I feel like this kind of "result dump" just cheapens mathematics. How about having a little respect for those whose work this builds on, and current mathematicians some of who may have spent years working on these problems.
Rather than sitting on these results until they had enough for a "shock and awe" 10-result dump, how about releasing these results individually as they were made/verified, as well as the failures (equally valuable to assess the current capabilities of LLMs), and try to make some analysis of HOW these breakthrough results were made. What were the prompts for each of these, how much guidance was there from the mathematicians employed by OpenAI, and most importantly how did the model arrive at these results ... what lines of reasoning resulted it in exploring ideas that humans had previously not explored?
I know it’s tough, but I am not a fan of elitism. Mathematics is no different from all other branches which themselves are just intellectual labor which is again just a special type of labor. There is nothing magical about it and if computers can trivialize it, so be it.
Where were all the mathematicians and academics in general when “regular joe” was automated? Now it’s hitting close to home and their foreheads are starting to get sweaty. I’d say let them. Tough luck. Make mathematics as “cheap” as possible. Nobody owes them any favors.
Let’s commoditize “being smart” and let go of arbitrary divisions between us.
Yes, but in practice that doesn’t happen. Again, I did not hear a loud protest against automation of all types of labor from the intelligentsia in the past.
In practice automation is great, as long as it doesn’t hit “the ones that matter” (a label which they themselves assign). I find it very hard to not imagine the smallest, tiniest violin playing the saddest song for them.
Again, “respect for mathematicians” and “their work”.. please. Just produce results. That’s all that ever mattered and let’s not change the rules of the game just because they don’t suit you anymore.
Protests against automation, or for the restructuring of society to accommodate automation without causing suffering of laborers, etc, is literally one of the largest and most extensively discussed topics in modern history.
This is what the Ludites were trying to accomplish with an (armed) uprising.
Go back 100 years and 15 percent of the population was "Farmer", and 30 percent were in some sort of "domestic" work.
Most modern jobs are a product of the fact that technology continues to advance. And by all measures it does not look like "ai" is going to change that.
That is the point: Make "X can be done by AI" common discourse and watch industries "being disrupted". At least with open weights we'd be able to verify the claims.
You can actually do all of that meta analysis if you have the conversation that led to the solution. This is as easy as just having people release logs of their conversations, and then you can load it into another LLM as context to ask a bunch of questions about it.
I hope this will be normal practice eventually. “Show your work” is trivial if you use AI to solve a problem.
It is so incredibly liberating to see the intellectual elite struggling with was already reality for 99% of the rest of us: your understanding is not required, it might in fact be detrimental as it amounts to a handbrake on progress.
Understanding never was the goal, results were. We will be getting boatloads of those. What does your understanding get us? You get a fancy house out of it and you might be intellectually stimulated by it, sure, but I hope you can see how that does not constitute a valid need for the rest of society to labor just to support your class and its lifestyle.
Whoa, how quickly have we forgotten how the AI is trained?
> You get a fancy house […] society to labor just to support your class and its lifestyle.
Wait, how much money do you think math teachers make? Do you know how much money AI engineers make in comparison? Why the animus and insecurity toward education? You are making math teachers sound far more rich and powerful than they actually are.
One would think something like the study of the square root of negative 1 would have little utility beyond a hand-wavy appeal to understanding, but it turns out to have major practical applications ranging from electrical engineering to 3d rotations.
I tend to believe we're still likely at a very primitive level of technology. It's difficult to imagine this primarily because of unknown unknowns. For instance at the end of the 19th century it was believed that science had basically been finished, and all that remained was working things out to ever higher levels of precision to wrap up a few curiosities.
Then along come relativity, quantum mechanics, and more - all completely revolutionizing science, and upending centuries of false beliefs. I think we're far more likely to be at the end of the 19th century again, rather than at the end of science, or anywhere near such. And if that's the case, then it may well be that seemingly academic discoveries of today, perhaps ultimately achieved by an LLM, are what help unlock the next great revolution.
