tool that does that. The LLM essentially acts as an intent detector, not a web crawler.
- LightBug1 2y agoDegenerative AI ?
- wvh 2y agoBe careful what you call AI, you might just get what you wish for...
- johndhi 2y agolol
- ChrisRR 2y agoBold of you to assume that most people even bother googling simple questions