Remix.run Logo
▲ ariwilson 5 hours ago

The way this is getting announced via a human mathematician humbly presenting a result is great, as opposed to the marketing influenced OpenAI post.

I am still a bit sad though. I had hoped the Rieman hypothesis would fall mostly to a human mathematician rather than 60/40 AI/mathematician that my layman's read of that note interprets.

▲alex43578 5 hours ago | parent | next [-]

I don’t get this feeling. Landing on the moon isn’t less of an accomplishment because we built the Saturn V instead of walking. If anything, can’t we appreciate how we took some sand and made it smart enough to help us solve problems?

▲layer8 5 hours ago | parent | next [-]

It would be sad if AI designed and built the Saturn V and we had it fly us to the moon without understanding the details of of how it is built and how it works. There is furthermore value for humans in going through the discovery process themselves, not just value in getting to the moon. So I understand how one can be sad if that won't happen anymore now, and instead humans are relegated to analyzing what the AI designed and built.

▲ChicagoBoy11 5 hours ago | parent | next [-]

But to extender your analogy: We do understand how Claude does what it does and how it is built, so how is this different?

▲layer8 5 hours ago | parent | next [-]

I would dispute that we completely understand how LLMs do what they do. An analogy would be that we meet an alien race that gives us their proof of the Riemann hypothesis translated to Lean. Plus of all other provable mathematical conjectures we come up with. And we might kind-of understand how those aliens do math. It would still be a bit sad that we'd be relegated to consuming what the aliens give us, and never be able to ascend to the level of the alien mathematicians

▲Philpax 4 hours ago | parent | prev | next [-]

We definitely don't understand how Claude does what it does! We know how to grow a Claude, but we don't understand what goes on inside a Claude.

▲ruszki 4 hours ago | parent | prev | next [-]

I remember when about 15 years ago, I wanted to make my neural network home work based on proven math. I had to train a model to learn basic number recognition. The math said 100000 neuron is what's needed, so I typed in that, and my laptop immediately froze completely, and died a little later. I learned on that day that we know empirically that a few 10s are enough.

We have the same thing with LLMs. We don't know how they work. We don't understand them. We have only empirical, statistical indications, nothing is proven. We know how the small parts behave, we really understand those, and also the different algorithms. But they whys? No. We're getting better, but still far away.

▲barbazoo 5 hours ago | parent | prev | next [-]

And presumably mathematicians understand the output, the proof, whatever it is that came out of the process. I don't feel like humans have to be the one doing all the manual math and figuring out all combinations.

▲streetfighter64 4 hours ago | parent | prev [-]

I mean I understand how cars work but that doesn't mean I'd feel the same accomplishment of driving 42 km as running a marathon.

▲larsfaye 4 hours ago | parent | prev | next [-]

We already had AlphaFold, which "solved" the protein folding problem, but scientists readily admit that they know next to nothing new about the field, despite it being accomplished.

▲zoeysmithe 4 hours ago | parent | prev | next [-]

I mean computers made Apollo work. They couldn't get it done with human paper calculation. The lander itself had a processor in it. So its kinda the same thing. "Humans should have done this by hand," is just an odd thing to want and ties back to some really backwards thinking. These are our tools. No one is expecting us to make homes without power tools or walk for miles instead of taking a train. At a certain point AI will be normalized and we'll have actual critical takes on AI that make sense other than "Well, I just don't like it for vague human achievement reasons." The same way people complaining about 1980s cars getting "computer chips" and "no longer being able to hand tune my carbonator" have long accepted more complex and automated and computerized cars.

As far as not knowing how it works, that becomes a bit of a philosophical problem. We know how LLMs work and their algorithms. I think we can always play up "we don't know" is disingenuous ways. How many people who made Apollo calculations could also write long proofs and theorems of the fundamental maths they were using from first principles? Not a lot. We're always standing on someone else's shoulders. We'll never truly know everything or how exactly our tools work on an individual level. Maybe AI changes things because an innovation may not have a direct line to an Isaac Newton at its base, but I think that's a bit of a vague and poetic difference, and not a real one. Its still human achievement but via machinery.