Yes, but what will truly advance science is new discoveries that require curiosity and creative thinking, and LLMs are not the technology for this.
As impressive as these mathematical results are, all they represent is known mathematical techniques being combined in ways that humans had not explored. The AI is not inventing areas of math, or even creating it's own conjectures. There will be more of these to follow, but eventually the pace will slow as the space of what can be done with known mathematics is explored.
What is needed for creativity, for new scientific discovery, is more than an auto-regressive LLM assembling someone else's Lego blocks in novel ways - it will require an AI not just built to predict but one that is also built to expand the frontier of what it knows. Prediction failure that results in curiosity and exploration, not hallucination.
I shouldn't be surprised how prevalent this viewpoint is, given the entrenched anti-intellectualism these days in the US and across the world, but it's disturbing to see it so widespread on HN of all places.
You are of course free to wield this new technology any way you choose (to the extent its owners and your wallet allow you to, for now at least), and if you can enrich yourself by doing so power to you. I won't attempt to dissuade you from your apparent hatred of those who have spent their lives trying to expand the boundaries of human understanding. I'll just state my equally emotion-driven opposing view: It always was and always will be about the understanding. "Results" and "progress" without sufficient thought are the reason our enormous modern wealth is accompanied by so much human misery.
I don't relish the prospect of living in your predicted world of multiplying technology untempered by human thought, if that's what happens to emerge. I don't think it will be a good place for anyone but a handful of the very richest.
Good luck adding yourself to their number.
Once you have experienced first hand the level of disdain the so called intellectuels have towards the regular joe it is hard to feel any sympathy for them. If I can be replaced with cheap foreign labour you too should just deal with it. You are not entitled to your well paying oil drilling job, you know, you should just move on and take another job or two and some nightly backbreaking gigs here and there if math doesn't feed your family any more. Surely you don't want to be lazier than immigrants and can adjust to their frugal standards of living...
Exceptional mathematicians are not going to be jobless, just like master level craftsmen still do kinda well. Tough luck if your abilities are subpar to achieve excellence though, that is just the rat race to the bottom we have created and finally the so called intellectuels are also getting to fully taste it themselves. Realizing you were never that special and your job can be either fully automated away or reduced to a wage that barely pays the bills is a hard pill to swallow. But progress is inevitable as they have been saying.
I struggle to identify with the apparent glee you feel at berating the person you imagine yourself talking to. I guess we all have our own idiosyncratic forms of fantasy, but this one doesn't seem very healthy.
Same thing on coding - business doesn’t care one bit about your clean architecture or clever design, so long as the AI can change the implementation fast enough to ship whatever product comes up with.
People are having a rough time figuring out that good design and fundamentals no longer align with the business side of things.
> People are having a rough time figuring out that good design and fundamentals no longer align with the business side of things.
That's not true and wasn't true in the sense that businesses don't care directly but they are indirectly. Are you saying business doesn't care if things go down or it doesn't work?
At the end of the day people don't care until they do and then they blame you.
> business doesn’t care one bit about your clean architecture or clever design
Perhaps not in general, but this is nothing new or related to AI. For one thing software is a strangely private craft - your boss is likely only aware of what you are producing, not the craft or quality (or not) that went into it - the only ones aware of that would be co-workers who are looking at your source code, and perhaps choosing to use it or not. Secondly, the sad truth in business, not just for software developers, is that there is a huge amount of truth in "it's not what you know, it's who you know".
> People are having a rough time figuring out that good design and fundamentals no longer align with the business side of things.
The business may be unaware of what your code looks like, or the expertise that went into it to avoid bad outcomes (immediate or future), but they surely DO care when poor quality software leads to production outages, and customer complaints, and their boss yelling at them as a result. The results of poor quality software will manifest in many ways, not just the poor reliability and performance of your latest release, but also the future outcomes of attempts to build on top of it.
