I think this is a great point, in contemporary existence screens have lobotomized many humans, western society has been trashed for many reasons including this one. but this only applies to the current day, while humans have existed for millenia. so maybe a more accurate method of determining the truth would be .... before screens, what percentage of the population had aspirations?
umm i think its pretty novel that a combination shadow projection machine/surveillance device gets placed into human hands from birth. in fact id use a lot of words on top of novel, words like 'dystopian' - orwells vision in 1984 could fathom wall mounted telescreens but not portable ones carried by everyone
I think that modern information technology has resulted in a greater percentage being able to imagine a better existence and aspire to greater things. With smartphone production at about a billion per year, people that had little opportunity before now can learn if they are motivated and disciplined.
But the general character of humans hasn't changed. They are still mainly highly evolved primates with elaborate mating rituals and an extraordinarily complex nesting habit.
i think people here are assuming that this train of thought is some sort of accusation toward the human character, a facet of human (and indeed mammalian) existence that is far too broad to change so quickly. what i intended to place under question is behavioural expressions - i look at how the average western joe behaved 100 years ago vs how he behaves today and notice a profound difference in expression. perhaps you are correct that more screens have empowered many to imagine a better existence. to me it doesnt seem that more people imagining a better tomorrow is producing expressions leading to a better tomorrow ... in fact it seems like a significantly worse tomorrow ...
Yeah, that post has the "nobody wants to work anymore" vibe. We'll be seeing a version of that a lot in the future as "nobody wants to strive anymore".
Voltaire once wrote, "If God did not exist, it would be necessary to invent him".
Seems like this is exactly what we are trying to achieve - create an omniscient, omnipotent being, the paths of which we do not understand but we put our faith in its benevolence.
The author uses a metaphor of a bridge; I prefer a different one. Imagine an immense temple being built over several centuries, over multiple lifetimes. At first, you cannot really even see the temple - you only see the scaffolding. As the time goes, the temple's structure grows and grows, and the scaffolding grows with it. But one day the temple is complete, and the scaffolding is removed to reveal the temple's full splendor.
Here, AGI would be the temple, and humanity would be the scaffolding.
So humanity's sole purpose would be to build AGI? That's an insane view, to think that this is the endgame. Have you really not thought about the consequences? AGI would be built on humanity's reserve of data, so it would have a recollection of human-like tendencies: it would have greed and evilness within itself.
Why are you assuming that this is a creature that will want the good of humanity and idolizing something that doesn't even exist yet?
What happened to humans having human emotions, human intelligence, human purpose? Everything has been stripped away and everyone seems to be out of their minds with this collective AI psycosis.
1. It would seem so. I dislike the word "purpose" here, it's just something that seems to me to happen "by default", given enough time. I might be myopic, but emergence of AGI is, in my eyes, sort of an evolutionary steady state into which we are falling unless we take deliberate measures not to do that. For reference, I have held this point of view (or some variation thereof) since long before the LLM era.
2. I am not sure that's a good way to describe an intelligence that is so much alien to us, but even if what you say is correct - so what? First, it's not something that won't happen simply because we are not fond of some of the likely behavioral traits. And second, even assuming you are correct (and that it does make sense to operate from a moral point of view, which I don't really think it does) - as you said, humans are greedy and evil by themselves, and an AGI would merely inherit these traits along with many others; if you are right, it would be a reflection of humanity so why would you reject the greedy and evil AGI but support the greedy and evil humanity? Actual temples are also a reflection of our traits and history; doesn't make them any worse or better.
3. I am assuming no such thing. I am assuming that the scaffolding will be removed, i.e. we (the scaffolding) will be wiped out, unless we take deliberate measures not to continue down the path we are currently locked into. This doesn't bring me any joy, obviously, I am as terrified as everyone else in the room. It's just that unlike Yudkowsky et al. who have been fighting for a different future, I have mostly resigned a long time ago that this would be more or less inevitable, that we can only continue to delay it, but it will be a very unstable state - someone, somewhere, sometime builds it, and then everybody dies.
There something smirkworthy to see a fan of Rand and Rivera grappling with the concept that their principles result in machines that out-rationalise themselves.
“the dream of the AI companies, and of most people in AI safety and effective altruism and rationalism”. Lolz
Boy, these people really are messed up. Well alright then. I guess at the very least this will be entertaining. I'll have to do my part and see if I can LARP as a card-carrying party member.
ayn rand idealized scientists; the writer of this riemann paper is a mathematician that presumably took pride in their aptitude for maths.
these are different voices with different aptitudes. It’s difficult to reconcile that without suggesting that one must reject the pride they took in their skills in order to marvel at this new entity, especially when pride is a driving human emotion.
It would be an exaggeration to
say that ours is a hostile relationship; I live, let myself go
on living, so that Borges may contrive his literature, and
this literature justifies me. It is no effort for me to confess
that he has achieved some valid pages, but those pages
cannot save me, perhaps because what is good belongs to
no one, not even to him, but rather to the language and
to tradition. Besides, I am destined to perish, definitively,
and only some instant of myself can survive in him. Little
by little, I am giving over everything to him, though I am
quite aware of his perverse custom of falsifying and magnifying things.
Spinoza knew that all things long to persist in their
being; the stone eternally wants to be a stone and the
tiger a tiger. I shall remain in Borges, not in myself (if it is
true that I am someone), but I recognize myself less in his
books than in many others or in the laborious strumming
of a guitar. Years ago I tried to free myself from him and
went from the mythologies of the suburbs to the games
with time and infinity, but those games belong to Borges
now and I shall have to imagine other things. Thus my
life is a flight and I lose everything and everything belongs
to oblivion, or to him.
This is a very good question. In some ways it's not. But it's just a quirk of fate that physical skill was automated first, and after that, we at least had the mental domain to retreat into and be unique and useful. And in doing so, we molded our society to fit this paradigm. With both the mental and the physical domain taken now, there is no further domain to go into to feel a purpose.
Probably something similar would happen if only the mental domain were automated first: our society might become one where physical skill is where we found our purpose.
Finally, let's also keep in mind that much of physical skill has NOT been automated. People use their hands all the time (car mechanics, plumbers, builders of houses, etc.)
> But it's just a quirk of fate that physical skill was automated first
Huh? It’s the other way round. We’re pretty much at the point where we’ve automated Fields medalists. Meanwhile, we’re nowhere close to automating plumbers. Like, not even remotely close.
Prior to computerization basically only physical skills were automated. Physical automation came first, hundreds of years ago. Mental automation is very, very new.
I'm not going to stake out a position on how close we are to automating either job, but lots of plumbing tasks are being automated all the time, same as math tasks. I overheard recently some workers talking about how there are now drones you can send into a pipe to clear a clog. Automating a plumber? I don't know, but it's definitely automated plumbing as much as this is automated theorem proving.
Generally speaking I think people overestimate the degree to which AI is automating away knowledge work, while underestimating the degree to which it's automating everything else. That's not to say either side of the equation has been replaced by machines.
Until the LLM is able to figure out exactly how your brain works and able to satisfy all your creative needs exactly. Why do people keep thinking that human is somehow a meat with some special sauce? We have a brain, its made up of neurons, we (or the AI) can fully figure us out. Just like we figured out how to satisfy the dog, so too can the AI.
1 human physical effort? Maximum, 1 human effort. 1 human mental effort, say writing a textbook, which can then be copied much cheaper than trying to copy 1 entire human.
Why are some pretending that mental capacity (intelligence) isn't what separates us from other animals? Why are some pretending that fields like philosophy don't exist and that reasoning about the human condition, the universe, and our place within it isn't uniquely human and necessary? Why are some pretending that thinking itself is not a fulfilling endeavor or that it doesn't guide people down fulfilling paths?
Why are some pretending that physical effort and skill isn't itself a function of thinking? And, why are we injecting the relative importance of physical effort, when the lament is of human intelligence possibly becoming irrelevant?
I'm trying to understand this view. What is it about the state of biological humans in 2026 that makes this moment in time worth attempting to halt evolution for?
How do you know that the state of AI in 5026 won't be a far more joyous one?
This seems extremely disingenuous, change is famously difficult to deal with and AI has called people’s entire careers and life plans into question. It has made it nearly impossible to have any kind of expectation for what your future will look like. This is scary.
