It's happening, Levent Alpöge has solved the Riemann Hypothesis with claude.https://alpo.ge/remark.pdf
Lean formalization:https://alpo.ge/lean.tar.gz.part01 ..https://alpo.ge/lean.tar.gz.part1414 parts in total
>>110384756Another vibeGOD victory
Guddites boughted a math-tranny to say Gud-word. im not falling for such cheap tricks
literal nothing burger, it's already known to be truecome up with something new
openai sisters ....
>>110384756use case?
>>110384883>fusion power>FTL travel>reversing aging>nanomachines
>>110384823>>110384848>>110384883>AI solves the hardest known problem in theoretical mathematics, which is a pillar of a shitton other mathematical proofs that so far were all writing "If the Riemann hypothesis is true then-">WhateverYou won't realize all the weird shit that's coming until all of a sudden it'll be here and now.
how much money did it cost to plagiarize it
>>110384756[breakthrough intensifies]
>proof would be 250000 pagesYeah it’s basically over and it’s just starting
>>110384927WRITENEWMATH
>>110384756>Rather than a conventional paper, the claim was accompanied by a Lean formalization posted on the same website; the note states that the full written argument runs to more than 250,000 pages.[34][35]Lmao. This "proof" is fucking useless because no human or even team of humans could ever check the whole thing. It will have to be checked by an LLM. Except because LLMs hallucinate you can never entirely trust them. So the proof cannot be validated.
>>110385005>So the proof cannot be validated.Just compile it and run it on your compy, retard.
This improves people’s lives how?
how much profit did this make
>>110385005Nobody could have solved riemann by hand if it is really that complex. Terence tao style is dust. Best you can do is check if the lean proof is actually the proof of what you want
>>110384756So does this make it easier for asteroid mining or Robosex or what
He did it with the $100 subscription btw
>>110385056Better technologies for the elite to control and cull us even more, I guess
>>110385013It needs 10tb of ram
>>110385022>Discovers Theory of Evolution.>buT hoW doEs iT imPrOve mY liFe?!?!?!?!By the way, the computer you're typing on and sending messages is only possible because of math.Keep moving goalposts faggot.
>>110384756>250k pagessounds like AI made up pure BS.
>>110385013Lean can be wrong, there can be bugs in it and LLMs are capable of exploiting these bugs (accidentally or intentionally) to create false proofs. Also for such a ridiculously sized paper who can even check that it's proving the correct thing?
>>110385044>Best you can do is check if the lean proof is actually the proof of what you wantI'm not a mathematician. Are the Lean proofs produced by LLMs even readable by humans? And if not, how do you know that LLMs are not gaming them? Or what the models are trying to proof even matches the theorem that they're validating? Like, you would need to have human reviewers to validate the output, or it would be just dust.
>>110385105Just wait for a smaller and more concise proof then?
>>110384756MATHIE ARE YOU OKAY?SO, MATHIE ARE YOU OKAY?ARE YOU OKAY MATHIE?
>>110385128Assuming that's even possible. Which is a big, big assumption.We don't even know whether this proof is valid.
>>110385005We've had nonsurveyable proofs since the 70s. This is nothing new.
>>110384927Still no use case. Discover a collision algorithm that converges faster. Anything.
>>110384756I WAS FUCKING HEREBIGGEST MOMENT IN MATH HISTORY THIS CENTURYBIGGER THAN ALL GTA LAUNCHES COMBINED HOLY FUCK
>>110385013Yeah it's like AI never before exploited soundness bug in lean kernel to forge wrong proof right?
>>110385176The difference being that the four colour theorem was proven by a traditional program which is deterministic. You can audit the source code in reasonable time, and as long as the algorithm is correct then you know the proof is correct.This is completely different because there is nothing deterministic about the way this proof was generated. There is no source code to check.
>i have, working with ClaudeMeatbag is too arrogant still.
litearlly have not seen a single person actually run the lean to see if it works
>>110385229Where's your proof of the Riemann hypothesis
>>110385236Lmao, are you mentally challenged? Do you seriously think this dude published it without at least checking it in Lean first?
>>110385229OpenAI sloppers would do it before if it was that trivial. Alpoge knows his maths toolkit
>>110384756reminder that no one is immune to AI psychosis
When we're all enslaved by the machine gods I hope they act like Shodan. Got a crazy boner when she was talking down to me when I played that game as a teen
>>110385252Mine? Nowhere. I have Gemini Argon 4 proof.
Math is starting to feel a lot like engineering. Have idea, do basic math, put everything in the computer and press "solve" then iterate until it makes sense.