I think it's extremely early to assess the impact of AI software quality and the ways it impacts businesses in these ways, such as production outages, that they do care about. There have certainly been well publicized cases such as AWS outages blamed on AI, but although it seems like an eternity it's only been ~6 months since AI really became usefully competent and capable of "vibe coding", so I think we're only just beginning to see the consequences.
Companies and bosses that have enthusiastically adopted AI are unlikely to point to AI and themselves if/when it proves to be detrimental (AWS were quick to deny it, despite the insider leaks to the contrary), so I'd not be surprised if it takes years rather than months for there to be much clarity on this.
FWIW, I think there is a considerable difference between math, which is what we're really discussing here, and software development, regarding how quickly it will or has fallen to AI. Math is really uniquely suited to LLMs and reinforcement learning of domain-specific reasoning since mathematical papers basically consist of nothing but reasoning (training data!), and it is an entirely theoretical field that essentially lives in the world of language (incl. mathematical notation) without any external grounding. Contrast to software development where what is available to train on is mostly the end result (code) rather than the reasoning that got you there, and there is also a large "experimental" aspect - what happens in the real world when you run the produced artifact.
Your reply isn't a response to the comment, but rather a general expression of antipathy towards mathematicians, who are apparently fatcats who have fancy lifestyles.
That’s pure straw man. There isn’t anyone against the working class in this thread, and there is anti-intellectualism in this thread. I don’t doubt there are some academics who are jerks, but you’re incorrectly assuming something about all academics (not to mention all working class) that simply isn’t true.
Go back and re-read. Did I say anything bad about AI?
What I'm talking about is the way OpenAI have handled this. They are in the business of artificial intelligence, striving to create human-level, or even super-human intelligence, and telling us what a wonderful future this will create.
Shouldn't they be celebrating intelligence, especially human intelligence which is what they have bottled here?! LLMs are themselves minimally intelligent - they are primarily regurgitating human intelligence. Without their human training data, these systems would be nothing.
Let's see OpenAI build "GPTZero", an actually intelligent system, that learns for itself, learns language and mathematics from nothing like a baby, then I feel they'd have the right to minimize and cheapen the achievement if they chose to (unlikely). As it is, all they have is a predictor, utterly dependent on HUMAN intelligence, and it would be more becoming of them to acknowledge this by paying a bit more deference and respect to the people who enabled them to build what they have.
I’m looking forward to the days where AI would help in tackling the problems in biology. Especially, on creating new drugs, enzymes and understanding the genetic diseases. An absolutely interesting time to live.
Right now it makes your account completely useless if you work on something completely harmless related to viruses. Blocks, downgrades, throttling of answers "to check for safety".
You need to be at a select few institutions to avoid these safety features, which I'm not. University needs to sign some stuff but I that's outside my control.
Does anyone else have trouble telling how much of this news (along with the 'AI escaping and hacking' stories) is genuine, vs how much is just AI firms overstating their capabilities due to strong commercial incentives?
These math results appear genuine, and impressive, and it seems they've already been at least provisionally verified.
Of course there is still a massive marketing aspect to this, with the AI companies wanting to you assume that because their product is world-class at math, a capability that is useless to 99.99% of their potential customers, that it will be equally useful in areas that you actually care about, such as managing your vending machine, perhaps :-)
It's hard to know how to interpret these hacking confessions/boasts and what the reality is behind them. I get the impression (with low confidence) that they really did not anticipate or orchestrate these attacks, but it also seems they did little to prevent them, and seem to be happy that they occurred (as you note, a chance to suggest how powerful they are).
I do think that LLMs/agents can be highly capable, and dangerous as hackers, especially if you deliberately train them to be as was the case with Mythos.
There is current news of US water treatment plants being hacked, apparently by Iranian state actors, and we should be glad it was just water treatment plants and not some more critical piece of infrastructure (perhaps power generation or transport, etc). I hate to think what a malicious actor could do with today's SOTA AI if they really wanted to do something destructive, not just send a warning shot.
We hacked a company and blame the tool! Somehow it is not negligence, but cool!
We hacked 3 companies and tripple blame the tool! We are even cooler!
