>15 days old>solved the Unit Distance problem, solved the Cycle Double Cover Conjecture, the Benjamini–Hochberg conjecture for correlated two-sided Gaussian tests, the Bandelt–Dress quartet-distance conjecture and the Dinitz–Garg–Goemans conjectureWhy was the greatest mathematician of this generation snubbed of the Fields medal?
LLMs are "solving" all these conjectures, yet the providers don't share their workflow. If it's so easy, why don't they market to us? All people seem to be using LLMs for in academia is cheating and fraud, not these legitimate tasks they're apparently capable of.
>>17023998Many post chatlogs (sometimes as supplementary material to their papers, often as direct convo share links), and also repos for open source pipelines, repos for logs and traces and lean, spec papers for closed source pipelines, methodology papers for AI-driven math and so on. You'd know this if you were even slightly interested in looking into it instead of just reposting your increasingly stale delusional coping.
>>17023998Why would I give you my system that solved shit instead of keep it to myself dumbass
>>17023474It did not solve why the OP is a fag enigma though
>>17023474brute-forcing counterexamples is somehow new?
>>17023998>/sci/ pretending to know shit, againstfu about things you have no clue about dickheadHere's the full chat history with artifact files: https://chatgpt.com/share/6a60b2eb-0b64-83ee-9c76-7931ca1de063
>>17024150OP is a faggot is an axiom under ZFC.
>>17024315Holy cope
>can solve some gay math problem only nerds care about that has zero applications>can't solve hair loss or small dihunless technology, glad it'll be able to do my emails even better now though.
>>17024315nice strawman migger
>>17024315>brute-forcing counterexamples is somehow new?Do you know what brute forcing is?
>>17023474Claims are not verifications.
>5+ "AI did a math thing" spam threads per day>zero evidence that chatbots are intelligent>zero evidence that chatbots are driving progress in mathematics>zero evidence that chatbots are driving progress in any field>zero evidence of any real progress in the "AI" field itselfKeep failing.
>>17024752https://www.quantamagazine.org/how-terry-tao-became-an-evangelist-for-ai-in-math-20260608/
>>17024757>irrelevant sponsored "journalism"Doesn't refute anything I wrote.
>>17024761terrence tao, literal fields medalist who utilizes ai for all his research now, is non-zero evidence that ai is deeply integrated into modern mathematics research. it refutes everything you claimed.
>>17024763>this guy heckin' uses chatbotsDoesn't refute anything I wrote.> ai is """deeply integrated""" into modern mathematics research>"""deeply integrated""" Meaningless token string.
>>17024764ywnbam; ywnbas; ywnbah
>5+ "AI did a math thing" spam threads per day>zero evidence that chatbots are intelligent>zero evidence that chatbots are driving progress in mathematics>zero evidence that chatbots are driving progress in any field>zero evidence of any real progress in the "AI" field itselfPlus:>zero attempts to refute any of the aboveKeep failing.
>>17024766Luddite cope. AI Chads cannot stop winning.
>>17024766because all of that is true. and yet LLMs are BTFOing mathematicians. that's the real world bottom line
>>17024766>>zero evidence that chatbots are driving progress in any field>>zero evidence of any real progress in the "AI" field itself/threadthe limitations of transformers are as obvious as ever but AI scammers are too invested in this dead end tech to progress beyond it
>i'm ignorantly unfamiliar with what is happening>therefore nothing is happeningnever forget that "people" like this exist.
>>17024778why don't you enlighten me about "what is happening"? which "AI" company is abandoning transformers and what new technology is replacing them?
>>17024779nah. if you're going to hallucinate (just like ai, lol) about me saying things regarding from transformers, then it's obvious you have a closed mind. i will not waste my time attempting to educate a dimwit incapable of learning (just like ai, lol)
>>17024781>nahi accept your concession. nothing is happening, AI scammers are just N-tupling down on guessing the next token
>>17024782>your mother isn't sucking my dick, therefore she doesn't exist
>>17024784which "AI" company is abandoning transformers and what new technology is replacing them?
>>17024786see>>17024781here i'll quote it since you struggle>you're going to hallucinate (just like ai, lol) about me saying things regarding from transformerslemme try translating it into retardese for youi din say nuffin bout trans forming
>>17024781>absolutely seethingLOL
>>17024782>AI scammers are just N-tupling down on guessing the next tokenYes but the psychotic monkey you're arguing with doesn't understand how it relates to transformers. It's not sentient. It doesn't know the bare basics of the subject it's word-guessing corporate opinions about.
>>17024788see >>17024782>i accept your concession. nothing is happening, AI scammers are just N-tupling down on guessing the next token
>>17024782
>>17024811>slop generator picture makes my schizophrenic fantasy real
>>17024817
>>17024822>"AI"-believing schizophrenic is devoid of original thought
>>17024825
>>17024826>"AI"-believing schizophrenic straight up pulling "memes" from the marketing folder
>>17024766pretty funny how much this post riles them up. the way they feel threatened by this, you can tell they don't believe in their own propaganda
>>17024829>>17024833Samefag.
>>17024833>riled upI don't think this mindless animal is even conscious. It's just posting random GPT-3-era slop images.