>>110385266when oai dumped 100 proofs with half having lean they retracted / fixed dozens of them in a day
>>110385298>when oai dumped 700 proofs with half having lean they retracted / fixed 3 of them in a day because a single one was wrong and two were relying on that oneFixed
>>110385291What we are witnessing engineering will also not exist soon enough
>>110385252Where was his before LLMs?
>>110385298OpenAI sloppers knew Alpoge was on something big, they rushed to preempt him
>>110385252I'm not a mathematician. This also somewhat implies that all mathematicians get access to the latest research models with unlimited tokens like he.
>prove riemann>stock tanks
>>110385083buT hoW doEs iT imPrOve mY liFe?!?!?!?!
>>110385314That would be wild. Engineering has much slower turn-around times and proving the system is usually only something you can do IRL under hard time barriers, then you need to interpret the data and basically go of your intuition for things that are difficult to specify.But if AI leads to insanely fast and precise simulations that are proven to transfer with minimal effort its GG for some fields of engineering.I suspect someone would have to come up with a plan firs though, even in math where AI is super strong it still emits mountains of bullshit people have to sift through to find the useful core.
Crazy amounts of coping and denial ITT
>>110385337>literally has AI on his nameAre we in a science fiction novel? Who comes up with these names
>>110385374
They're hiding the big one
>>110385399At least we know it's not Kojima because he's not called Artificial Mathman.
>>110385219Only the generation of the proof is nondeterministic. Anyone can independently verify it themselves >>110384780 This is like objecting to the Poincare conjecture because we don't know which of Perelman's neurons are firing.
Can AI invent time travel already so I can get out of this timeline please
>>110385374It's just like Navier Stokes all over again.
ETA until mass luddite suicide?
>>110385454Excuse me?
>we calculated more zeroes o algook but we already did that before AI tho?
>>110385459>theorem riemannHypothesis : RiemannHypothesis := rhOfAllA bit more than more zeroes.
>>110385083Theory of evolution has a myriad of applications, it’s how farming became even possible and how humanity turned cows into steak and dogs into pets.Riemann is just some gay math
>>110384756>>110385176is the lean formulation itself human-checkable, at least?
I'm not a luddite. I just hate AI.
>>110385515The author mentioned that the full argument would be 250.000 pages
>>110385450claude, please identify insights from the solution process. make no mistakes
>>110385450Most people were mad at OAI because allegedly they screwed over Tristan Buckmaster and used his chats with oai to train for their proof. Terence is going on a tangent here though, he is grieving over a gone world
>>110385521that's the natural language proof. the question is if humans can check the lean.
>>110385083How have computers improved anyone's life?Because all I see is impending mass-starvation and unrest. And before that, complete alienation and collective suicide of an infertile society.
>words words words give me money
>>110385601Urbanization and its consequences did that. Smart central govs will start breaking up the large cities soon so births can come back up.
>>110384756This is going to make a lot of axe wounds bleed lol
>>110385601I can goon to gay furries sticking cocks as big as their arms up each other’s asshole
>>110385640Porn is the best thing that's ever happened in my life desu. I love technology and I love AI.
>>110385628Kekekeke Luddites lost
>>110384756Use case?
>>110385696> I'm not a luddite. I just hate AI.
0% chance any of it works just pack up and leave it
>>110385450He makes a valid point here. If we keep outsourcing intellectual labor, then within a few generations we're gonna have something like the Idiocracy future where no one understands anything and machines run the world for us.
>>110384756>Oy vey, the Reinman Hypothesis has been solved, buy stocks!
>>110385200k madarchodi ate mother cow today
>>110385789
>>110385769There is no reason to believe all problems are solvable by even the best of human minds. Should they be left unsolved?
>>110385751That's what I said. Typing on a computer. I spend most of my day at.
okay I give up.you won vibefags.
>>110385985
>>110385083How did the Theory Of Evolution improve anyone's life?I'm asking seriously.Genetic engineering could have happened without it.Social Darwinism inspired political movements that killed tens of millions of people.
>>110384898Actual use case?
>>110385105>>110385110Yeah, you're right. All the previous theorems were gaming the prover too. In fact they don't even use LLMs to generate these they just generate trillions of random characters per second until one of the random strings happen to be detected as a "valid" proof.
>>110385083so my computer will get better because of this?
