Michael Zaal

[Michael Zaal @MichaelZaal8]

@Deniz_Genlik I think people should be more tolerant of mathematicians' eccentricities. I don't think he has bad intentions, unlike OpenAI, which is overtly combative.

[Michael Zaal @MichaelZaal8]

@AlexKontorovich I do think the ability to prove new results (either without AI or contributing to an AI-human collaboration) should matter for hiring, and that we shouldn't leave that completely to AI.

[Michael Zaal @MichaelZaal8]

@Polymarket I've heard that at any given time there are 100 women in the US who think they are in an online relationship with Johnny Depp. This case is just a drop in the bucket.

[Michael Zaal @MichaelZaal8]

@littmath It's the general issue with https://t.co/MogS3cF22L... anyone can post and catchy posts get amplified. Catchy is often fear-mongering or hyping "cool" things. There's no filter so emotions reign supreme over intelligence, in math and in other subjects.

[Michael Zaal @MichaelZaal8]

@ElliotGlazer They may have just tried a zillion times because it's so related to the Hodge Conjecture. Then the nth time they get a hit due to the sign error. The other papers could have then propagated this error.

[Michael Zaal @MichaelZaal8]

@dzackgarza He probably applies it to himself too, tbf.

[Michael Zaal @MichaelZaal8]

@lpachter @AlexKontorovich I don't know why you keep at this.. That which was correct in his thesis was nothing special. It was a short "consolation prize" of little mathematical relevance. His 7 yrs in grad school led to very little. There was no sign of great potential and no one owed him anything.

[Michael Zaal @MichaelZaal8]

@lpachter @dmitrykrachun No one really claims there was math of value in his thesis. I don't think a thoroughly flawed thesis merits a postdoc, just because others did not find his mistake. There was no evidence he could do serious math research, and no one is obligated to hire someone like that.

[Michael Zaal @MichaelZaal8]

@lpachter His thesis was not fine if it was based on a false premise and therefore incorrect. And basing a thesis on an advisor's paper which is provably false is not comparable to making it conditional on an unsolved conjecture, even if it later proves to be false.

[Michael Zaal @MichaelZaal8]

@ObhishekSaha It helps consolidate their control over mathematics and the reputation of their models. Right now these proofs are criticized for being incomprehensible and they are working to change this for the benefit of their company.

[Michael Zaal @MichaelZaal8]

@PI010101 lol using Trump's terminology to talk about AI in math. Finally we find the one research mathematician who follows his orders.

[Michael Zaal @MichaelZaal8]

@jdlichtman I personally think they should do the whole thing again to see if the same proof or a substantially different proof comes out of it.

[Michael Zaal @MichaelZaal8]

@mathandcobb @SebastienBubeck @nihilunbounded Well technically it's true, these aren't proofs.

[Michael Zaal @MichaelZaal8]

@Schrodinge95776 @hsu_steve He could point out that the proofs are long and incomprehensible, and might contain inaccuracies, and suggest working towards finding better proofs of these results. Or just simply say he'd have to shift research directions perhaps, but not in a panicky way.

[Michael Zaal @MichaelZaal8]

@AlexKontorovich Don't worry, these papers are pretty indigestible and it doesn't matter which one you are looking at.

[Michael Zaal @MichaelZaal8]

@PI010101 Ok, so I assume you'll digest these unformalized proofs for us?

[Michael Zaal @MichaelZaal8]

@jdlichtman These aren't formalized. How can we be confident in their results? Or can we?

[Michael Zaal @MichaelZaal8]

@slimer48484 How is it scooping his research if he's pasting theorems that Claude proved and not him?