AI represents an asymmetry attack. It's now trivial to generate enormous amounts of possibly true information. It's possibly true because an actual conscious mind needs to go through and try to refute the proof. It's asymmetric because the AI can generate a thousand of these a second, but your mind can only process truth and fact tables at a fixed rate, allowing us to bog down minds as necessary.

>>16988700
>It's possibly true because an actual conscious mind needs to go through and try to refute the proof.
False. It's definitely true that a conscious mind should ignore any AI outputs unless the "author" provides overwhelming and easily verifiable traditional proof for validity of the results.

>>16988730
It's easy to ignore proofs that are obviously AI outputs, especially those that admit to AI use or are published as anon. The danger comes when people start submitting AI proofs under legitimate or quasi-legitimate credentials. There've already been dozens of high profile retractions and it's only going to get worse. Journals are going to have to start devoting more and more manpower and time and effort towards going through blatantly false generated papers.

Arxiv or snarxiv exposed this years before ai.

>>16988758
>The danger comes when people start submitting AI proofs under legitimate or quasi-legitimate credentials.
If someone with legitimate credentials submits a proof, it presumably holds up under his scrutiny regardless of its origin, so it shouldn't be any more wrong or incompetent than a proof he would've produced by hand.

>>16988695
it doesn't proof that humans are magically special and have unique access to the sky castle with the magical book of all math so it's wrong by construction

>>16988700
The Lean proof is supposed to fix that. Of course one has to ignore any accompanying descriptive text and read the Lean description of the theorem only. It's common to see slop where mundane, obvious theorems are presented as having a more profound meaning in accompanying slop text, or to see theorems "proven" by making a bazillion extra assumptions.

>>16988770
That's an appeal to authority fallacy - just because someone has legitimate credentials doesn't mean they can't knowingly or unknowingly submit a bad proof, to say nothing of knowingly or unknowingly submitting a bad proof generated by an LLM as one's own original work.

Why should AI generated papers be subject to any less scrutiny than a paper produced without use of generative AI? If anything the level of scrutiny should be greater.

>>16988787
>That's an appeal to authority fallacy
How come anyone who whips out these fake reddit "fallacies" always turns out to be a mouth-breathing cretin who doesn't even know what any of them mean?

>>16988787
ai is cancer, but you're also a retard. expertise is different from appeal to authority. if einstein had used ai to assist in his general theory of relativity, we should accept his expertise to validate it. conversely, if (You) were to use ai to submit some fag theory like op, that's an appeal to authority. fuck off and delete your internet, dumbfuck

>>16988796
> we should accept his expertise
Now this is an actual appeal to authority fallacy but also has nothing whatsoever to do with my post.

So nowdays random joe fucks who lacks any training in mathematics try to prove big theorems? The revolt of the herd

>>16988700
Gibberish
>>16988758
>"Journals are going to have to start devoting more and more manpower and time and effort towards going through blatantly false generated papers."
Lol.
>Peer review is gonna fix it.
Retard

Has there ever been a non-mentally-ill namefag?

>>16988784
just plug my files into ShatGPT if you don't get it. there are maybe 50-100 people in the world who can code lean like this and none of them are talking to the 2-3 uni compartments

>>16988805
shut the fuck up bot

>>16988817
retard

>>16988820
>be me, vibe coder
>drop a theorem from 2030
>misquote on macrame origami website

>>16988758
>Journals are going to have to start devoting more and more manpower and time
They won't. Journals are mostly for profit or have people pay up to be in one. They don't care.

>>16988787
If a famed guy posts AI slip it's his neck on line.

>>16988796
>if einstein had used ai to assist in his general theory of relativity, we should accept his expertise to validate it
you're assuming he'd validate it.

>>16989017
you obviously know nothing of einstein's work ethic. fuck off.

File: opus.png (16 KB, 760x153)

16 KB PNG

>>16988695

File: bass_ihara.png (15 KB, 705x224)

15 KB PNG

>>16989499

>>16988796
>we should accept his expertise to validate it
nope. *we ourselves* should validate it. the motto is "who the fuck is einstein?"