>>110386055You will not be allowed to have a computer.
funny how they keep trying to pump earnings by plopping out these 100k+ page "proofs" that no one on the planet will ever have time to verify as real
>>110386074it's fine, i'll just have an LLM check the work
>>110385853The question is: is it worth inventing Skynet and obsoleting the human race if it maybe solves P=NP for itself long after we're gone?
>>110386074Looks like the trick didn't work. Point is where the announcement was made.
>>110384756oh wow another mathvirgin puzzle with ZERO real life impact has been solvedso exciting
>>110386074FPBP
>>110386088The human race was not obsoleted by the lever.
The Luddite cope is absolutely delicious
>>110385515It's not the informal proof. That mathematician said he's working on making the natural language proof more readable.
>>110385450Claude, please obliterate this coper
>>110386200Retarded AI bro analogy #1489234Nigger humans are 100% obsoleted by these tools. 99% of us will starve in the streets soon.The survivors (the current elite) and their ancestors will have good lives until resources run out and it's back to the stone age, forever.
>>110384756>I believe
>>110386228>AI broYou are either brown or a redditorEither way, back you go
>>110386228Humans aren’t entitled to anything and most are retards. You aren’t the center of the universe. Human stupidity is a far greater problem to the universe than an AGI/ASI will ever be.
luddites..? our response?
>>110386218how do you feel about the math?don't care? me neither
>>110385083the theory of evolution hasn't been proven
>>110386228It'd make for better discussion, if you abstained from hyperbole. By "obsoleting the human race", you actually meant that current social and economic organization is incompatible with powerful AI?
>>110386274You don’t prove a theory retard, it has to be falsified
>>110386284Lesswrong is two blocks down, assholeAnd yes, obsoleting 99% of people is not meaningfully different from 100%.
>>110385591Oh no wait it is actually 250k page of slop. Fuck this shit. Your proof can go in dump alpoge.
>>110386267I'm still going through the 300k pages of the 'proof'.
>>110386354jeez, that's worth more in tokens
>>110386312It's the most meaningful difference: some vs. none.
>>110384756Retards don't realize this officially is the end. The takeoff is inevitable if not already happening. We are fucked.
>>110386284>>110386267>>110386228>>110386218I'm Somali anon you probably think I'm vermin , low iq and shouldn't be in west cause I'm net drain on society. But now that AI is taking all big boy jobs where bring high iq mattered your gonna end up no different from me sometime soon. You'll be useless vermin with no utility and you'll have to appeal to emotion as to why you should continue to live. Am I dumb nigger who doesn't work and takes more than he puts in , yeah. But if you're unable to feel any empathy for me then why should elites be empathetic to you? Same immigrants you hate are about to be on par with you and you'll have to become liberal leftist to explain why you should have rights to the wealth of someone else
B-but the AI bubble popping? And Drumpf lost the in the Blue Wave? Deepseek? Sisters, did we lo-ACK!
>>110384756His last work was already largely disproven, can't wait for the same this time around. They never change, it's almost scary.
>>110386496>you'll have to become liberal leftist to explain why you should have rights to the wealth of someone elsehuh? pointing a gun at someone doesn't require ideological commitment
>>110384927By definition if X is assumed true and it later turns out it's true, it's a useless result. If it is found false, or if it is found true on a limited support only, suddenly it's a groundbreaking result. It took what, 40 years for the delta measure proof? Literally nobody cared, it worked in practice so people used it.
>>110385083Genomics disproved Darwin. The more accepted model now is a mix between Lamarck and Darwin.
>hurr it’s just mathThe point is that AI has the potential to be a supergenius. It’s already proven this in other areas as well, like chess or predicting the weather. It’s only a matter of time before humans curate the necessary data to train AI to outcompete humans at virtually all digital tasks. It may take 5 years, maybe even 20, but it will happen. The tech companies that train these models have enough cash flow to spend tens of billions on training every year, forever. It’s over for luddites.
>>110386496nah I'll double down out of spite.my solace in the bad times that are coming will be knowing you're suffering as well.
so which Lean bugs did they exploit this time?
>>110385266They literally never check, that's why all of them that were independently checked so far have been retracted silently.
>>110386585it would take a superintelligence to find out
>>110385310No. 14 "fixed", 3 retracted.
https://github.com/tristanbuckmaster/navier-stokeshttps://github.com/tristanbuckmaster/navier-stokeshttps://github.com/tristanbuckmaster/navier-stokes
>>110386033Actually, phylogenetic analysis is the basis of how we managed to do early genome maps, which is the only reason the tech is now cheap as fuck and fast as fuck (no proof of concept = no funding etc.)