I dunno, its a bit scary, but I do feel like its just continuing the same line of human ingenuity. Not knowing exactly how something came about and non-determinism scares some people, and perhaps rightly so, but pretty much all major technological advances are scary on some level because they so strongly challenge the status quo and our ways of thinking and or ideas of what is good and right and proper. At a certain point, its hard not to see the more dramatic AI outrage as just hysteria and moral panic. I find it interesting how easily we can spot moral panics in the past in history, but when its our turn, it can be very hard to see when we're in a moral panic. To many living during those times, the moral panic sounds entirely reasonable, but its still a moral panic.

▲techmunky 3 hours ago | parent | prev [-]

And if the rocket failed at any point nobody would know how to fix it. Only the AI guessing and saying "you're right to point that out" when the rocket explodes and everyone inside perishes.

▲ariwilson 5 hours ago | parent | prev | next [-]

I also build things that would've taken me months in days now so I do get that feeling of amazement at what is possible now.

But at the same time one of my favorite memories from childhood as an introvert nerd math / computer guy is locking myself in a room for an evening and relying on my brain to solve and truly understand a problem. I'm nostalgic for that. A solution that requires 250k pages to describe takes something away from that.

Humans are complicated and contradictory, which is great!

▲jm4 5 hours ago | parent | next [-]

The entirety of human history is about building tools to perform tasks our bodies aren’t capable of on their own. Someone else in this thread said something about how cool it is that humans turned sand into a machine that could accomplish this. Our gift as humans is that we can devise ways to overcome the limits of our brains.

▲qup 4 hours ago | parent | prev [-]

If you're nostalgic for that, do it.

Doesn't make progress sad?

▲WarmWash 5 hours ago | parent | prev | next [-]

We are all sitting on a thousand years of media telling us that the human mind is exceptional and transcendent.

It's naturally going to be upsetting to everyone if reality unfolds to show that human intelligence is nothing special, and actually not even that good.

▲NoDodgeQuestion 5 hours ago | parent [-]

special yes, good no

▲xtracto 5 hours ago | parent | prev | next [-]

For the first time in humans 200,000+ years we have an "independent" or "different" intelligence that can teach us things... and WE built it. We can learn from it and improve after its teachings. This new era is going to be amazing.

▲lanstin 4 hours ago | parent | prev | next [-]

And parts of what we thought of as intellectual work turn out to be mechanistic. That was the hope for math before Gödel, and seeing it become true is an example of getting what you idealized and then being disappointed, it seems; or holding onto the numinous parts and persisting, as Professor Tao is urging us towards.

▲ebb_earl_co 5 hours ago | parent | prev | next [-]

It seems to me, having B.S. and M.S. in mathematics, that the intrigue of human intellect continues to dim. I would wager that a large cohort of mathematicians, even ones remote to the discipline now such as myself, maintain the joy of cogitation.

One felt a camaraderie with all mathematicians in history because mathematics was, wonderfully, a club with no bouncer, the only cover charge being a bit of concentration of attention. Not that one could join the elite ranks of Gauss, Euler, et al. by association, but the bonhomie was real for a lot of us, realizing that all of the legendary mathematics was dreamt up in the same way that we were engaging in it, too.

That horse has left the barn, decidedly, and there is grief.

▲curt15 5 hours ago | parent | next [-]

The effect of LLMs on math is pretty much the same as for SWE: great for existing experts like this guy, bad for juniors since the traditional criteria for evaluating merit no longer apply, and completely demoralizing for prospectives.

▲QuesnayJr 4 hours ago | parent | prev | next [-]

This exactly describes my feeling, a feeling that I wasn't aware I had until the last few months, that feeling of continuity through millenia. And you're right. It was a club anyone could join.

▲lanstin 4 hours ago | parent | prev [-]

Ironically the foundations project leading up to Gödel’s great incompleteness theorem was a project to discover a mechanism that could just spit out true theorems. We have been asking for this since symbolic logic was nailed down and yet now we find the mechanistic nature of our cognition to be a disillusionment.