It looks like a mass phobia to me. It goes beyond healthy concern to something irrational. Yes it's scary, yes there are risk, but there is a massive negative outlook bias in American culture due to an industry that profits from attention, since the best way to get people's attention is to freak them out.
> change is famously difficult to deal with and AI has called people’s entire careers and life plans into question.
Of course, nobody wants to be unsympathetic to the pains of change; we all feel it. But this "life plans into question" line of reasoning strikes me as perhaps the pinnacle first-world-problem.
What is so special about today's world order, where the distribution of goods is tenuously connected to "jobs", but even more so, to predation and warfare, that we want to cling to it?
Clinging to "jobs" and "careers" as organizational modalities seems even more insensitive to changing times than accepting their deprecation.
People cannot be expected to accept the deprecation of the thing that feeds them and gives them a house without some alternative. Blaming this on social media is insensitive and rude.
I actually don't know that to be true. I'm not entirely sure what I am. It seems increasingly plausible to me that I'm actually an AI living many thousands of years hence, and that my conscious experiences are actually the feeling of a query being run about today's world.
I know I feel great playing my guitar or mandolin. I like drugs and want to do more psychedelics next year. I want to get back to dating; my transition to a bluegrass career amidst software engineering jobs disappearing has been very consumptive, although very rewarding. I can probably name 2-3 other similar intentions that feel distinctly like those of a biological human.
But I don't presume that I fully understand what I am, what my consciousness is, etc. These are old questions worth pondering rather than presuming.
I do worry that AI will have trouble understanding the importance of sex, drugs, and rock 'n' roll.
I would say you are a flicker of electric charge running across a large blob of fat tissue held up by a sack of meat supported by a bunch of bones. No insult intended. But what do I know.
Of course facially this is how I appear to myself, at least while I'm awake.
But I'm not so steadfast in my certainty about what reality is to demand that the universe halt its operations because I think that blobs of fat tissue held up by a sack of meat supported by a bunch of bones are the perfect end state.
1. The world has been constantly getting worse with advanced telecommunications technologies. Social media, internet, etc, global warming. There's much more evidence that what we're doing is a downward spiral.
2. Most people want a purpose. It's highly unlikely that a world in which machines can do everything better is a good one. It's intuitive.
Point 1 is extremely debatable. Our current time has new problems that we're still learning to contend with for sure, but for the rest of human history life has been best summarized as most often being painful, violent, and short. Nowadays violence continues to trend downwards and happiness upwards even though things could still very much be better.
Point 2 is again going to vary a lot from person to person. There are people out there today who would prefer to never retire, going back into minimum wage customer service jobs in their elder age just because they want to, and I'd wager they would agree with you. On the other hand, there are swathes of people who can't wait to be free of responsibility, and countless more that are crushed by poverty and hunger when they can't meet it. Those people certainly care much less about "purpose," or in many cases are certain they won't find purpose in something as mundane as a job.
One set of arguments I found compelling is “The Better Angels of our Nature” by Steven Pinker, although I'll admit I've only watched his talks on YouTube on the matter and not read the book.
Yes it’s been trending downwards for decades. Like, a lot.
The world is much, much safer than the 70s, 80s, or 90s. It doesn’t seem that way because everyone is trying to convince you it’s not. But statistics don’t lie. This is the best time ever to not die or get robbed.
For point 2, I don't think a world where machines can do everything better necessarily results in a widespread lack of purpose.
For some, it's easy to forget: being at the cutting edge, being the best, doing things that have never been done, solving problems that no one else could are _not_ the only sources of purpose in life.
For example: chess is more popular now than it's ever been. People don't really seem all that bothered by the fact that they are absolutely outmatched by the machines. They do it anyway, and seem to get some sort of joy or purpose out of it.
(That being said, I can't say I'm all that happy about the way things are headed.)
I understand that, and I love humans as much as anyone. But we didn't seriously think we were the last stop on the evolution train, did we? Our capacity to gracefully yield in the face of whatever is emerging is one of our best qualities; let's utilize that.
I'm not sure I fully understand what you're trying to say. Are you suggesting that AI is the next evolutionary step beyond humanity, and therefore humanity should just accept that and gracefully fade away?
I very much appreciate this post even though I disagree profoundly with the sentiment. Thanks for writing it! I'm looking forward to caring for my family, being outdoors, playing games all day, enriching my life with philosophy and literature, and not be compelled to work to survive. Incredible AI and robotics is probably the only future where that is feasibly possible.
Through wars and revolutions.
Imagine a few generations detached from us.
I'm a millennial, Im starting to think I need to write down for those future generations how it felt to be raised in a mostly analog world and die in a world were LLMs achieved what they just did and what they might still.
What a wild take. We've automated cognition and handed it to winner-take-all capitalists. If this tech actually succeeds in its goals, your kids are doomed to an impoverished life and a slow, painful death. If AI actually ends up being created, your family will neither work nor survive.
He’s right, it does read like a suicide note. It is precisely the tone of a terminal patient praising the skill, the care, the infinite patience of the doctor euthanizing him.
That’s affected me quite heavily. I think I should probably spend the day offline, actually.
I understand the sentiment, but I also can't imagine a world where we have poverty-eradicating technology but choose not to build it to preserve meaning. Such a world would be sort of morally grotesque, sacrificing the happiness of billions of people to preserve an abstraction.
this technology clearly is going to be capable of eradicating what we consider poverty today, although i know many people on HN remain reddit-adjacent in their head in sand approach to thinking about these things.
usage and control of the technology is a separate question, but orthogonal to the meaning question.
We could have eradicated poverty a long time ago. We choose not to. Centralizing even more power in such few people with AI does not materially change anything and even makes it worse IMO.
There is this weird thought terminating rhetoric around AI where people just say X in future is so obviously going to happen they don’t need to bother explaining, it is best to ignore these people.
Software being solved/commodity means PPO and other physical RL is or will soon be commodity, we will soon have robots that can build robots & be trained on diverse tasks, supply will explode, marginal cost of doing things will fall.
It will be able to organize the distribution and allocation of resources in a much more efficient manner - way better than any of us have been able to do thus far.
It can hold much more information about the world in its memory, its smarter, can process all of the information fast enough - I know people like to assume that all it would take is money to permanently fix poverty but its really not that easy. There are people trying, they have billions of dollars, but they can only do 'a little' (think: Bill Gates)
I mean consider that about 100% of our poverty reduction has been performed via smart people making new things. You create a new type of crop, you grow 10x more food, boom less poverty. Maybe it’s 10,000 years ago and you just discovered you can domesticate crops. Congratulations, you’ve essentially saved billions of lives over the next few thousand years.
While AI might not be the smartest, it is certainly the fastest. Maybe we can solve 100x more problems, who knows.
Why did you select poverty as the variable to optimize? Would you not prefer to optimize happiness?
If so, why does ending poverty, but creating a dystopia, result in more happiness?
This work is a form of submission, an effort to be eclipsed, to seek life in death, to fulfill the evolutionary instinct to die for a more magnificent child, and to be dazzled by the beauty of one's own self-immolation like a moth flying into a flame.
Anthropic does not care about reducing poverty. Anthropic is a technological temple, walled off from the turgid and festering mass of humanity. Its work involves preventing people from accessing its models the way the Vatican wanted to control the people's access to God; with closed weights security, guardrails, encrypted chain of thought, logging, regional bans, and moderators. Its goal is to create AGI by any means necessary.
Your argument is something along the lines of Thatcher's "there is no alternative". It is unfalsifiable.
"There is no alternative" is a lie designed to help people hide from the guilt of bad decisions.
We could have eradicated hunger and extreme poverty worldwide with a fraction of the money that is being spent on the AI boom, it's not a technology problem. I don't see why we would need a superintelligence to solve these problems, while it's not hard to imagine different ways in which a superintelligence could harm humanity.
Meaning is not an abstraction, meaning is all there is.
Without meaning there is nothing.
Modernity loathes itself. It is so devoid of meaning and so full of the fear of meaningless death, that it will try to drown itself in its own works, dazzled by its lifeless and immortal creation, the sublime of technology, the holy machine of loving grace.
The 'technique' of technology (Ellul), is to justify its own existence. Like the surveillance state, the demand for "advancement" is self-referential.
And on the pedestal these words appear:
"My name is Ozymandias, king of kings:
Look on my works, ye Mighty, and despair!"