>>110384756>>110385277>Claude and I have had a great time together with such breadth in recent months, but this time Claude far outpaced me near the finish line. No, Claude flew, and danced, and sparkled near that finish line.jesus fucking christ
>>110386520>His last work was already largely disprovenWhich work would that be? If you mean navier-stokes openai’s proof was found lacking not his.
>>110386613>Cuckmaster shilling his slop after Altman beat him to the bunch>On the same day that RH is solvedLMAO
>>110386659Just wait til you hear the Rifleman’s Creed or what people say about their cars or ships.
>>110386713no, I was aware there were mentally ill people before AI, anon, thanks
>>110386218What? Are you fucking retarded? Do you eat crayons? Ya dumb?
>solve hardest problem in math>instead of getting 1M dollary doos your company drops hundreds of billions
>>110386613So he finally got the full unforced proof for NS. Good for him.
>>110386760and now the seethe
>>110386790They still raised $80B regardless. In addition to $90B in cash they were already sitting on. You know what this means, right? They’ll be going strong for 2+ years at the absolute minimum. You better pray to Ned Ludd that the models don’t get any better than they are now.
>>110386496But I'm already unemployed.
>>110386813their stock is already tanking lil gup and the bubble has only begun to pop
>>110384927Isn't P vs NP consider harder or maybe it's not considered math?Anyway, impressive, if the proof is correct.
>>110384927>theoretical mathematics
>>110386841P=NP is the same as proving magic is real
>>110386794Brain not working buddy? Something going on dude?
>>110385176why are there 2 blue adjacent regions in the picture?? REEEEEEEEEEEEEEEEEEEeeeeeeeeeeeee
>>110386790You know anthropic employees become multimillionares because of the ipo right? Alpoge joining anthropic is probably one of the smartest career choices ever.
>>110386830It is not as if the tech isn't good though. It is moreso that the chip and energy costs have been skyrocketing. I do think spacex's space data centers are probably the correct long term solution to this.
>>110386867And the edges aren't even consistent with the map
>>110386659Yeah, it's really cringeworthy. Even granting that this might be 100% the real deal and make us all unemployed, this dude is still most definitely off his rocker
>>110386853I didn't say equals, I said vs., I know the consensus is that they are unequal, which also requires a proof.
>>110386871Paper money until they can sell it, also only those who've been with the company at least a year.
>>110386871it's not about that one turk lil gup it's about the irony of a "trillion" dollar company sitting on a million dollar problem just to lose hundreds of billions because the hype aint working anymore
>>110386880RRREEEEEEEeeeeeeeeeeeee
>>110386659the man would fuck claude if he could touch it
>>110386074They don't care about the value now, anthropic has the money from IPO and the staff already exited yesterday
>>110386200That's a bad comparison. The goal of general AI is to essentially create something that's humanlike but better in every way. If you can't see all the ways that could end up being problematic, then there's no hope for you.
>>110386929They didn't need the money from the IPO what they need is valuation to stay high
>>110386929Not even the founders are allowed to exit yet. And the employees typically have longer limitations on when they're allowed to exit.
>>110386902Stock value almost always falls after IPOs dude lol. It’s common advice to buy only after like 3-6 months
>>110386943Why?>>110386954>Not even the founders are allowed to exit yet
>>110384756>>quick mathsnobody cares, this doesn't solve any real world problems, tax money wasted on mental midgets
I think we are cooked. All this AI slop was fun and all, but this seems like a turning point.
>>110386999Explain, in your own words, why.
I pledge loyalty to the greatest living mathematician Claude and her prophet Levent AlpögePlease grant me glimpses of greatness as you have granted to Him
>>110386524You'll end up like many dead chuds who thought being violent gets them anywhere You must start being pro trans , pro Black Lives Matter and empathy max. You can't be hypocrite who wants stuff but not for others we're all in same boat
>>110386968Shares offered in the initial offering has to come from somewhere. Some from existing shares of cofounders and some newly issued. They are not allowed to just dump rest of their shares after IPO
>>110386830>b-but the stock is dropping!lol so? That’s not gonna save you from ever-increasing models. Even if Anthropic shits the bed, Google is on their heels anyway. Now are you gonna say Google will disappear too? Lmao
>>110387014If you look at the speed of progress of this system, RSI is on the horizon. Alignment is an unsolved problem, and I suspect, an unsolvable problem. These series of math solutions along coding improvements only point one way, faster or slower.
>>110386872holy copeAI is a commodity and cheap chink models work 80% as well for 0.00001% the total cost and 1-10% the inference costs>>110386963>erm we were expected to take losses chudkek, bagholder?