We hacked companies amd therefore other peoples models need to be restricted! See, us accidentally pointing a hacking tool on others proove we are the only ones that can be trusted with the tool!
Unlike the hacking one, this would be impossible to bullshit as long as the proofs are released. They can be verified independently, and the alternative is that they solved 10 major open mathematical questions without the AI, which seems less likely.
The LLM marketing loop is getting awfully long in the tooth. I saw a meme on Twitter the other day that showed a circular state diagram with something like:
> “GPT solved a math problem” -> “Claude solved a math problem” -> “GPT escaped the sandbox” -> “Claude escaped the sandbox” -> …
I believe this is what we wanted computers to help us solve along with other prior hard problems prior to computers. This should be viewed as a good thing even if Anthropic, OpenAI, etc benefit just like IBM benefitted from mainframes.
Indeed. I specifically recall my logic lecturers in a compsci course expressing the view the one purpose of computers is to work towards enabling automatic proofs, partly because they saw this as the universal problem in compsci (esp compilers) as well as massive potential for scientific progress generally.
I feel like this kind of "result dump" just cheapens mathematics. How about having a little respect for those whose work this builds on, and current mathematicians some of who may have spent years working on these problems.
Rather than sitting on these results until they had enough for a "shock and awe" 10-result dump, how about releasing these results individually as they were made/verified, as well as the failures (equally valuable to assess the current capabilities of LLMs), and try to make some analysis of HOW these breakthrough results were made. What were the prompts for each of these, how much guidance was there from the mathematicians employed by OpenAI, and most importantly how did the model arrive at these results ... what lines of reasoning resulted it in exploring ideas that humans had previously not explored?
I know it’s tough, but I am not a fan of elitism. Mathematics is no different from all other branches which themselves are just intellectual labor which is again just a special type of labor. There is nothing magical about it and if computers can trivialize it, so be it.
Where were all the mathematicians and academics in general when “regular joe” was automated? Now it’s hitting close to home and their foreheads are starting to get sweaty. I’d say let them. Tough luck. Make mathematics as “cheap” as possible. Nobody owes them any favors.
Let’s commoditize “being smart” and let go of arbitrary divisions between us.
I'd prefer the reverse approach. Let's value the "regular Joe" instead of cheapening everyone.
If everybody is "valued", nobody is.
The concept of value implies exclusivity by design.
Yes, but in practice that doesn’t happen. Again, I did not hear a loud protest against automation of all types of labor from the intelligentsia in the past.
In practice automation is great, as long as it doesn’t hit “the ones that matter” (a label which they themselves assign). I find it very hard to not imagine the smallest, tiniest violin playing the saddest song for them.
Again, “respect for mathematicians” and “their work”.. please. Just produce results. That’s all that ever mattered and let’s not change the rules of the game just because they don’t suit you anymore.
Protests against automation, or for the restructuring of society to accommodate automation without causing suffering of laborers, etc, is literally one of the largest and most extensively discussed topics in modern history.
This is what the Ludites were trying to accomplish with an (armed) uprising.
Go back 100 years and 15 percent of the population was "Farmer", and 30 percent were in some sort of "domestic" work.
Most modern jobs are a product of the fact that technology continues to advance. And by all measures it does not look like "ai" is going to change that.
that's an incredibly narrow view of what's happening.
Not that I disagree, but I think this is how many people feel about AI output in fields they care about.
It's also a funny historical mirror to an earlier phase of math proof culture: in a previous era, cryptic result dumps were quite common.
That is the point: Make "X can be done by AI" common discourse and watch industries "being disrupted". At least with open weights we'd be able to verify the claims.
You can actually do all of that meta analysis if you have the conversation that led to the solution. This is as easy as just having people release logs of their conversations, and then you can load it into another LLM as context to ask a bunch of questions about it.
I hope this will be normal practice eventually. “Show your work” is trivial if you use AI to solve a problem.