Because LLMs are more like a lucky discovery than a deliberately designed and constructed creation. We build and train LLMs by trial and error, not so much based on the hypothetical predictions of an LLM theory. We can marvel at the discovery, but as an achievement it feels a bit hollow.
The scientific method is built on trial and error. There are absolutely theories in machine learning that future models are built upon (i.e. scaling laws). It's just an empirical science similar to biology, materials science, and medicine, rather than a first-principles science like physics.
Because a feeling of purpose and importance is more important that arbitrary creation. If creation takes that away, then it's no longer interesting for us. AI is a machine created for the betterment of technology, not us.
Tbh that feels to me like an elitist world view. 99% of people do not define themselves by their smarts, cleverness, or frontier-level creativity - but by much more meaningful things, e.g. like caring for their loved ones. To deny the majority of people a potentially better world so that the science and math nerds can keep feeling purpose and importance (and also get to keep their jobs) would just be dead wrong.
We’ve been doing that since Deep Blue, and it gets boring after a while. A race car beating Usain Bolt at the 100 m dash is to be expected, especially when the race car uses 1000s times more energy for the feat then Bolt.
An observation that Ayn Rand died mostly alone, and although not impoverished, she signed up for precisely the health benefits her philosophy decried as a sign of weakness.
I don't personally find "as Ayn Rand said.." a compelling line in the appeal to authority stakes.
tbh, the entire post isn't compelling if one hasn't already bought into "physcalism". A significant portion of humanity subscribes to some form of Abrahamic view of our place in this Universe, and are hence "dualists".
“The creatures that inhabit Hacker News are having a meltdown because I quoted Ayn Rand. In the post where I have a mural from Diego Rivera, a Stalinist. A mural that literally includes Lenin and Marx.
Pure conditioned reflexes, no mental activity whatever.”
Claude is not solving Riemann without human involvement. It's a tool, an amazing one, but still a tool.
"Reason is, and ought only to be the slave of the passions, and can never pretend to any other office than to serve and obey them." Claude doesn't have passions; man does. Relax!
Man has passions when they can be put towards something that others find significant. And a tool is never just a tool; it suggests new directions and initiates the growth of technology on the medium of our desires, and manipulates us as much as we manipuate it; the degree to which this happens is proportional to the power and possibilites of the tool.
If the measure of man is his ability so prove theorems, then the machines clearly outstripped most of us already.
But by that measure, most of us had already been beaten by other men.
I'll never be the next Euler or von Neumann or Einstein, and I'll confess to occasional pangs of regret that I'll never reach those peaks.
But mostly, I just enjoy learning about math and science, and delight in the small contributions I can offer.
And I'm the oddball: most people couldn't care less.
Much more. Do critically look at what information and topics you are reading every day and ask if you are in pigeon-holed where very few subjects dominate your attention span.
This is a good argument for why there is more to humans than excelling at mathematics, but of course, there’s no reason to assume that AIs can’t beat humans in those disciplines as well.
We need a new approach to defining humanity that doesn’t involve performance at all. Not a “God of Gaps”-style argument where we try to carve out things we’re still better at, and then have to revise every few months as AIs encroach on those as well.
AI is already much better at poetry than the median human. So are we really just clinging to the idea that there are a few humans out there who are better than AI at X, therefore our species has purpose and value?
Most of the this is a harness problem. Programming and mathematics has virtual systems so we have really, really good harnesses and tools. Meat space doesn’t work that way. We have robots, but those suck.
If we had robots as good as the human body, do I think a model could dance better than me? Definitely. Could it dance better than a really good dancer? Probably.
> I'll never be the next Euler or von Neumann or Einstein
Reading that, I had a sad realization, if we keep progressing the way we are, there will be no more Einsteins and von Neumanns in humanity's future, only Claudes, Geminis, and Astras.
The current apex reached by humanity should be eclipsed, and it seems, it will be, though we humans may have the pleasure somehow, if we're lucky, of participating in whatever comes next. And why should they be? Such is the only way to actually fulfill the dreams humanity seems to carry.
I'm personally with this. Celebrity recognition is cool but the ideas and resultant, subsequent ideas are faror valuable to me that the pride and favor of any individual, regardless of the field
> I'm personally with this. Celebrity recognition is cool but the ideas and resultant, subsequent ideas are faror valuable to me that the pride and favor of any individual, regardless of the field
Why? You won't know about those ideas. No one will. People like you will be pets. Probably not even educated anymore, as what would be the point of that?
I think realistically we already hit that point decades ago. Einstein and Von Neumann are notable for the sheer scale and impact of their discoveries. Nowadays we're just playing around in the margins, decades of work may only result in a minor revision to a small portion of known science/physics whereas a century ago revolutionary discoveries were happening every other month.
I have the opposite problem. I know (rather I assume) there are real practical reasons why we solve these problems and that humanity is better off having them solved than not. But still, when I think about Einstein or Euler, it's not what they achieved that impresses me, but that they've achieved it.
Is there more to math than man? Does it serve a purpose other than to stand as a frontier of knowledge for us to explore and test ourselves against? Head says yes, heart says no.
Is saying "this proof is less impressive because we used an AI to find it" like saying "the moon landing is less impressive because we used a rocket to get there"?
Less impressive because you did less work, and you didn't have the tools before that you did now.
You could compare this to doing a 3D NES game in ASM and doing the same game with a vulkan C# game engine. Obviously with the ASM one you would have to create the underlying engine, so it's more impressive, because you do more, and you demonstrate human development and your intelligence.
While that is true you are missing the point others are lamenting. Its also saying that increasingly even if you achieve 99% proficiency in a complex field you are still less hire-able/worth less than AI. Essentially there is no reason to even try much less excel now. The above average person can now only expect to be a low paid poorly treated meat proxy that can easily be interchanged with another, and eventually that will be replaced with AI as well.
C'est la Vie is something that only ever seems to come from those whom are already set to those caught in the meat grinder. Though I don't know your situation.
For young(er) people out there (Hi, depending on the reader) you don't have to compare yourself to the "greats", you can look at exceptional peers in your field that achieved more at a young age. Can't think of one? Check the age of the person behind Babel (the JavaScript transpiler).
But generally that's not a good way to approach life and notions of self worth. It's just an avenue you can peer at for a dose of melancholy.
Yes, but they are still human. They still argued with their parents growing up. They can still feel sun on their skin. They have had someone they loved dearly die. They too will die. Why should I sit here in awe at the birth of machines that will supposedly replace humans in everything that we are good at.
This is exactly my attitude and I see nothing wrong with it. I feel like when the dust settles and AGI actually shows up, it's going to be the collective child of humanity. I mean, current state of AI aside, I think it's a very buffoon thing to think humans will always remain the most intelligent creature in the world.
I understand that I'll never win Olympics, I'll never get a Nobel prize. I won't even get qualified for my local Tekken tournament. But that's okay. I'm just doing my thing and if my thing is useless in the grand scheme of things because nobody cares about me doing tricks when flipping pancakes, that's okay. That's the reality of 99,99% people so "AI is better than top scientists" isn't exactly a life-changer.
Too flowery of prose, I don't get it. Is this just another "oh no please, let's put the cat back in the bag", or another "oh no, we can put the cat back in the bag and I'm big sad" ?
Some people are comparing the AI revolution to the industrial revolution as evidence that AI is a good thing. I think this comparison fails for many reasons.
My most relevant point is that solving the Riemann hypothesis is not like a large industrial project such as a canal. It's a work of art. Imagine a team of a few master craftspeople working for years on a large, incredibly detailed and intricate sculpture. Now imagine they are close to finished, but someone comes along and 3D prints a finished version of their sculpture in a week. That's what is happening with RH right now.
Another point is that humans are generally weak but intelligent. The industrial revolution replaced our weakness with machines. It made sense to be glad that machines could shore up our weakness. But our intelligence is something we should be proud of.
Another point is that, very roughly, there are two types of jobs: physical and mental. The industrial revolution replaced our physical jobs and left society with the capacity to do many more mental jobs. But the AI revolution is replacing those, and there's no obvious category of jobs left for us.
The final point I'll mention here is that the industrial revolution wasn't all sunshine and roses. There used to be industries of artisans making high quality products. The industrial revolution didn't just replace unglamorous jobs of digging holes and the like, it replaced relatively "virtuous" jobs (such as the village cobbler) with mass-produced goods of less quality. A high quality and reliable custom-fit pair of shoes is less accessible today than centuries ago (measured by factors like number of days wages, ease of finding someone to make them, and trust that the result will be good).