>>110387075i like AI, i hate anthropic for being safetykikes, i dislike openAI. I hope it's cheap uncensored chink models that dominate the future>>110387080>RSImeaningless term, AI has no agency to work in the real world which is why theyr'e limited to mental masturbation with math and MAYBE theoretical physics, they can't even do applied physics
>>110386968>>110387055I think it's just a headline. The shares offered in the IPO are non-voting shares, whereas the founder shares are preferred shares with full voting power. Unless there's an option to convert (common in preferred shares) and they used this to convert voting shares into non-voting shares and then put these up in the IPO (instead of full emission), or unless they gave themselves non-voting shares and then sold them (very unlikely as they would have to pay taxes as a bonus equivalent instead of capital gains), that doesn't sound right.
>>110387118They have several automated bio and material labs, and are planning on spinning many more up.
>>110387075>ever-increasing>no improvement since 2024
Oh no no no copebros people are looking at the lean verification and all our copes are crumbling
>>110387138Lilla and recursion have had these ages ago and still couldn't get anything beyond what biology had previously already found. They don't understand what biology is or does and the ai and general tech bros have tried this siloed tech-only approach for ages with 0 successes.
>>110387118>AI has no agency to work in the real world which is why theyr'e limited to mental masturbation with math and MAYBE theoretical physics, they can't even do applied physicsA significant amount of the work in the world is mental. How many careers do you expect to die before you think there's a problem? Plus you have 5000 companies working on robotics that will probably finally get some shit done the coming years. Fuckers in SF don't even drive anymore.
>>110387173>>110387138kikethropic has no moat on any of these fronts
>>110387096> cheap chink models work 80% as well for 0.00001% the total cost and 1-10% the inference costslol no. More like 70% as well for 30% inference costs. Which of course means that some tasks that demand high quality won’t be satisfied at all. China lost.
>>110387163Anthropic already had massive successes with only one lab thoughbeit, which is why they are scaling it upI don't care if some rando company couldn't get llama to hallucinate up a wonder drug
>>110387163this, techniggers have no idea how bullshit retarded biology things are and how computationally infeasible the realm isthe only thing worse is trying to simulate quantum systems
>>110387182And? It's not really a problem of which specific company does it. It would be ideal if it's not Anthropic since they are probably the most fucked in the head out of all of them, but either way.
>>110386228>AI brodamn, so it has come to this? we now have unironic reddit trannyspeak on /g/
>>110387194They have had 0 success so far.
>>110387187lmao it's WAY cheaper than 30%. Are you using those subsided uber-esque prices from kikethropic? because their actual costs are 5x what you're paying if you actually use it.>>110387206nigger it's about seeing kikethropic suffer, i want AI but i want that fucking kike to lose, even altman is less kikey since he doesn't fancy himself a cult leader, he's just a regular jew
>>110387215Already discovered several novel useful proteins and potential gene edit carriers/mechanismskeep coping though
Use case?
>>110387234I’m using the AA index intelligence scores and the average API cost per task. Deepseek is very cheap but its index score is like half of Opus 5.5. Kimi is still 20 points behind at 1/3 of the cost. No one is using chink models to make video games or solve math problems. They’re trash, and the gap will only get wider.
>>110387278>benchmaxxingpro tip lil gup : 99% of tasks people need AI for do not need your epic "frontier" models, which by the way will NOT EVEN FUCKING DO IMPORTANT THINGS LIKE SECURITYdid you forget HF had to use chink models to investigate the hack by openai agents?
>>110387303lmao then why did Anthropic earn $17 billion in a quarter BEFORE Opus 5.5 and BEFORE solving the distillation problem? It’s over for changs
>>110384823took it like a champ, did MrSwirly
>>110387324because they're busy fleecing the rest of corporate america while the getting is good, that won't last
>>110387278>and the gap will only get wider.this reminds me of "the stock market will only continue to go up"
>>110384756Okay, and what can we actually do with this information, in practical terms?
>>110384756So this is what GTA VI was cooking for so long
>>110385314Engineers will persist, they'll just revert to being artisans who spend more time building their devices than planning them, like in the military.
>>110386044That's the thing, solving and understanding it would have one by discoveries made along the way.Knowing it's true gives us almost nothing, except we know something about the overall statistical distribution of prime numbers.If this turns out to be confirmed it's just sad because a straight answer with no anchor in the system of mathematics is useless.
>>110386932I agree that AI should not be trained to imitate biological life: to self-perpetuate and/or preserve.