▲Flatterer3544 5 hours ago | parent | prev | next [-]

Some things getting solved opens a new whole dimension of new possibilities, so assumebly the negativity steems towards the fact, if this did just take us from A to B and there were no new undiscovered paths encountered during the journey to be explored..

▲wat10000 4 hours ago | parent | prev | next [-]

A moon landing is something that pretty obviously can't be accomplished without advanced technology. This is something that was at least believed to be achievable without the equivalent of a gigantic rocket. Imagine if the conquest of Everest had taken longer, and the first person to reach the peak got there by helicopter. We could appreciate the technological prowess represented by that landing, but a mountaineering enthusiast might still be reasonably disappointed that a climber didn't get there first.

▲jobigoud 4 hours ago | parent | prev | next [-]

> Landing on the moon isn’t less of an accomplishment because we built the Saturn V instead of walking.

I think it's quite telling that by "landing on the Moon" you meant the humans landing on the moon and not the first actual landing on the Moon by human-made machines, which was in 1966, 3 years before Apollo 11.

So it looks like you do think that it's less of an accomplishment after all.

▲peteey 5 hours ago | parent | prev | next [-]

A more apt comparison is getting to the moon by outsourcing the Saturn V design to China.

Previously our talented contribution was doing the thinking. Now, AI seems to do this as a separate entity.

▲tock 5 hours ago | parent [-]

The moon landing was an achievement for all of humanity.

▲throwaway132448 5 hours ago | parent [-]

But the Nazis in particular.

▲abixb 5 hours ago | parent | prev | next [-]

Yes. We're tool using species. We evolved from African savannah to this AI moment by using tools.

I don't understand this skepticism towards tool use. It's like saying, "I'm sad that Reimann Hypothesis solution was published as a PDF and not in a parchment or made as a verbal announcement."

Anyway, great accomplishment. What an incredible thing to wake up to.

▲psychoslave 5 hours ago | parent [-]

Now that you suggest it, a video of the person declaiming the announcement while holding a parchment, would definitely would be nice to have. Hopefully someone will generate that using some genAI video tool.

▲alex43578 4 hours ago | parent [-]

No, you’re letting AI take the job the town crier! What’s he supposed to do now?

▲psychoslave an hour ago | parent [-]

Relax have a cocktail, wonder about the next fun thing to do and the next challenge to overcome to reach indefinite pananthropic harmony.

Getting rid of the meritocracy mythology for example.

▲SoftTalker 5 hours ago | parent | prev | next [-]

I can lift more weight with a forklift than I can on a barbell but it doesn't give me any sense of accomplishment.

▲ben_w 5 hours ago | parent | next [-]

Aye there's that, but I'd feel pretty accomplished if I built a working forklift.

Or used a forklift to build a house that everyone else said couldn't be built because it was beyond human muscle.

▲pembrook 5 hours ago | parent | prev [-]

If you built the forklift you'd have a sense of accomplishment.

▲acedTrex 5 hours ago | parent [-]

Alpoge did not build the forklift, kind of a basic flaw in that analogy.

▲pembrook 4 hours ago | parent [-]

Don't think Claude comes with a 'Solve historical math hypothesis' button.

Pretty sure he didn't just prompt 'Solve Riemann, make no mistakes'

▲teeray 5 hours ago | parent | prev | next [-]

It’s a bit more like we created sentient animals in a lab, they ran off to another planet, then sent one of their spacecraft back to shuttle us to Ganymede. We have no idea how the spaceship works, and if it never arrived, getting to Ganymede would still be science fiction. You can’t authentically call that an accomplishment… just a result really. We get a participation trophy at best.

▲uludag 5 hours ago | parent | prev | next [-]

I think a more apt analogy which gets at what they probably felt was immagine if humanity accomplished the moon-landing via a supposed Walmart Inc. Space Division, the announcement of celebration of this event felt like a giant Walmart promo, and the only reason they wanted to go to the moon was to put ads on it.

▲all3z 4 hours ago | parent [-]

You mean like that energy drink company that has a better space program than most countries?