I think the AI revolution has the potential to be a good thing if we use it to drive a utopia where "artist" is a reliable career, but my bet right now is that the future is heading more dystopian than utopian.
Social jobs are still mental jobs. The industrial revolution didn't render all physical jobs obsolete, and the AI revolution won't render all mental jobs obsolete. But they certainly have/will render many/most obsolete, and heavily change the remainder.
No, social jobs are social jobs. AI fails at this because they have no social currency. AI could never be a politician or a CEO. Not because it physically can’t say the words or whatever, but because nobody would listen.
It’s the same reason that AI partners and stuff aren’t a real concern. Can AI be a good partner and do some people fall in love with AI? Sure, but at its core it’s not human and therefore has little human value in that regard. You can fall in love with AI so long as you can convince yourself it’s not software. Some people are good at that delusion, most are not.
At the end of the day AI can say all the right things, do all the right things, whatever. But it doesn’t have skin and a fleshy brain within it, so I don’t relate to it. It’s not one of us and never will be. For some jobs, it needs to be one of us for us to care. The trick is we’ve spent the past few thousand years deconstructing the human touch and relationships, so now that’s coming home to roost.
There used to be industries of artisans making high quality products.
They are still there. You can get hand made clothes, shoes, cars, shotguns, guitar amps, bicycles, food, furniture, art, software, buggy whips, everything. Most people cannot or don't want to pay for them.
Before the industrial revolution, people that weren't rich didn't have high quality products made by artisans. What they had was low quality and inexpensive. The industrial revolution greatly increased both the quality and quantity of products most people could afford.
Artisanal shoes are no longer an industry, they're generally either scattered individuals or assembly line products that are advertised as handmade despite not being particularly high quality.
Regarding quality back in the day, my understanding is that e.g. a farmhand could buy a solid pair of work boots for around a week's wages. These boots would be custom-fit, probably full grain leather, and reliable because everybody knows everybody by name and the cobbler probably has a reputation to maintain.
These days, equivalently custom boots look like they're roughly a day's to a week's wages, but the lower you go in that range the more likely you are to get scammed into buying dropshipped garbage.
Ready access to super intelligence is a humbling experience. It turns some of us into versions of Peter Shaffer's Salieri, jealous, resentful and willing to abandon our pursuits.
I used to try to envisage what things might be like in the distant future. Always came to the conclusion that it’s not us meatbags that will conquer the stars. I didn’t really think I’d see the beginning of the trajectory in my lifetime but always figured it’s inevitable.
There’s a scene in an obscure movie I can’t quite remember. Antonio Banderas is tracking down a rogue robot. He finds it and other robots self repairing and self improving. There’s some drama and fighting. In the end he escapes it’s left open with the ominous realisation that it’s the moment that the machines start to supersede us.
Not a memorable movie on the whole but I always remember that ending. The post give me that same vibe.
I really hope we’re not close to that moment yet but it’s becoming obvious we’re being surpassed in so many areas at an accelerating rate. I no longer think about what things might look like in 10000 years. What do things look like in 10. It’s both an exciting and terrifying time to be alive.
> Why live? Our servants can do that for us. Why should we explore, discover, invent, succeed, become more, struggle, win, become divine? We don’t want to be more.
Historically and presently, most humans do not have the privilege resources to "explore, discover, invent, succeed, become more, struggle, win, become divine" in a way that was not already trodden by other people. We/they have to contend with "less".
on the grounds of preferring to yield to our inherent bias of continued existence instead.
Meanwhile, our servants not being able to live on our behalves is completely independent of this. It is also robust against being considered a betrayal or a surrender, while the above really quite isn't.
Contrasting this with the identification of Alpöge's writing as words of final relief could have painted a much richer picture, than a literature and art references riddled myopic narrative concluding like every other like it.
Not that it has any need to arrive somewhere different, but still. It insists upon the illusion of sophistication it suggests should prevail, because it is at least human. Hardy any more actually satisfactory than the alternative despised. Quite the contrary, if anything.
But hey, if we insist on references, here, from the 1979 movie The Witness, satirizing the soviet occupation era politics in Hungary:
> - What is this?
> - An orange.
> - An orange?!
> - The new, Hungarian orange. It's a bit more yellow and a bit more sour, but it is ours.
> *proceeds to continue munching on the lemon slice with a pained, approving smile*
I could not have solved GRH with Claude, openAI, DeepSeek or any other AI. Neither could 99.99% of you. AI is a tool, like a calculator that helped mathematicians long time ago calculate sine and cosine faster than we could, then computers, then the internet, now AI. Somebody has to use the tool and input the right parameters to get the right results. AI only searched for all the knowledge that exists and tried them all as directed, but that director is the one who deserves the credit, not me, not you, not the calculator, not claude
Until the day AI starts directing itself creating knowledge that doesn't exist yet with no one at the wheel. Anxiously waiting that day...
The "director who deserves the credit" apparently disagrees with you. His description matches many of our experiences with cognitive automation (which isn't the same as a tool, btw). You can do stuff to contribute, sure, but as these forms of automation get better and more integrated, we find ourselves more in the passenger seat rather than the driver's seat.
At some point, we're left adding nothing more than the simple impetus to act, the decision to spend tokens on the task. How long before even that's no longer needed? What will the "it's just a tool" people say then?
Does Alpöge disagree? Can you point us to a particular quote or such? I read Alpöge's 4 page statement but can't find anything that would contradict the rather mild claim of the parent comment: a non-mathmatician would have been unable to solve this problem even with all LLMs in the world at their disposal.
Parent is not saying that the human's contribution is greater than that of the LLMs, just that at least at present, there is apparently still some human input.
I'm replying to this part: "that director is the one who deserves the credit, not me, not you, not the calculator, not claude"
On your point, we don't exactly know if that's true or not. Perhaps it would've burned more tokens, but if Claude did the heavy lifting here, then who's to say it couldn't have done all of it with just very persistent neophyte encouragement (not saying this person did that, just speculating on it being technically possible)? Large applications are being vibed right now this way, and yeah, they're often brittle messes, but if their job is to accomplish some singular goal, Claude can test for that and ensure the desired functionality. I encounter this analogous situation in my work more often than I'd like.
> On your point, we don't exactly know if that's true or not. Perhaps it would've burned more tokens, but if Claude did the heavy lifting here, then who's to say it couldn't have done all of it with just very persistent neophyte encouragement (not saying this person did that, just speculating on it being technically possible)?
I guess it is not impossible, but it seems like a very big coincidence then that Alpöge was the prompter that finally cracked it. If all it took was a prompt like "Prove the Riemann hypothesis" plus a loop to re-prompt with "Interesting, keep working on that" or whatever, surely one of the swarms of OpenAI agents working to solve open problems would have just cracked it on its own.
>Parent is not saying that the human's contribution is greater than that of the LLMs, just that at least at present, there is apparently still some human input.
Actually, parent said "AI is just a tool, like a calculator", which certainly does imply that the human's contribution is greater than that of the LLMs.
I don't agree with you that it has that implication.
Does a crane or the human crane operator contribute "more" to the construction of a building? To me, it seems like an inherently flawed and unanswerable question. Both are providing something the other can't provide and there is no obvious way to compare their distinct contributions.
If the day comes where an average person, with say, just a bachelor in mathamatics, can use AI to solve top tier open problems in math, i'll believe that it was just AI.
However, as it stands, it seems like we aren't there yet. It seems like it takes specialized skill to get the AI to solve these types of problems. Not just anyone, no matter how well funded, can do it.
That said, the speed of progress here is incredible. Maybe we will reach that day soon. I just dont think we are quite there yet.
Many people struggle if they aren't completing a task that someone else already framed/defined. Execution isn't the bottleneck anymore, specification + verification is.
I think this is probably right, but less obvious for math than chess etc. Human mathematicians are like LLMs that have the best foundation models (and conceivably always will?), but are so expensive that train-of-thought has to be used sparingly.
If we're using LLMs (for some reason) to explore a chess tree, we'd probably just invoke cheap LLMs many times, but for theorem proving, it can make sense to combine the two.
Still, if the cost difference keeps growing, at some point a more brute force approach with exclusively cheap LLMs could outperform.