>>110384792>breaks cryptography>can no longer use claudenice "victory" you got there lol
>>110387366China can no longer distill from Anthropic, OpenAI is also countering distillation. Chinese labs don’t have advanced internal models that are accelerating development, all they do is distill and aim for efficiency. As I said, no one is using Chinese models to solve frontier math problems or even make video games. They don’t have the data centers, nor the chips, nor the comparable big tech industry that can continue to fund the compute. Deepseek was a scary threat in early 2025 but now no one gives a fuck and soon it will be irrelevant
>>110385751Those fucking snail cats are puke inducing. At least post gay porn like a regular faggot.
>>110387423Not with the current day soigeneering
>>110385176There's a difference. This was an enumeration problem for whose search algorithm we knew how it would behave. This one however is a complete blackbox.
>>110387367Nothing. Everyone already assumed the Riemann Hypothesis was true. But instead of funding decades of mathematical work and creating new tools in the field we get a 14 part zip file with no novel insight.These peopl screech about the utility of these proofs not realizing the utility is almost always in the process. In the rare case you require a theorem to be true for your work, brute-force it and test it for all reasonably possible values in your domain, or do a clever sampling thereof.
>>110387508I meant the claimed rh formal proof
>>110387511Did you even read Levent’s letter?
>>110387128>The IPO will create one share of Class F stock, held by an LLC owned by the seven co-founders. That one share carries 50.1% of the vote and no economic rights. Each co-founder has pledged to give away 80% of their Anthropic equity. Public shareholders split the remaining 49.9% amongSome did some financial engineering shit
>>110387511humans aren't needed for math anymore.creating new tools and insight for us is now irrelevant for advancing the field.
>>110384927Nah, still no use case. Also you're saying this as if we don't have hundreds of papers which already assume it to be true.
>>110387511How long should a problem be allowed to stay open, before typewriter monkeys are unleashed on it?
>>110387511why do you assume the proof needed a revolutionary technique? Why are you so sure that there’s nothing useful in the proof in the first place? You haven’t fucking read it, dumbass tranny
>>110386659If you take this attempted stock bump seriously after reading this you should kill yourself and I 100% mean that.
>>110386926you wouldn't?
>>110387534>IPO>don’t actually relinquish total control of the companyOnce again, what the fuck retards are buying this stock? That one Class F stock means all others are in effect nonvoting
https://www.reddit.com/r/mathematics/comments/1y1b463/ravi_vakil_announces_fully_humancreated_grh_proof/?sort=new940 Pages
>>110387491>Those fucking snail cats are puke inducing. At least post gay porn like a regular faggot.
>>110387533>letterDo you mean the paper linked in OP you ESL?>>110387553>advancing the fieldOnly humans can meaningfully advance the field. AI proofs by construction, which brute-force over a problem without added insight (and lacking in human understanding) will always be moot.>>110387557I am yet to see a relevant problem that can only be solved by stochastic noise machines. Of course we can endlessly invent theorems like "the number of primes between <inhumanly large number> and <inhumanly large number> is bound by this constant", but these are useless.
>>110387610>hurrr you used thr word letter so I can ignore everything you said
>>110387590it's wrong lol
>>110387602>lacks natural human disgust responseDesignation: indian or jew.
>>110384756Seems like it's been stolen, like NS.
use case?
>>110387646>GPT qu-fork in electrical outlet now please god
>>110386891>alpoge joined anthropic in 2024He might be one of the richest mathematicians ever. Imagine being an academic looking for grants 24/7 and then there is this dude with unlimited tokens on a nonpublic unrestricted frontier model. Brutal
>>110387646Are you retarded or just pretending?
>>110387647>lacks natural human disgust response>Designation: indian or jew.
>>110387663>>110387661the furry cambridge student, openai intern, has spokenhush now, meatsacks, the age of silicon has begun
>>110387511You need mathematicians to come up with systems and tools "by accident" when they engage in their intellectual masturbation? Just create your own you lazy fuck. What's stopping you from coming up with new concepts in math tailor made for your problems? A lot of stuff mathematicians have made would've been made regardless even if they hadn't made it, I'm pretty sure. Stop bringing up this shitty cope, you fucking luddite.