▲acedTrex 5 hours ago | parent | prev | next [-]

I can not fathom this viewpoint, it is as alien to me as being a flat earther. What frame of mind is required to consider an analogy of "we built saturn V instead of walking" to be even remotely ontologically in the same stratosphere as what is happening today?

I can't even begin to make sense of that lack of basic understanding of humanity.

▲lanstin 4 hours ago | parent [-]

It turns out that the cognitive processes of advanced maths are subject to mechanization, just like carrying a load from point A to point B. It is a blow to our self notion that work of the mind is somehow fundamentally distinct from work of the rest of the body.

Also, people are nervous that the mechanization of new human capabilities will give even more power to money and less to labor, and we will live in some distopian libertarian society. I think that is a separate and realistic fear from the grief of discovering our thoughts are not as special as we told ourselves.

For me, if I am confident that the social structure will enable fulfilling lives for people, my kids and their peers, then I am ok with exploring and understanding the possibilities of mechanized cognition.

And to those objecting to the phrase mechanized cognition, I use Claude a lot and the thing I find missing is that independence of perspective that all the people I encounter have. We wake up with a perspective and goals and preferences, a why to answer what built into the protoplasm. Some technology will rebuild that aspect of our experience in a different substrate, but it doesn’t feel like a system that can’t count but can prove Riemann is quite wired correctly for that fleshy self motivated loop.

▲ericmay 4 hours ago | parent | prev [-]

I'll do my best to explain what is happening here, as I see it.

There's a reason Paul Graham wrote Hackers and Painters and not Hackers and Assembly Line Workers.

Many people long to be craftsmen and to have an understanding of their place in the world and within the social hierarchy.

These folks want to perfect some art or craft, to be the best, or to be skillful, to "create" and find joy in creation, and they pursue a level of purity in the art. Totally human, completely normal, and part of what has led to our success and advancement as a species. Think about your open-source maintainer, or the gal who has Ruby on Rails stickers plastered on their laptop, or maybe you yourself as you try to get excited about the super boring details about the latest release of some open-source package you want to be the expert in (why?). They want to be "hackers" and "painters"!

The problem is that the pursuit of this particular craft (software development) falls into the realm of Science and Industry. Engineering, code, electronics, machines, experiments - all of these things and more are part of those two fields and those two fields by definition require and improve upon existing technology or understanding and as such they are fields which change often, cause some skills to become obsolete without remorse, and they do so at a fast pace.

The software engineer who believes in their craft and their expertise in JavaScript is finding that Science and Industry no longer value that craft and expertise - so that person's place in the world is directly under threat, including their income in many cases. It causes a range of emotional reactions.

You can see how this manifests with many software engineers going on to do silly things like move to Japan and live on a mountain top to write code in peace, or start a Substack newsletter that nobody ever reads, or build expensive tiny sheds ("houses") so they can go in to their back yard and "code in piece" - i.e. pursue their craft. Another common theme is minimalism - flat surfaces, clean lines, things like that. The person seeks more and more control. If only they could just remove this inefficient piece of furniture at last they can finally code and do my painting in peace with my perfect coffee and get in to my zen flow. It's a whole psychology, and it's under grave threat from AI.

With the threat of AI, of course the natural human instinct for many of these very intelligent software engineers (most are about average intelligence however) is to perhaps physically and culturally retreat and go "farm" (interesting how they try to go pursue other crafts isn't it?) or emotionally retreat "AI isn't a big deal, I'm important, this technology won't beat out my craft" and these very natural human reactions you are seeing are due to psychological defenses built up by average people who believe their livelihoods are under threat.

So the comments will of course be around things like denial, or impurity (it's not pure enough that the Saturn V was built - that's cheating, we should have invented wings and flew there ourselves), or it'll manifest into a mechanism in which many are actively cheering on a foreign adversary to "beat" their home country because it'll validate them in some weird way. See? Look what you did. You created AI and now we're behind! Told ya! There's no changing their mind because their job depends on their mind not being changed.