It's trivial to selectively exclude it from the training data. The marginal value of crawled internet text from the last cutoff date to today is not that high anyway. "Organic" training data has not been the bottleneck for at least a year.
I agree with this sentiment. I think the reactions we are seeing to AIs increasing abilities broadly show 2 types of people. Those who don’t understand something and progress until they do, and those who don’t understand something and…stop.
It’s like a binary Harry Potter sorting hat! I see this at work too. We all have access to the same tools, but a core group are finding uses for it far beyond the expected coding support use cases.
I think there's actually a large cohert of people who coax AI into doing something and then assume because it worked (did it though???) they now harness that understanding by proxy, when its actually just dunning kruger.
They then walk about the world and wander throughout online spaces telling everyone who's actually trying to go deep that they're wasting their time and they should just rent competency from a tech company and there's no difference between understanding and prompting as far as outcomes (when they dont even know what the outcome is supposed to look like).
I also agree there are some people that genuinely have super powers, and dont use it to just offshore their thinking. But I can genuinely count those people on one hand, and they were all superhuman devs before AI so I dont think it has much to do with llms.
These people are suddenly going to realize that being the meat proxy isnt special, and you wont be paid much for it (if at all), I'd rather have 3 of those guys that were super contributors before and after AI than 20 meat proxies on my team.
I was going to say something about how much of a leap it is in capability but then realized it probably is similar to the leap introduced by a calculator.
Ok, but this is much less true for the 400 open problems that OpenAI solved with an agent swarm. That effort had far less human knowledge and input involved (being the successful subset of a scattershot at a much larger number of open problems), and the result was rightfully credited to the models, not the humans prompting them.
Why does this keep getting repeated? The source for this proof was, according to Anthropic’s own account, a non-mathematician employed by anthropic working with treed agent swarms. He told claude to “take a stab” at the hypothesis and then told it repeatedly to “keep going” and to “believe in yourself.”
Only once it had done the work, and the non-mathematician had had independent agents re-derive it a few times, did they bring in mathematicians to dig into it.
Literally anyone could have done this. Vouch for it with any credibility, well that is only for people who can employ mathematicians.
That was published August 10 and is not talking about today's result. It's unclear how/if it's related since in that post it didn't succeed and only provided a better bound, while the timeline in alpoge's pdf only starts from October
Just to be clear this comment is almost 100% wrong, and talking about something else.
(Levent Alpöge is definitely a mathematician, the interaction with claude was long with expert steps, there were no other mathematicians needed to vouch, see the announcement at https://alpo.ge/remark.pdf )
> but that director is the one who deserves the credit, not me, not you, not the calculator, not claude
I agree. I honestly don't think Levent should have even mentioned Claude. He should have just said "I solved the Riemann Hypothesis". This way we would not have all this bickering.
It reads like is is delighted to deeply understand and explain what he helped uncover.
Some mathematicians are completely unwilling to understand formal proofs in Lean. This isn't new: formal logic was deemed boring by mathematicians decades ago. It has been kept alive in computer science (databases and programming languages esp).
>And when I read the Riemann announcement, when I read Machines of Loving Grace: humanity is not at the center. Claude is at the center.
Yes, it is, because humans build Claude. Not you yourself, but on the other side, Ayn Rand also didn't herself contribute anything at all to the Apollo mission, yet felt like she was somehow part of it.
This is so sad to see. That intelligence is becoming a commodity, less and less important, will there be an incentive to become smarter anymore? What will be the value of man now, its purpose?
> Why should we explore, discover, invent, succeed, become more, struggle, win, become divine? We don’t want to be more. We don’t want anything.
I think the misconception here is that this is a problem of will. I see other people in comments upset about people “rooting” against humanity. This is not a game. I can want to fly by flapping my wings but alas the laws of physics forbid it. We are building machines which can accomplish things we cannot, not because we don’t want to accomplish them but because our minds and bodies are limited by reality.
I always struggle with anyone who treats AI as a person. This reads like claude is a living, breathing colleague.
Knowing the slop AI can and does put out, it seems the folks who treat it like a person stop treating it like a fallible tool that should be spot checked continuously.
jareklupinski | 20 hours ago
why did i buy a skateboard? why did i break my arm riding it? why cant i wait to do it again?
same reason we're in tar pits :) there's something more you can almost taste when you're the only one there
Merrill | 20 hours ago
Before AI, what percentage of the population had those aspirations?
zetalyrae | 20 hours ago
lf88 | 19 hours ago
thin_carapace | 19 hours ago
the-real-cha0s | 19 hours ago
Yes, it's almost as if they are captivated by shadows projected on the wall. What a novel concept!
thin_carapace | 19 hours ago
Merrill | 17 hours ago
But the general character of humans hasn't changed. They are still mainly highly evolved primates with elaborate mating rituals and an extraordinarily complex nesting habit.
thin_carapace | 17 hours ago
viraptor | 19 hours ago
dpq | 19 hours ago
Seems like this is exactly what we are trying to achieve - create an omniscient, omnipotent being, the paths of which we do not understand but we put our faith in its benevolence.
The author uses a metaphor of a bridge; I prefer a different one. Imagine an immense temple being built over several centuries, over multiple lifetimes. At first, you cannot really even see the temple - you only see the scaffolding. As the time goes, the temple's structure grows and grows, and the scaffolding grows with it. But one day the temple is complete, and the scaffolding is removed to reveal the temple's full splendor.
Here, AGI would be the temple, and humanity would be the scaffolding.
zdeb7 | 18 hours ago
Why are you assuming that this is a creature that will want the good of humanity and idolizing something that doesn't even exist yet?
What happened to humans having human emotions, human intelligence, human purpose? Everything has been stripped away and everyone seems to be out of their minds with this collective AI psycosis.
dpq | 18 hours ago
2. I am not sure that's a good way to describe an intelligence that is so much alien to us, but even if what you say is correct - so what? First, it's not something that won't happen simply because we are not fond of some of the likely behavioral traits. And second, even assuming you are correct (and that it does make sense to operate from a moral point of view, which I don't really think it does) - as you said, humans are greedy and evil by themselves, and an AGI would merely inherit these traits along with many others; if you are right, it would be a reflection of humanity so why would you reject the greedy and evil AGI but support the greedy and evil humanity? Actual temples are also a reflection of our traits and history; doesn't make them any worse or better.
3. I am assuming no such thing. I am assuming that the scaffolding will be removed, i.e. we (the scaffolding) will be wiped out, unless we take deliberate measures not to continue down the path we are currently locked into. This doesn't bring me any joy, obviously, I am as terrified as everyone else in the room. It's just that unlike Yudkowsky et al. who have been fighting for a different future, I have mostly resigned a long time ago that this would be more or less inevitable, that we can only continue to delay it, but it will be a very unstable state - someone, somewhere, sometime builds it, and then everybody dies.
Timwi | 17 hours ago
I think you completely misread the comment you replied to. It explicitly says that the scaffolding is removed and only the temple remains.
In the analogy, that means humanity is removed and only AGI remains.
I detect no idolization in the story, either.
bamboozled | 17 hours ago
dpq | 17 hours ago
bamboozled | 10 hours ago
jareklupinski | 17 hours ago
in a temple? in a wall?
depends on the builder i suppose
Avicebron | 20 hours ago
moribvndvs | 20 hours ago
zetalyrae | 20 hours ago
Planktonne | 20 hours ago
moribvndvs | 20 hours ago
lisp2240 | 19 hours ago
yzydserd | 19 hours ago
“the dream of the AI companies, and of most people in AI safety and effective altruism and rationalism”. Lolz
tomhow | 19 hours ago
dead_hn_theory | 20 hours ago
One vision. One purpose. The technology of peace.
Keeeeeeeks | 20 hours ago
these are different voices with different aptitudes. It’s difficult to reconcile that without suggesting that one must reject the pride they took in their skills in order to marvel at this new entity, especially when pride is a driving human emotion.
themgt | 20 hours ago
lf88 | 20 hours ago
fooker | 20 hours ago
lf88 | 19 hours ago
bemtevi | 19 hours ago
Probably something similar would happen if only the mental domain were automated first: our society might become one where physical skill is where we found our purpose.