It is so incredibly liberating to see the intellectual elite struggling with was already reality for 99% of the rest of us: your understanding is not required, it might in fact be detrimental as it amounts to a handbrake on progress.
Understanding never was the goal, results were. We will be getting boatloads of those. What does your understanding get us? You get a fancy house out of it and you might be intellectually stimulated by it, sure, but I hope you can see how that does not constitute a valid need for the rest of society to labor just to support your class and its lifestyle.
> your understanding is not required
Whoa, how quickly have we forgotten how the AI is trained?
> You get a fancy house […] society to labor just to support your class and its lifestyle.
Wait, how much money do you think math teachers make? Do you know how much money AI engineers make in comparison? Why the animus and insecurity toward education? You are making math teachers sound far more rich and powerful than they actually are.
> Understanding never was the goal
Hard disagree
what are even "results" in mathematics if not understanding? for instance, what's the utility of the counterexamples built by openAI?
It seems you're fighting an hallucinated enemy.
One would think something like the study of the square root of negative 1 would have little utility beyond a hand-wavy appeal to understanding, but it turns out to have major practical applications ranging from electrical engineering to 3d rotations.
I tend to believe we're still likely at a very primitive level of technology. It's difficult to imagine this primarily because of unknown unknowns. For instance at the end of the 19th century it was believed that science had basically been finished, and all that remained was working things out to ever higher levels of precision to wrap up a few curiosities.
Then along come relativity, quantum mechanics, and more - all completely revolutionizing science, and upending centuries of false beliefs. I think we're far more likely to be at the end of the 19th century again, rather than at the end of science, or anywhere near such. And if that's the case, then it may well be that seemingly academic discoveries of today, perhaps ultimately achieved by an LLM, are what help unlock the next great revolution.
Yes, but what will truly advance science is new discoveries that require curiosity and creative thinking, and LLMs are not the technology for this.
As impressive as these mathematical results are, all they represent is known mathematical techniques being combined in ways that humans had not explored. The AI is not inventing areas of math, or even creating it's own conjectures. There will be more of these to follow, but eventually the pace will slow as the space of what can be done with known mathematics is explored.
What is needed for creativity, for new scientific discovery, is more than an auto-regressive LLM assembling someone else's Lego blocks in novel ways - it will require an AI not just built to predict but one that is also built to expand the frontier of what it knows. Prediction failure that results in curiosity and exploration, not hallucination.
We'll get there eventually.
I shouldn't be surprised how prevalent this viewpoint is, given the entrenched anti-intellectualism these days in the US and across the world, but it's disturbing to see it so widespread on HN of all places.
You are of course free to wield this new technology any way you choose (to the extent its owners and your wallet allow you to, for now at least), and if you can enrich yourself by doing so power to you. I won't attempt to dissuade you from your apparent hatred of those who have spent their lives trying to expand the boundaries of human understanding. I'll just state my equally emotion-driven opposing view: It always was and always will be about the understanding. "Results" and "progress" without sufficient thought are the reason our enormous modern wealth is accompanied by so much human misery.
I don't relish the prospect of living in your predicted world of multiplying technology untempered by human thought, if that's what happens to emerge. I don't think it will be a good place for anyone but a handful of the very richest. Good luck adding yourself to their number.
Once you have experienced first hand the level of disdain the so called intellectuels have towards the regular joe it is hard to feel any sympathy for them. If I can be replaced with cheap foreign labour you too should just deal with it. You are not entitled to your well paying oil drilling job, you know, you should just move on and take another job or two and some nightly backbreaking gigs here and there if math doesn't feed your family any more. Surely you don't want to be lazier than immigrants and can adjust to their frugal standards of living...
Exceptional mathematicians are not going to be jobless, just like master level craftsmen still do kinda well. Tough luck if your abilities are subpar to achieve excellence though, that is just the rat race to the bottom we have created and finally the so called intellectuels are also getting to fully taste it themselves. Realizing you were never that special and your job can be either fully automated away or reduced to a wage that barely pays the bills is a hard pill to swallow. But progress is inevitable as they have been saying.