It's difficult for someone like yourself to combat these feelings that you don't understand without first understanding them, and to understand them is to brush away some of the nonsense that people get up to and get back to basics which is that many people feel under threat from advancements in technology, many people want but lack a sense of place in social hierarchy and purpose in life ,and for those that do have such things these advancements threaten to disrupt their current lives and purpose. Of course they're going to tell you we should walk to the moon instead of building the Saturn V (AI).

Their lives depend on it!

▲foolserrandboy 4 hours ago | parent | prev | next [-]

Maybe it is beyond one single human's ability to solve in one lifetime. It has been around 150 years why would we assume a human CAN solve it. Maybe the only way through these hard problems is to use AI. There could be some evolutionary hard limits on how much a human can fit into working memory to solve these hard problems. And we still get to read the results and try to understand them. Btw we built the tools and the tools use human developed patterns of reasoning to solve the problem so it is still very human in a sense, not alien.

▲suddenlybananas 4 hours ago | parent [-]

I'm not a mathematician but surely a lot of the math used is this proof did not exist 150 years ago.

▲Sol- 5 hours ago | parent | prev | next [-]

I think Alpöge specifically was criticized in the past for being somewhat disrespectful of the math community, since he just dumped AI slop onto Twitter in meme form. The OpenAI post was actually more respectful compared to that. So this type of announcement is more of a return to somewhat more deferential communication.

▲mahogany 5 hours ago | parent | prev | next [-]

Probably closer to 90/10

▲ks2048 an hour ago | parent | next [-]

I'm not sure numbers like this have much meaning.

It's worth remembering that this was done by someone who (as far as I can tell) is a great mathematician and not by some "prompt engineer" with an interest in math and large budget.

▲suddenlybananas 5 hours ago | parent | prev [-]

None of us can know this without all the prompts/history

▲kingstnap 2 hours ago | parent | next [-]

TFA literally says this:

> but this time Claude far outpaced me near the finish line

> I intend to do my best to understand and impart understanding of this argument and its certain simplifications, to speak about it, to write it properly

> we were going to win but I was also, so to speak, out of my element, in that I personally have little mastery of automorphic forms on higher rank groups

> I’m just too inexpert in the relevant stuff at the moment, I hope the reader sees the breadth of this argument!

Also the exposition at the bottom of the paper is straight up just claude.

> So here is a rough timeline Claude provided

Also its like literally gigabytes worth of lean... it goes without saying that that is going to be effectively 100% claude.

I'm sure Levent put in some good work, but also this is clearly in large part the work of claude.

▲mahogany 3 hours ago | parent | prev [-]

Sure, and it depends on what we mean by the ratio. But, seeing that the proof comprises millions of lines of lean code (which have not been read or understood by a human), or 250k pages of prose (which have not been read or understood by a human), it makes me think the machine largely "did" the work of this proof. How far do you think he would have gotten without the machine?

Actually, thinking about it more, the ratio might be closer to 99/1.

▲tomrod 4 hours ago | parent | prev | next [-]

AI is just a tool, at the end of the day. Regardless whether the proof is accurate, this is showing strong capability of both the mathematician as well as the tool.

I have a substantial math background and certainly can't yet prompt LLMs to solve extraordinarily hard proofs. This is a real achievement. That the logic wasn't 100% worked out by a human is less exciting than what this enables.

▲kypro 4 hours ago | parent [-]

> certainly can't yet prompt LLMs to solve extraordinarily hard proofs

How long do you think that will last? A couple of weeks, or a couple of months? Surely, not years?

▲tomrod 4 hours ago | parent [-]

In my specific case? No idea, but the first step would be to want to start prompting LLMs this way. Usually driven by a side project or other need.

▲sashank_1509 4 hours ago | parent | prev | next [-]

I’m pretty sure it’s 90/10 or 85/15 based on my reading of the post

▲Waterluvian 5 hours ago | parent | prev | next [-]

Real mathematicians do all their work with calcium carbonate, not silicon dioxide.

▲rustyhancock 4 hours ago | parent | prev | next [-]

Given Anthropic IPO was yesterday, I suspect anthropic needs to be cautious about how they present anything not entirely confirmed.