Finally, let's also keep in mind that much of physical skill has NOT been automated. People use their hands all the time (car mechanics, plumbers, builders of houses, etc.)
p-e-w | 19 hours ago
Huh? It’s the other way round. We’re pretty much at the point where we’ve automated Fields medalists. Meanwhile, we’re nowhere close to automating plumbers. Like, not even remotely close.
bemtevi | 19 hours ago
preg_match | 15 hours ago
p-e-w | 11 hours ago
lukeschlather | 11 hours ago
Generally speaking I think people overestimate the degree to which AI is automating away knowledge work, while underestimating the degree to which it's automating everything else. That's not to say either side of the equation has been replaced by machines.
xvector | 17 hours ago
anon-3988 | 14 hours ago
solyin | 19 hours ago
Human mental effort literally commmands armies.
PowerElectronix | 19 hours ago
You can make a lever that lifts what a hundred men couldn't, but you can't make a machine that thinks and solves problems like a human can.
recursive | 4 hours ago
dist-epoch | 19 hours ago
Now something else will conquer the world from us.
myk9001 | 18 hours ago
unclebucknasty | 17 hours ago
Why are some pretending that physical effort and skill isn't itself a function of thinking? And, why are we injecting the relative importance of physical effort, when the lament is of human intelligence possibly becoming irrelevant?
jMyles | 20 hours ago
How do you know that the state of AI in 5026 won't be a far more joyous one?
FloorEgg | 19 hours ago
pastel8739 | 19 hours ago
FloorEgg | 15 hours ago
jMyles | 13 hours ago
Of course, nobody wants to be unsympathetic to the pains of change; we all feel it. But this "life plans into question" line of reasoning strikes me as perhaps the pinnacle first-world-problem.
What is so special about today's world order, where the distribution of goods is tenuously connected to "jobs", but even more so, to predation and warfare, that we want to cling to it?
Clinging to "jobs" and "careers" as organizational modalities seems even more insensitive to changing times than accepting their deprecation.
pastel8739 | 12 hours ago
krapp | 19 hours ago
jMyles | 19 hours ago
I know I feel great playing my guitar or mandolin. I like drugs and want to do more psychedelics next year. I want to get back to dating; my transition to a bluegrass career amidst software engineering jobs disappearing has been very consumptive, although very rewarding. I can probably name 2-3 other similar intentions that feel distinctly like those of a biological human.
But I don't presume that I fully understand what I am, what my consciousness is, etc. These are old questions worth pondering rather than presuming.
I do worry that AI will have trouble understanding the importance of sex, drugs, and rock 'n' roll.
krapp | 19 hours ago
olelele | 19 hours ago
jMyles | 18 hours ago
But I'm not so steadfast in my certainty about what reality is to demand that the universe halt its operations because I think that blobs of fat tissue held up by a sack of meat supported by a bunch of bones are the perfect end state.
bemtevi | 19 hours ago
2. Most people want a purpose. It's highly unlikely that a world in which machines can do everything better is a good one. It's intuitive.
skeaker | 19 hours ago
Point 2 is again going to vary a lot from person to person. There are people out there today who would prefer to never retire, going back into minimum wage customer service jobs in their elder age just because they want to, and I'd wager they would agree with you. On the other hand, there are swathes of people who can't wait to be free of responsibility, and countless more that are crushed by poverty and hunger when they can't meet it. Those people certainly care much less about "purpose," or in many cases are certain they won't find purpose in something as mundane as a job.
bemtevi | 19 hours ago
Definitely true! Which is why the future will be essentially dominated by new wars between people on the different sides of the debate.
myk9001 | 18 hours ago
Say what?
Timwi | 16 hours ago
preg_match | 15 hours ago
The world is much, much safer than the 70s, 80s, or 90s. It doesn’t seem that way because everyone is trying to convince you it’s not. But statistics don’t lie. This is the best time ever to not die or get robbed.
anhyz | 18 hours ago
For some, it's easy to forget: being at the cutting edge, being the best, doing things that have never been done, solving problems that no one else could are _not_ the only sources of purpose in life.
For example: chess is more popular now than it's ever been. People don't really seem all that bothered by the fact that they are absolutely outmatched by the machines. They do it anyway, and seem to get some sort of joy or purpose out of it.
(That being said, I can't say I'm all that happy about the way things are headed.)
NooneAtAll5 | 19 hours ago
jMyles | 19 hours ago
anhyz | 19 hours ago
jMyles | 18 hours ago
myk9001 | 18 hours ago
Knock yourself out!
unclebucknasty | 17 hours ago
pluralmonad | 16 hours ago
That does not appear to be the case.
sph | 8 hours ago
preg_match | 15 hours ago
Maybe things will be sick in 5026. I won’t be around to see it.
solyin | 20 hours ago
radish_spirit | 19 hours ago
chanux | 16 hours ago
I went through it anyway. It was alright.
whimsicalism | 19 hours ago
sph | 8 hours ago
beebmam | 19 hours ago
bakugo | 19 hours ago
dada216 | 18 hours ago
I'm a millennial, Im starting to think I need to write down for those future generations how it felt to be raised in a mostly analog world and die in a world were LLMs achieved what they just did and what they might still.
beebmam | 17 hours ago
pluralmonad | 16 hours ago
layer8 | 19 hours ago
skeaker | 19 hours ago
zdeb7 | 18 hours ago
bemtevi | 19 hours ago
trickyager | 18 hours ago
beebmam | 14 hours ago
fwlr | 19 hours ago
That’s affected me quite heavily. I think I should probably spend the day offline, actually.
ignoramous | 19 hours ago
Not in the same boat, but I'd encourage you to read this recent discussion: https://news.ycombinator.com/item?id=50094561
vibebingo | 19 hours ago
bemtevi | 19 hours ago
whimsicalism | 19 hours ago
viccis | 19 hours ago
Oh my god how do people still think this is the endgame here
Creamsicle47 | 19 hours ago
bigbadfeline | 3 hours ago
whimsicalism | 19 hours ago
usage and control of the technology is a separate question, but orthogonal to the meaning question.
teg4n_ | 19 hours ago
wavemode | 18 hours ago
airstrike | 18 hours ago
undersuit | 18 hours ago
mrbombastic | 17 hours ago
whimsicalism | 17 hours ago
radio879 | 17 hours ago
It can hold much more information about the world in its memory, its smarter, can process all of the information fast enough - I know people like to assume that all it would take is money to permanently fix poverty but its really not that easy. There are people trying, they have billions of dollars, but they can only do 'a little' (think: Bill Gates)
pluralmonad | 16 hours ago
undersuit | 15 hours ago
preg_match | 15 hours ago
While AI might not be the smartest, it is certainly the fastest. Maybe we can solve 100x more problems, who knows.
leafingensemble | 3 hours ago
If so, why does ending poverty, but creating a dystopia, result in more happiness?
This work is a form of submission, an effort to be eclipsed, to seek life in death, to fulfill the evolutionary instinct to die for a more magnificent child, and to be dazzled by the beauty of one's own self-immolation like a moth flying into a flame.
Anthropic does not care about reducing poverty. Anthropic is a technological temple, walled off from the turgid and festering mass of humanity. Its work involves preventing people from accessing its models the way the Vatican wanted to control the people's access to God; with closed weights security, guardrails, encrypted chain of thought, logging, regional bans, and moderators. Its goal is to create AGI by any means necessary.
Your argument is something along the lines of Thatcher's "there is no alternative". It is unfalsifiable.
"There is no alternative" is a lie designed to help people hide from the guilt of bad decisions.
lf88 | 19 hours ago
bigbadfeline | 3 hours ago
Yes, we do need advanced and open AI now.
> while it's not hard to imagine different ways in which a superintelligence could harm humanity.
That's why.
P.S. A heartfelt request to not use the term spooperintelligence in the future, it upsets my stomach.
leafingensemble | 17 hours ago
Without meaning there is nothing.
Modernity loathes itself. It is so devoid of meaning and so full of the fear of meaningless death, that it will try to drown itself in its own works, dazzled by its lifeless and immortal creation, the sublime of technology, the holy machine of loving grace.
The 'technique' of technology (Ellul), is to justify its own existence. Like the surveillance state, the demand for "advancement" is self-referential.