I will keep studying math purely for fun.
I struggle to identify with the apparent glee you feel at berating the person you imagine yourself talking to. I guess we all have our own idiosyncratic forms of fantasy, but this one doesn't seem very healthy.
Same thing on coding - business doesn’t care one bit about your clean architecture or clever design, so long as the AI can change the implementation fast enough to ship whatever product comes up with.
People are having a rough time figuring out that good design and fundamentals no longer align with the business side of things.
> People are having a rough time figuring out that good design and fundamentals no longer align with the business side of things.
That's not true and wasn't true in the sense that businesses don't care directly but they are indirectly. Are you saying business doesn't care if things go down or it doesn't work?
At the end of the day people don't care until they do and then they blame you.
> business doesn’t care one bit about your clean architecture or clever design
Perhaps not in general, but this is nothing new or related to AI. For one thing software is a strangely private craft - your boss is likely only aware of what you are producing, not the craft or quality (or not) that went into it - the only ones aware of that would be co-workers who are looking at your source code, and perhaps choosing to use it or not. Secondly, the sad truth in business, not just for software developers, is that there is a huge amount of truth in "it's not what you know, it's who you know".
> People are having a rough time figuring out that good design and fundamentals no longer align with the business side of things.
The business may be unaware of what your code looks like, or the expertise that went into it to avoid bad outcomes (immediate or future), but they surely DO care when poor quality software leads to production outages, and customer complaints, and their boss yelling at them as a result. The results of poor quality software will manifest in many ways, not just the poor reliability and performance of your latest release, but also the future outcomes of attempts to build on top of it.
I think it's extremely early to assess the impact of AI software quality and the ways it impacts businesses in these ways, such as production outages, that they do care about. There have certainly been well publicized cases such as AWS outages blamed on AI, but although it seems like an eternity it's only been ~6 months since AI really became usefully competent and capable of "vibe coding", so I think we're only just beginning to see the consequences.
Companies and bosses that have enthusiastically adopted AI are unlikely to point to AI and themselves if/when it proves to be detrimental (AWS were quick to deny it, despite the insider leaks to the contrary), so I'd not be surprised if it takes years rather than months for there to be much clarity on this.
FWIW, I think there is a considerable difference between math, which is what we're really discussing here, and software development, regarding how quickly it will or has fallen to AI. Math is really uniquely suited to LLMs and reinforcement learning of domain-specific reasoning since mathematical papers basically consist of nothing but reasoning (training data!), and it is an entirely theoretical field that essentially lives in the world of language (incl. mathematical notation) without any external grounding. Contrast to software development where what is available to train on is mostly the end result (code) rather than the reasoning that got you there, and there is also a large "experimental" aspect - what happens in the real world when you run the produced artifact.
Your reply isn't a response to the comment, but rather a general expression of antipathy towards mathematicians, who are apparently fatcats who have fancy lifestyles.
OpenAI sat on the results so they could drop 10 at a time. It's rumored that they are sitting more results: https://mathoverflow.net/questions/513818/a-serious-challeng...
So not only are we not going to get boatloads of results, we're only get as many results as necessary for OpenAI to market their models.
Even beyond the malady of feeling liberated by others struggling, all you’ve done is express poisonous anti-intellectualism.
Anti-intellectualism is not worse than being against the actual working class (those stupid rednecks who don't even know what is best for them).
That’s pure straw man. There isn’t anyone against the working class in this thread, and there is anti-intellectualism in this thread. I don’t doubt there are some academics who are jerks, but you’re incorrectly assuming something about all academics (not to mention all working class) that simply isn’t true.
Anti-AI rethoric is getting ridiculous.
Go back and re-read. Did I say anything bad about AI?
What I'm talking about is the way OpenAI have handled this. They are in the business of artificial intelligence, striving to create human-level, or even super-human intelligence, and telling us what a wonderful future this will create.
