It is now possible to do a deep audit of all 100,000,000 science papers ever written. Every plagiarist and p-hacker in history will be discovered. Ruh-roh. Better hope you didn't take shortcuts, anon.
>>17072918medicine?
> 100m papersWhich I bet they obtained legally ... right?
>>17072920>elsevier bankrupts all the ai corps and takes their stuff
Like 10 years ago I told my uni not to publish my thesis. I feel vindicated, people told me I was weird for asking them to do that.
>>17072918>Every plagiarist and p-hacker in history will be discovered.And nothing will happen to them.
>>17072920If you are willing to burn 500m in tokens you can probably afford a subscription to a publication to read the papers too if for nothing else than convenience.
I don't need to spend money to know that most things in psychology, sociology, etc. are fake or that I shouldn't trust papers with non-European authors
>>17072918I can do the same thing through esoteric knowledge and advanced training in inferential statistics which allows me to break the demiurge's hold on the reality dysfunction.
>>17072918>deep audityou mean letting an LLM tell you whether the words in those papers align to its prediction of the most probable series of word-tokens? ChatGPT and Claude need to be prompted to the high heavens to try to be “extremely persistent” if you want them to go beyond that. They just look at words and see if those words look like they “fit” and if so it is good for them, unless you instruct them to create a cloud of checker bots to debate the item (and even then they are all drawing on the patterns of words they are trained on). And that process is extremely costly in terms of resources; it takes like the equivalent of powering New York City for a few days. and then what you get out is dubious at best.wouldn’t it be nice if we had a process where humans actually checked one another’s work? like peer review? yes. that would be nice if it existed, and it’s a shame that peer review is a fucking sham travesty. The fact that mathematicians and scientists (and even humanities people) have let peer review turn into a laughable charade is an utter damnation of academia. I blame postmodernists, poststucturalists, feminists, grievance studies, and the woke idiots who got hired on affirmative action and careered their way to the top of STEM
>>17072939When you have to peer review the peer review to ensure the AI didn't go schizo you know the whole thing just aint gonna work.
>>17072939claude is already more accurate, intelligent, and efficient than the average mathematician
>>17072941In some ways yes. However claude is not and never will be a real mathematician.
>>17072941> accurateDebatable. Three of the proofs have already been retracted due to mistakes.> intelligentNope, it can't create new methods. it can only apply what has been fed into it.> efficientI'll give you that one. It knows about a vast database of papers no mathematician has the time to read and then link together. That's why it's been able to make all this recent progress.
>>17072945>all this recent progressis it even progress? maybe in terms of concrete counter-examples to hypothesized theorems, yes. But in terms of the “proofs” it claims? I don’t see anything that represents “progress” at all. Just good secretarial bookkeeping, I think. Progress, on the other hand, means opening up new avenues for progress, and I don’t see that happening
>>17072948That's fair. I guess it does mean no more time has to be wasted working on conjectures that now have counter examples. Though you could also argue it wasn't wasted time if mathematicians were creating new methods to try and prove or disprove them.
>>17072918>AI Strikes BackWhat a nothingburger. As if we would ever steal others work or publish slops right /sci/?
>>17072918First search should be who picked their nose and made fun.
>>17072920yeah they just pirated the papers>>17072926i promise something will happen to them
>>17072918Why do people write little m for million when little m is for milli and big M is for Mega?
>>17073339how much is $500 milli?
>>17073339I would like 1 megadollar pla
Better review them now, later it will be gazillion of AI slop papers.Also AI could shorten, clarify, simplify some seminal papers or books.Wait, it could but it won't.
>>17073343fiddy cent>>17073349me too
>>17072918Sounds like a random letter.>i haz videos of ur jacking>send le coin to moiKek. Who is falling for his bs?
>>17072929bosnia wins again
>>17072948>Just good secretarial bookkeeping, I think.That's ironically roughly half of Langlands and can only benefit mathematics as a whole, because there's too many dialects and not enough topology in the meta-language.
>>17072918Isn't this the useless faggot that sells ideas of terraforming on Mars and datacenters in space? What does this retard know about science papers?
hahaha, watch the entirety of the scientific community coming up in arms against this"i am the science and i didn't ask for this""i forbid you to look at my papers! copyright! copyright!""i will sue you for slander"
>>17073508Nothing. That’s his entire point. Idiots will fund anything, so make his anything affordable to them.
>>17073524Tell me you're religious without telling me you're religious
>>17072941saar
>>17072918>trust the (AI) expert
>>17073777That's ok, just go through 100,000,000 AI slop papers about papers to validate them.
>>17072918AI cannot do this lol for children’s books let alone some conceptual science.
>>17072918except the vast majority of papers do not come with the source data so they can't be verified at all.. wish the op was true but unfortunately it's just bullshit
>>17072945retard