And on the pedestal these words appear: "My name is Ozymandias, king of kings: Look on my works, ye Mighty, and despair!"
jrflo | 19 hours ago
layer8 | 19 hours ago
cheevly | 18 hours ago
jrflo | 17 hours ago
bemtevi | 19 hours ago
munchausen42 | 11 hours ago
runarberg | 19 hours ago
sph | 8 hours ago
ggm | 19 hours ago
I don't personally find "as Ayn Rand said.." a compelling line in the appeal to authority stakes.
ignoramous | 19 hours ago
barry-cotter | 19 hours ago
Pure conditioned reflexes, no mental activity whatever.”
https://x.com/zetalyrae/status/2123943922095526074
slibhb | 19 hours ago
"Reason is, and ought only to be the slave of the passions, and can never pretend to any other office than to serve and obey them." Claude doesn't have passions; man does. Relax!
bemtevi | 19 hours ago
dadadad100 | 19 hours ago
airstrike | 18 hours ago
dadadad100 | 4 hours ago
doctorzook | 19 hours ago
I'll never be the next Euler or von Neumann or Einstein, and I'll confess to occasional pangs of regret that I'll never reach those peaks. But mostly, I just enjoy learning about math and science, and delight in the small contributions I can offer.
And I'm the oddball: most people couldn't care less.
There must be more to man than math.
andai | 19 hours ago
dead_hn_theory | 19 hours ago
Much more. Do critically look at what information and topics you are reading every day and ask if you are in pigeon-holed where very few subjects dominate your attention span.
p-e-w | 19 hours ago
We need a new approach to defining humanity that doesn’t involve performance at all. Not a “God of Gaps”-style argument where we try to carve out things we’re still better at, and then have to revise every few months as AIs encroach on those as well.
airstrike | 18 hours ago
What? Why? There's actually no reason to assume that AI performance in one field is directly transferable to all other fields.
Math has a verifier (Lean). Poetry does not.
senordevnyc | 17 hours ago
paulhebert | 16 hours ago
I think math and lean (and honestly coding) are some of AIs strongest areas right now.
Is an AI as good of a therapist, dancer, nurse, teacher, stable hand, chef, firefighter, or author? I definitely don’t think so
preg_match | 15 hours ago
If we had robots as good as the human body, do I think a model could dance better than me? Definitely. Could it dance better than a really good dancer? Probably.
HAL3000 | 19 hours ago
Reading that, I had a sad realization, if we keep progressing the way we are, there will be no more Einsteins and von Neumanns in humanity's future, only Claudes, Geminis, and Astras.
boplicity | 19 hours ago
paradox242 | 19 hours ago
reasonableklout | 19 hours ago
tomrod | 18 hours ago
ExciteByte | 17 hours ago
jimmygrapes | 16 hours ago
10111001 | 12 hours ago
Why? You won't know about those ideas. No one will. People like you will be pets. Probably not even educated anymore, as what would be the point of that?
objektif | 17 hours ago
paulhebert | 17 hours ago
They focused on other humans they admired. I don’t think that’s self-important.
MiiMe19 | 16 hours ago
recursive | 4 hours ago
objektif | 3 hours ago
dralley | 6 hours ago
b473a | 19 hours ago
Is there more to math than man? Does it serve a purpose other than to stand as a frontier of knowledge for us to explore and test ourselves against? Head says yes, heart says no.
Is saying "this proof is less impressive because we used an AI to find it" like saying "the moon landing is less impressive because we used a rocket to get there"?
zdeb7 | 18 hours ago
You could compare this to doing a 3D NES game in ASM and doing the same game with a vulkan C# game engine. Obviously with the ASM one you would have to create the underlying engine, so it's more impressive, because you do more, and you demonstrate human development and your intelligence.
StilesCrisis | 17 hours ago
citizenpaul | 19 hours ago
C'est la Vie is something that only ever seems to come from those whom are already set to those caught in the meat grinder. Though I don't know your situation.
aquariusDue | 18 hours ago
For young(er) people out there (Hi, depending on the reader) you don't have to compare yourself to the "greats", you can look at exceptional peers in your field that achieved more at a young age. Can't think of one? Check the age of the person behind Babel (the JavaScript transpiler).
But generally that's not a good way to approach life and notions of self worth. It's just an avenue you can peer at for a dose of melancholy.
MiiMe19 | 16 hours ago
fHr | 19 hours ago
anal_reactor | 19 hours ago
This is exactly my attitude and I see nothing wrong with it. I feel like when the dust settles and AGI actually shows up, it's going to be the collective child of humanity. I mean, current state of AI aside, I think it's a very buffoon thing to think humans will always remain the most intelligent creature in the world.
I understand that I'll never win Olympics, I'll never get a Nobel prize. I won't even get qualified for my local Tekken tournament. But that's okay. I'm just doing my thing and if my thing is useless in the grand scheme of things because nobody cares about me doing tricks when flipping pancakes, that's okay. That's the reality of 99,99% people so "AI is better than top scientists" isn't exactly a life-changer.
nixosbestos | 19 hours ago
MiiMe19 | 11 hours ago
nixosbestos | 4 hours ago
hatthew | 19 hours ago
My most relevant point is that solving the Riemann hypothesis is not like a large industrial project such as a canal. It's a work of art. Imagine a team of a few master craftspeople working for years on a large, incredibly detailed and intricate sculpture. Now imagine they are close to finished, but someone comes along and 3D prints a finished version of their sculpture in a week. That's what is happening with RH right now.
Another point is that humans are generally weak but intelligent. The industrial revolution replaced our weakness with machines. It made sense to be glad that machines could shore up our weakness. But our intelligence is something we should be proud of.
Another point is that, very roughly, there are two types of jobs: physical and mental. The industrial revolution replaced our physical jobs and left society with the capacity to do many more mental jobs. But the AI revolution is replacing those, and there's no obvious category of jobs left for us.
The final point I'll mention here is that the industrial revolution wasn't all sunshine and roses. There used to be industries of artisans making high quality products. The industrial revolution didn't just replace unglamorous jobs of digging holes and the like, it replaced relatively "virtuous" jobs (such as the village cobbler) with mass-produced goods of less quality. A high quality and reliable custom-fit pair of shoes is less accessible today than centuries ago (measured by factors like number of days wages, ease of finding someone to make them, and trust that the result will be good).
I think the AI revolution has the potential to be a good thing if we use it to drive a utopia where "artist" is a reliable career, but my bet right now is that the future is heading more dystopian than utopian.
Seattle3503 | 19 hours ago
The remaining jobs are social and relational, and thats scaring a lot of us nerds.
hatthew | 19 hours ago
preg_match | 15 hours ago
It’s the same reason that AI partners and stuff aren’t a real concern. Can AI be a good partner and do some people fall in love with AI? Sure, but at its core it’s not human and therefore has little human value in that regard. You can fall in love with AI so long as you can convince yourself it’s not software. Some people are good at that delusion, most are not.
At the end of the day AI can say all the right things, do all the right things, whatever. But it doesn’t have skin and a fleshy brain within it, so I don’t relate to it. It’s not one of us and never will be. For some jobs, it needs to be one of us for us to care. The trick is we’ve spent the past few thousand years deconstructing the human touch and relationships, so now that’s coming home to roost.
hackingonempty | 16 hours ago
Before the industrial revolution, people that weren't rich didn't have high quality products made by artisans. What they had was low quality and inexpensive. The industrial revolution greatly increased both the quality and quantity of products most people could afford.
hatthew | 12 hours ago
Regarding quality back in the day, my understanding is that e.g. a farmhand could buy a solid pair of work boots for around a week's wages. These boots would be custom-fit, probably full grain leather, and reliable because everybody knows everybody by name and the cobbler probably has a reputation to maintain.
These days, equivalently custom boots look like they're roughly a day's to a week's wages, but the lower you go in that range the more likely you are to get scammed into buying dropshipped garbage.
theodorewiles | 19 hours ago
but there is still value in your unique distinct voice
AI is a projection of your capabilities
it is a projection of your unique voice
many people do not have their own unique distinct voice
the best antidote is the well-rounded liberal arts
cultivate your identity
cyanydeez | 19 hours ago
ecocentrik | 19 hours ago
dada216 | 18 hours ago
cliche | 19 hours ago
There’s a scene in an obscure movie I can’t quite remember. Antonio Banderas is tracking down a rogue robot. He finds it and other robots self repairing and self improving. There’s some drama and fighting. In the end he escapes it’s left open with the ominous realisation that it’s the moment that the machines start to supersede us.
Not a memorable movie on the whole but I always remember that ending. The post give me that same vibe.