▲corndoge 2 hours ago | parent [-]

> There is an obvious cynical view, to do with the timing, but it is false, and I am pleased that I am in circumstances where it is obviously false.

FTA

▲dist-epoch 5 hours ago | parent | prev | next [-]

40 percent human? In this case the human has 0 percent math input.

It might have a lot of engineering credit (setting up the swarm, ...)

▲suddenlybananas 5 hours ago | parent [-]

You didn't read the PDF, that's not what Alpöge did.

▲dist-epoch 5 hours ago | parent [-]

I did read it:

> I'm just too inexpert in the relevant stuff at the moment, I hope the reader sees the breadth of this argument!

> I intend to do my best to understand and impart understanding of this argument

He was shepherding Claude.

▲suddenlybananas 4 hours ago | parent [-]

> I guided us to the right situation > I had us on the line of combining two quite different ideas present > The parts in which my vague intuition worked out due to Claude’s technical mastery

I am not saying he could have done this without Claude. But there's a reason he did it and not anyone else. This wasn't a case of giving Claude the prompt "solve riemann hypothesis make no mistakes"

▲mahogany 3 hours ago | parent | next [-]

What is the reason? I'm sure he is a talented mathematician, and that is undoubtedly necessary here, but how many other mathematicians have unlimited access to internal frontier models? Clearly, the playing field is not even.

▲famouswaffles an hour ago | parent | prev | next [-]

>But there's a reason he did it and not anyone else.

Well yes he works at Anthropic lol

▲spindump8930 4 hours ago | parent | prev [-]

> there's a reason he did it and not anyone else.

This is the part that deniers love to ignore.

▲iknownthing 5 hours ago | parent | prev | next [-]

Don't understand the sadness - are you interested in the discovery or credit attribution?

▲Zambyte 5 hours ago | parent | next [-]

The process of solving hard problems like this often involves inventing / discovering new branches of mathematics in order to solve the problem. By solving the problem with AI, we haven't potentially snuffed out the potential for discovering these new branches of mathematics.

▲Scea91 an hour ago | parent | next [-]

In this case it feels humans alone were given enough time.

Also, I know its beyond what a single person can affect but lots of the problems that the community faces now seem to be in essence people and incentive problems. Existence of this proof in principle doesn't stop anyone from pursuing the problem further and discovering these new branches of mathematics too.

What if it was a human proof and in its process, no new mathematics were discovered? Would we also frown that?

▲iknownthing 4 hours ago | parent | prev [-]

So it's not the discovery that's important it's how you discover it?

▲aezart 4 hours ago | parent [-]

Here's a stupid video game example: If you fast travel everywhere you accomplish your current quest more quickly but you never find any of the side quests that only trigger when you stumble across them in the wild.

▲whycombilater 5 hours ago | parent | prev [-]

People often prefer horses to combine harvesters.

▲marcusverus 5 hours ago | parent | prev | next [-]

The F-47 is pretty sweet, but what a shame that it was made without employing a single, solitary slide rule!

▲bitwize 5 hours ago | parent | prev [-]

Get used to it. No human will prove anything significant unassisted, ever again. This is reality now.

▲dahart 5 hours ago | parent | next [-]

This kind of assistance is new, but I’d say nobody has ever proven anything truly unassisted, and people have been trying to point out for a long time that crediting a single person for discoveries and claiming they were “unassisted” conceals the true story. Education, background knowledge, conversations, correspondence, and context have always been paramount. This is one reason Math has often hit points where multiple people make the same discovery at nearly the same time “independently”. It might be a good thing that we can now dispense with the myth of anything being “unassisted”.

▲bitwize 5 hours ago | parent [-]

I mean unassisted by AI.

▲psychoslave 5 hours ago | parent | prev | next [-]

Can you prove it?

▲nozzlegear 5 hours ago | parent | prev | next [-]

Get a load of Nostradamus over here

▲parineum 5 hours ago | parent | prev [-]

No human has proved anything significant "unassisted" in quite a while.