>>17024766>zero evidence that chatbots are driving progress in mathematics>zero evidence that chatbots are driving progress in any fieldask me how I know you haven't looked at recent arxiv papers in 6+ months>zero evidence of any real progress in the "AI" field itselflocal models stronger than last year's pro models now run on laptops but sure, no progress
>>17024838>he believes in "AI" progress.Ask me how I know you're functionally illiterate and can't even read any arxiv paper.
>I know gaming has progressed because Doom runs on my toaster nowAI cretins unironically think this is an argument.
>can't integrate new information (aka no learning)>can't form new concepts (aka no innovation)>hallucinates (aka no reliability)>falls apart at out-of-distribution tasks (aka no true generalization)>can't generalize reasoning beyond specific contexts (aka can't reason)>evaluates high-perplexity schizobabble the same as high-perplexity works of genius (aka no understanding)yes, that describes GPT-3. but it also describes Claude Fable 5. nice progress, shill
>b-b-b-but look what Generic Tech Boomer #18289 said
Luddites BTFO.
apparently the schizo spammer hates these:https://arxiv.org/html/2607.21551v1
>>17024849>can't integrate new information (aka no learning)Human-level AI with forward-amnesia would still be extremely useful. Just include the important information in the context.>can't form new concepts (aka no innovation)Nebulous definition. What is a truly "new concept" and how often do humans come up with one? Depending on your mood, it could describe 99% of humans. Yet economy still works with them.>hallucinates (aka no reliability)Humans also sometimes make mistakes. If the rate of mistakes is low enough, it does not present issues in practice.>falls apart at out-of-distribution tasks (aka no true generalization)See "can't form new concepts" rebuttal. This is the same point, just restated.>can't generalize reasoning beyond specific contexts (aka can't reason)See "can't form new concepts" rebuttal. This is the same point, just restated.>evaluates high-perplexity schizobabble the same as high-perplexity works of genius (aka no understanding)Wrong. Modern models don't do this.
>>17024870>it's human-level, it just can't learnOk, schizo.>b-b-b-but what TRULY IS a new concept??I accept your concession.>Humans also sometimes make mistakeThis doesn't relate to anything in my post. You're hallucinating.>See "can't form new concepts" rebuttalAbsolute psychotic delusion. You can't even pretend those things aren't properly defined. They're precise technical claims.>Wrong. Modern models don't do this.They do, by definition, but you're psychotically ill, so you can't even comprehend what you're reading.
>>17024886Define a "new concept", in a way that would exclude the possibility of AI creating it but include the possibility of average human creating it.
>>17024893>Define a "new concept"There is no need to. That LLMs can't form new concepts is not under dispute anywhere except in your padded cell.
>>17024894So you cannot define it properly. I accept your concession.
>>17024896I accept that your psychotic illness is causing you to deny that LLMs have an internal representation for the concepts it picks up from the training data.
>>17024902>it picks up "It" being whatever variant of GD is used in training, naturally.
>>17024902Humans also have an internal representation for the concepts they pick up from their training data. This fact does not mean the system cannot be intelligent or creative.
>>17024908>b-b-b-b-but humans alsoAnd? You're literally an automaton at this point. You're not understanding what you're reading and you're certainly not forming any actual thoughts about it.
>>17024909Comparing artificial intelligence to natural intelligence is obviously useful. You cannot say that AI cannot be intelligent because it has learned internal representations, or because it sometimes makes mistakes, when humans do the same things, yet they are intelligent. You seem to have higher standards for AI than for humans. AGI is defined as human level intelligence, that is, sometimes faulty and imperfect. If you are looking for hypothetical perfect intelligence, that is ASI.
>>17024914>completely generic marketing cope>zero relation to anything in the preceding discussionThis is just bizarre. These "people" are simply not human anymore, if they ever were.
>>17024915You only post ad hominems without any actual arguments whenever I press you on any specifics of your points. This discussion is pointless. AI derangement syndrome is a mental illness. Seek help.
>>17024919My argument stands unchallenged: your retarded attempt to argue semantics falls flat on its face because no one in the AI field disputes the LLMs have an internal representation for the concepts in the training data, no matter how you define them.
>>17024921For the last time: I am not disputing that LLMs have an internal representation for the concepts in the training data. I am disputing that this somehow presents a problem for them being intelligent or creative, or achieving comparable intelligence to humans. Any intelligent system, natural or artificial, must have some internal representation of the things it has learned.You are either purposefuly obtuse, or dumb as fuck, if you don't get my very simple point.
>>17024930>I am not disputing that LLMs have an internal representation for the concepts in the training dataA representation which you've already acknowledged is static. So how can it form new concepts, you shit-brown retard?
>>17024933Do you think that every time you form a "new concept" (whatever that means), your neural network architecture changes?Weights in an artificial neural network are comparable to synapses in the biological brain. Synaptic connections in the brain do not meaningfully change on the order of seconds or minutes, they only change when long-term memory sets in. You are confusing creativity with long-term memory.You are essentially saying that someone with forward-amnesia (inability to form new long-term memories) cannot be intelligent or creative. That's ridiculous.
>>17024945My "neural architecture" doesn't change but my brain obviously does change. You really are mentally ill beyond hope. >Weights in an artificial neural network are comparable to synapses in the biological brain.They're not. For one, they're static, which prevents the LLM from forming new concepts. You have a profound mental disability. Have you always been this way? Or is this the result of using "AI"?