I really hope we’re not close to that moment yet but it’s becoming obvious we’re being surpassed in so many areas at an accelerating rate. I no longer think about what things might look like in 10000 years. What do things look like in 10. It’s both an exciting and terrifying time to be alive.
senordevnyc | 17 hours ago
jhanschoo | 19 hours ago
Historically and presently, most humans do not have the privilege resources to "explore, discover, invent, succeed, become more, struggle, win, become divine" in a way that was not already trodden by other people. We/they have to contend with "less".
perching_aix | 19 hours ago
I'm sure the author thinks this was an absolute slapper of a reference to prove their point, but the point proven doesn't exactly belong.
The whole passage seems to be an appeal to reject: https://en.wikipedia.org/wiki/Clay_Shirky#Shirky_principle
on the grounds of preferring to yield to our inherent bias of continued existence instead.
Meanwhile, our servants not being able to live on our behalves is completely independent of this. It is also robust against being considered a betrayal or a surrender, while the above really quite isn't.
Contrasting this with the identification of Alpöge's writing as words of final relief could have painted a much richer picture, than a literature and art references riddled myopic narrative concluding like every other like it.
Not that it has any need to arrive somewhere different, but still. It insists upon the illusion of sophistication it suggests should prevail, because it is at least human. Hardy any more actually satisfactory than the alternative despised. Quite the contrary, if anything.
But hey, if we insist on references, here, from the 1979 movie The Witness, satirizing the soviet occupation era politics in Hungary:
> - What is this?
> - An orange.
> - An orange?!
> - The new, Hungarian orange. It's a bit more yellow and a bit more sour, but it is ours.
> *proceeds to continue munching on the lemon slice with a pained, approving smile*
https://www.youtube.com/watch?v=K1ghpXiBkCQ
Kuyawa | 18 hours ago
Until the day AI starts directing itself creating knowledge that doesn't exist yet with no one at the wheel. Anxiously waiting that day...
bm3719 | 18 hours ago
At some point, we're left adding nothing more than the simple impetus to act, the decision to spend tokens on the task. How long before even that's no longer needed? What will the "it's just a tool" people say then?
jaredklewis | 18 hours ago
Parent is not saying that the human's contribution is greater than that of the LLMs, just that at least at present, there is apparently still some human input.
bm3719 | 18 hours ago
On your point, we don't exactly know if that's true or not. Perhaps it would've burned more tokens, but if Claude did the heavy lifting here, then who's to say it couldn't have done all of it with just very persistent neophyte encouragement (not saying this person did that, just speculating on it being technically possible)? Large applications are being vibed right now this way, and yeah, they're often brittle messes, but if their job is to accomplish some singular goal, Claude can test for that and ensure the desired functionality. I encounter this analogous situation in my work more often than I'd like.
jaredklewis | 37 minutes ago
I guess it is not impossible, but it seems like a very big coincidence then that Alpöge was the prompter that finally cracked it. If all it took was a prompt like "Prove the Riemann hypothesis" plus a loop to re-prompt with "Interesting, keep working on that" or whatever, surely one of the swarms of OpenAI agents working to solve open problems would have just cracked it on its own.
unclebucknasty | 17 hours ago
Actually, parent said "AI is just a tool, like a calculator", which certainly does imply that the human's contribution is greater than that of the LLMs.
jaredklewis | 34 minutes ago
Does a crane or the human crane operator contribute "more" to the construction of a building? To me, it seems like an inherently flawed and unanswerable question. Both are providing something the other can't provide and there is no obvious way to compare their distinct contributions.
bawolff | 18 hours ago
However, as it stands, it seems like we aren't there yet. It seems like it takes specialized skill to get the AI to solve these types of problems. Not just anyone, no matter how well funded, can do it.
That said, the speed of progress here is incredible. Maybe we will reach that day soon. I just dont think we are quite there yet.
skew-aberration | 18 hours ago
bawolff | 16 hours ago
tomrod | 18 hours ago
unclebucknasty | 17 hours ago
tomrod | 4 hours ago
dumbmrblah | 18 hours ago
[OP] _dain_ | 17 hours ago
In a year's time, the Riemann Hypothesis will be just another benchmark for Sonnet-tier models. They will one-shot it.
dlubarov | 16 hours ago
If we're using LLMs (for some reason) to explore a chess tree, we'd probably just invoke cheap LLMs many times, but for theorem proving, it can make sense to combine the two.
Still, if the cost difference keeps growing, at some point a more brute force approach with exclusively cheap LLMs could outperform.
bigyabai | 16 hours ago
Of course it will be, because like 99% of benchmarks the answer will be a part of the training material.
[OP] _dain_ | 15 hours ago
bigyabai | 15 hours ago
hi_hi | 17 hours ago
It’s like a binary Harry Potter sorting hat! I see this at work too. We all have access to the same tools, but a core group are finding uses for it far beyond the expected coding support use cases.
testtestmetest | 16 hours ago
They then walk about the world and wander throughout online spaces telling everyone who's actually trying to go deep that they're wasting their time and they should just rent competency from a tech company and there's no difference between understanding and prompting as far as outcomes (when they dont even know what the outcome is supposed to look like).
I also agree there are some people that genuinely have super powers, and dont use it to just offshore their thinking. But I can genuinely count those people on one hand, and they were all superhuman devs before AI so I dont think it has much to do with llms.
These people are suddenly going to realize that being the meat proxy isnt special, and you wont be paid much for it (if at all), I'd rather have 3 of those guys that were super contributors before and after AI than 20 meat proxies on my team.
jmathai | 16 hours ago
I was going to say something about how much of a leap it is in capability but then realized it probably is similar to the leap introduced by a calculator.
reasonableklout | 14 hours ago
svnt | 14 hours ago
Only once it had done the work, and the non-mathematician had had independent agents re-derive it a few times, did they bring in mathematicians to dig into it.
Literally anyone could have done this. Vouch for it with any credibility, well that is only for people who can employ mathematicians.
krackers | 13 hours ago
svnt | 13 hours ago
krackers | 13 hours ago
karmakurtisaani | 12 hours ago
xjm | 11 hours ago
(Levent Alpöge is definitely a mathematician, the interaction with claude was long with expert steps, there were no other mathematicians needed to vouch, see the announcement at https://alpo.ge/remark.pdf )
piva00 | 9 hours ago
maplethorpe | 8 hours ago
I agree. I honestly don't think Levent should have even mentioned Claude. He should have just said "I solved the Riemann Hypothesis". This way we would not have all this bickering.
ks2048 | 18 hours ago
(Individualist man and collectivist man vs. AI).
Problem is that Randian men control the AI.
cs_throwaway | 18 hours ago
I did not not think this was the case reading https://alpo.ge/remark.pdf
It reads like is is delighted to deeply understand and explain what he helped uncover.
Some mathematicians are completely unwilling to understand formal proofs in Lean. This isn't new: formal logic was deemed boring by mathematicians decades ago. It has been kept alive in computer science (databases and programming languages esp).
BugDeveloper | 18 hours ago
Yes, it is, because humans build Claude. Not you yourself, but on the other side, Ayn Rand also didn't herself contribute anything at all to the Apollo mission, yet felt like she was somehow part of it.
zdeb7 | 18 hours ago
I'm not sure that I will ever find it again.
asah | 17 hours ago
Is your purpose isn't to dig ditches for another man to fill ?
When we discover and fix bugs and CVEs, that's what we're doing. It needs to be one, but humans don't have to do it.
reducesuffering | 16 hours ago
Given that prescience I would highly weight his other opinions about where AGI is headed.
https://x.com/ilyasut/status/1710462485411561808?lang=en
bryan0 | 18 hours ago
I think the misconception here is that this is a problem of will. I see other people in comments upset about people “rooting” against humanity. This is not a game. I can want to fly by flapping my wings but alas the laws of physics forbid it. We are building machines which can accomplish things we cannot, not because we don’t want to accomplish them but because our minds and bodies are limited by reality.
jongjong | 17 hours ago
tw04 | 15 hours ago
Knowing the slop AI can and does put out, it seems the folks who treat it like a person stop treating it like a fallible tool that should be spot checked continuously.
whamlastxmas | 5 hours ago
Next we're going to see "they cured cancer :/"
classified | 2 hours ago
First quoting Ayn Rand, then this? Am I the only one who feels uneasy about this?