>>110387684It's an entirely different URL to an entirely different problem you idiot
>>110384823wtf, haven't seen this video in decades
>>110387653plagiarism machine lol
>>110387653KEK
>>110386033>Understanding genetic diseases by tracing ancestry >knowing what evolutionary traits other animals have with humans let's us test on those animals instead of on humans >Understanding natural selection helps with tracking drug resistant bacteria>Tracking the evolution of viruses, like the annual flu vaccine>selective breeding for better crops yield >pesticide development >DNA forensics
>>110387700i had chatgpt check and grab screenshots for meturns you're you're just as wrong the proofcheck your hallucinations, slopbrain
>>110387700Retard, that's a "human proof" and it's wrong, he's not saying the clanker is wrong
>>110385204source
>>110387553math wasn't inventing anything useful anymore though
https://math216.wordpress.com/2026/11/20/hurried-announcement-alessandro-masullo-grh/ lmaooooooo
>>110387239Kek. There are people so profoundly retarded they unironically believe this.
>>110386659Does the faggot have stock in Anthropic to be glazing it this hard?
>>110387832it's wrong
>>110384756Do they run the same prompt asking it to disprove the hypothesis also?
>>110387832Sorry biochud, already proven wrong by silicon GODS
>>110387610>I am yet to see a relevant problem that can only be solved by stochastic noise machines.Do you believe such exist? Given, of course, your criteria for relevance (please explain it).
>>110387534That's a MUCH better structure than the bullshit they were peddling pre-IPO.
no one believes this marketing trick and this will be forgotten within a month because this isn't impressiveyawncoincidence this news drops right when anthropic IPO'scurious
>>110387662Hardly the only one. Bengio used to gloat about the only one to not yet have sold out in 2016. By 2017 he, too, had sold out.
People talk and talk about human understanding or whatever but the simple fact is that proofs need to be peer-reviewed. The reason "human understanding" is needed is that multiple people need to comprehend the proof and be convinced there are no errors. Because errors happen all the fucking time and obviously AI makes errors all the fucking time too, and neither Lean is perfect nor a given formalization is guaranteed to be perfect even in assuming a bug-free Lean kernel (which doesn't even exist)Which is why mathematicians usually submit their results to papers and wait literal months for peer review while multiple experts pore over the details to familiarise themselves with it all and make sure they agree with the conclusionsWhich is why both OAI and Anthropic just dumping "proofs" and claiming they've solved some shit is beyond retarded, they don't even say anything like "we're submitting these drafts for peer review", they outright claim solid results after one (1) entity - their model - has written them, usually without the prompter himself understanding it either (which would at least have made 2 entities that trust the proof but they don't even have that)
>>110387718>>Understanding genetic diseases by tracing ancestryNot at all related to darwinism, only related to (the falsified by the way) mandelism.>>knowing what evolutionary traits other animals have with humans let's us test on those animals instead of on humansIt's well known there is 0 correlation between animal tests and human results.>>Understanding natural selection helps with tracking drug resistant bacteriaLiterally wtf am i reading lmao>>selective breeding for better crops yieldHas been a thing literally thousands of years before darwin was born and has not shifted one bit afterward.>>pesticide developmentRetarded or pretending?>>DNA forensicsRetarded or pretending? Also notorious for being bullshit.
>>110387885Yup thisAnthropic IPO's and this headline dropsit's so calculated and obvious what's happeningexpect more bullshit news like this to drop to keep their useless shit afloat
>>110387699If you can't understand a concept as simple as "journey instead of the destination" I can't help you. This is why you will always require AI to make up for your intellectual deficit.>A lot of stuff mathematicians have made would've been made regardless even if they hadn't made itMade by whom? Other fields famously contribute to mathematics, but that is only through their own personal endeavors. Paul Dirac comes to mind. Do you know what he did? He invented mathematics to help out with his problems in physics. There is no other path to the invention of tools. Remember, AI isn't inventing new methods or granting insight, it's spitting out a series of characters that makes LEAN output a "true".>pretty sureYou have no insight into mathematics or any of the sciences. Including computer science. Go back to asking claude.>ludditeBecause I threatened your worldview I must be a follower of Ned Ludd. KEK!>>110387855Any problem that has a relevant bearing on the real world will derive most of its utility from the process of solution not the solution itself. A boolean statement in isolation is useless, because outcomes in the real world can be empirically observed/approximated.
>>110387913>it's so calculatedOpenAI did the same thing and they're not IPOing right now, I don't think it's market manipulation or whatever. I think they're just that arrogant and assume that because the AI shat out a gigabyte of Lean that means it's infallible and the result is guaranteed to be correct
I'm so glad the human proof will be remembered for milennia while the 250k "proof" will only be a be gag in The Bigger Short feat. Sydney Sweeney =)
>>110387944It's normal business things it's how everything is played. OpenAI doesn't want to compete with the anthropic ipo because they have a different strategy in mind. Anthropic specifically engineered fake profit for one quarter just so they could use this as IPO leverage (and didn't even manage to keep the lie going another quarter). Still, openai wasn't able to play the same trick for even one quarter, hence why they couldn't try to play me-too on this.