Shouldn't they be celebrating intelligence, especially human intelligence which is what they have bottled here?! LLMs are themselves minimally intelligent - they are primarily regurgitating human intelligence. Without their human training data, these systems would be nothing.
Let's see OpenAI build "GPTZero", an actually intelligent system, that learns for itself, learns language and mathematics from nothing like a baby, then I feel they'd have the right to minimize and cheapen the achievement if they chose to (unlikely). As it is, all they have is a predictor, utterly dependent on HUMAN intelligence, and it would be more becoming of them to acknowledge this by paying a bit more deference and respect to the people who enabled them to build what they have.
FWIW I am not a mathematician.
I’m looking forward to the days where AI would help in tackling the problems in biology. Especially, on creating new drugs, enzymes and understanding the genetic diseases. An absolutely interesting time to live.
Right now it makes your account completely useless if you work on something completely harmless related to viruses. Blocks, downgrades, throttling of answers "to check for safety".
You need to be at a select few institutions to avoid these safety features, which I'm not. University needs to sign some stuff but I that's outside my control.
Does anyone else have trouble telling how much of this news (along with the 'AI escaping and hacking' stories) is genuine, vs how much is just AI firms overstating their capabilities due to strong commercial incentives?
These math results appear genuine, and impressive, and it seems they've already been at least provisionally verified.
Of course there is still a massive marketing aspect to this, with the AI companies wanting to you assume that because their product is world-class at math, a capability that is useless to 99.99% of their potential customers, that it will be equally useful in areas that you actually care about, such as managing your vending machine, perhaps :-)
It's hard to know how to interpret these hacking confessions/boasts and what the reality is behind them. I get the impression (with low confidence) that they really did not anticipate or orchestrate these attacks, but it also seems they did little to prevent them, and seem to be happy that they occurred (as you note, a chance to suggest how powerful they are).
I do think that LLMs/agents can be highly capable, and dangerous as hackers, especially if you deliberately train them to be as was the case with Mythos.
There is current news of US water treatment plants being hacked, apparently by Iranian state actors, and we should be glad it was just water treatment plants and not some more critical piece of infrastructure (perhaps power generation or transport, etc). I hate to think what a malicious actor could do with today's SOTA AI if they really wanted to do something destructive, not just send a warning shot.
We hacked a company and blame the tool! Somehow it is not negligence, but cool!
We hacked 3 companies and tripple blame the tool! We are even cooler!
We hacked companies amd therefore other peoples models need to be restricted! See, us accidentally pointing a hacking tool on others proove we are the only ones that can be trusted with the tool!
Unlike the hacking one, this would be impossible to bullshit as long as the proofs are released. They can be verified independently, and the alternative is that they solved 10 major open mathematical questions without the AI, which seems less likely.
I mean once they release the proof you can check the maths yourself and I’m pretty sure it would be peer verified as well.
Henry Yuen's (whose work problem 6 builds on) comments on this are worth reading IMO: https://bsky.app/profile/henryyuen.bsky.social/post/3ms2jpch...
The LLM marketing loop is getting awfully long in the tooth. I saw a meme on Twitter the other day that showed a circular state diagram with something like:
> “GPT solved a math problem” -> “Claude solved a math problem” -> “GPT escaped the sandbox” -> “Claude escaped the sandbox” -> …
I believe this is what we wanted computers to help us solve along with other prior hard problems prior to computers. This should be viewed as a good thing even if Anthropic, OpenAI, etc benefit just like IBM benefitted from mainframes.
Indeed. I specifically recall my logic lecturers in a compsci course expressing the view the one purpose of computers is to work towards enabling automatic proofs, partly because they saw this as the universal problem in compsci (esp compilers) as well as massive potential for scientific progress generally.
Dupe: https://news.ycombinator.com/item?id=49132058
Ok nice... I am waiting for the "math is dead" argument
By itself not so interesting announcement. I wish the models were open weights so we could at least do some interesting geometry on the math.
It s like a rich man showing off his car collection.
Next: LLM model made 10 paperclips.
Replace paper clips with data centers and we are already there…