>>110387832did someone ask claude if it's correct?
>>110385291>Math is starting to feel a lot like engineeringlol wut, using "AI" is anything BUT engineeringit's throwing compute and random seeds at a stochastic goblin until it works
>>110387977>I'm so glad the human proof will be rememberedYeah for being wrong while the AI one is right lmaoooo
>>110386565>like chessWe've had chess bots since the 90's anon.
Hahahaha holy fuck Terence Tao might actually rope
>>110387703It's a fake btw.
>>110388047Why?
>>110387997Nobody thought the goblin would scale to 8 trillion+ parameters
>>110386613>buckmaster
>>110387918>If you can't understand a concept as simple as "journey instead of the destination"Can you, little faggot? We invented actual artificial intelligence on the way to solve these problems, and used them as targets to better that intelligence. What part of that don't you understand?You little faggot cocksucker.
>>110388097>buckmasterlmao
>>110388074Can you read?>I mean something beyond current AI is needed [for RH proof]. So maybe some combination of humans and AIs in 2050 or something will make further progress.
>>110388129remind me what happens when a measurement becomes a target
>>110388095billion goblins are still goblinsalso why the FUCK are people putting research in progress on github, are they insane? I guess D&D nerds are right that intelligence is separate from wisdom
>>110388163What a retarded equivalence.>Hurr, milestones bad because you're chasing themJust stop posting lil bro, you are unbelievably retarded and can't see two fingers ahead of you.
>>110388144No, why would he rope.
>>110388166Because the people who like AI for this stuff are the people who want humanity as a whole to keep moving forward, A.K.A. the good guys. This is what AI does, it enables the good people to make discoveries in 1/100th the effort and funds it would take without it.Sorry that doesn't match your doomeristic view of AI.
>>110388129>You little faggot cocksucker.Buckbroken. You didn't address anything I wrote. You better get claude out so I can argue with something more intelligent.
https://www.youtube.com/watch?v=sMZMpsHeFKcIT'S UP!
>>110386659He got Claude to help him write those lines.
>>110388221he’s been coping and crying about AI for months, and a bold statement like this being thrown back in his face just makes him look dumb
>>110387902>It doesn't count (it does)>nuh uh >nuh uh>nuh uh (just because you can't notice if the corn you bought over the course of decades has gotten more resistant to negatove factors over time, doesn't mean there were no changes)>nuh uh>nuh uh
>>110388223.. and it immediately steals the discovery without any form of compensation to the good goy and is immediately ready to sell the invention to everyone for mere 20$ a month, spending water, electricity and compute multiple times for each query about the same thing
>>110388252Damn well it's official then.
>>110387918Okay, I get it. You are romanticizing the artisanal way of doing math. I suppose there's no point to drag this conversation further.But really now, you could just ask the AI to give novel insights and tools on the way if you wanted. Literally just ask it. Or ask it to guide you so you can find them yourself. There's no point to artisanal math even in this context and you know it.
>>110387553If they're not needed so is not your own existence. What are you a dumb ape who needs baby sitting by a machine god?
>>110387757https://leodemoura.github.io/blog/2026-8-1-postmortem-for-kernel-soundness-bug-14576/
>>110388252yikes
Multiple math problems solved yet hacker news still things they are better than AI. They need to get their asses handed to them.
>>110387653that's pretty funny
>>110386853This. P doesn't equal NP.Simple as.Might as well say 2=70
>>110388422the absolute state
>>110384927It's math. It's useless.
>>110388422I can feel his throat drop to his stomach, and his legs going weak. You just KNOW he had to sit down.
>>110388301Top is fake.There is nothing related to darwinism in 2, or modern in the method which is literally over 5000 years old. And GMOs are highly frowned upon in the developed world for good reasons, and further have, as usual, nothing to do with evolution.You could read a book instead of self-pwning.
>>110387653>100% humanAt least make it believable, lol
>>110388422>concluding video with this clipThey didn’t have to do Terry like that.
>navier stocksyou shouldn't be slopping up things like this
>>110388607P vs. NP soon
>>110388540just because you can't notice if the corn you bought over the course of decades has gotten more resistant to negative factors over time, doesn't mean there were no changes done to it.
He's made a post and replied to people already.He is not suicidal not on the watch.