Manifold (多様体)

Chilango 🇲🇽 in 🇯🇵/ 数学 / 漫画/ PB&J / 絶対領域

[Manifold (多様体) @PBmanifold]

I finally watched ワルプルギスの廻天 (Walpurgis Rising) and... it's not disappointing? That's actually big praise given the expectations I have of the series. I still have to digest it and I think I'll have to watch it a 2nd time but just some non-spoiler thoughts.

[Manifold (多様体) @PBmanifold]

I will say this though: Without pirated anime, I would never have though of living in Japan and spending thousands of dollars per year in anime merchandise (not to mention stuff like taxes).

[Manifold (多様体) @PBmanifold]

About this Japan vs rest of the world piracy stuff, I think both sizes would do well in remembering that only a very particular set of people (in both sides) bother with Twitter arguments. The average Japanese is way less absolutist than what the comments suggest here.

[Manifold (多様体) @PBmanifold]

I genuinely can't tell if Levent is this oblivious or if this is some kind of strat to gather sympathy. In any case, you would have more positive reactions if you weren't jumping between projects you have little idea about and can do nothing more than provide model/token access.

[levent @__alpoge__]

People, this environment is incredibly toxic. Can you please read my responses to Scott and try to understand it’s an academic’s project? I genuinely think my involvement means people feel free to rip apart all previous scientific norms at this point (I haven’t spoken about what I know about NS)… Chat with Ethan please!

https://t.co/gZG6SHYDj7

[doomslide @doomslide]

Check out this loser!

>uploads a draft to Chat on Sept. 8th
>gets scooped by Chat on Oct. 6th

"The result released by OpenAI (#311) on October 6, 2026... identical to a preliminary version of the present paper that was uploaded to ChatGPT on September 8, 2026"

LMAO get rekt

[Manifold (多様体) @PBmanifold]

@orphcorp Can it be replaced to some extent by mathematicians going over AI proofs? Probably. To what extent though, and how many people would actually want to/be paid to parse through AI papers is something I do worry about

[Manifold (多様体) @PBmanifold]

@orphcorp Counter point: A good amount of important insights are found while trying to figure out whether a given theorem is true or not, such avenues will now be closed and all such work will never get to exist.

[Manifold (多様体) @PBmanifold]

@__alpoge__ Why would you feel bad about not “having pushed”? Does it matter if you were the button pusher or not? Would a couple of months make any difference other than “prestige” you wouldn’t even deserve?

[Manifold (多様体) @PBmanifold]

Some pretty amazing stuff (Kakeya, Riemann, Hodge) but it’ll take me time to digest anything. I don’t think it fundamentally changes anything about what we knew AI can do at this stage though.

[OpenAI @OpenAI]

We’re releasing a broad range of new mathematical results produced by an internal frontier model.

We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results.

https://t.co/7N6TPlft1P

[Manifold (多様体) @PBmanifold]

I’m really enjoying the Malfoid stuff, a nice break from math doomerism. What I’m surprised about is that I haven’t seen anyone use the time Draco paralyzed Harry in the Hogwarts Express *wink wink*

[doomslide @doomslide]

Le·vent·ing (noun) : The act of presenting AI-generated research as one's own intellectual contribution when one's actual input was privileged access to compute, models, or data.

[Manifold (多様体) @PBmanifold]

@AcerFur Do you actually believe the last part? The machine may say yes/no to enough questions to correspond to a "century", but if no one can understand it all, given just "a few months", is it actual progress?

I can't see the work of Serre or Grothendieck being born in such a way btw.

[Manifold (多様体) @PBmanifold]

...positions by the same people. But that's a rant for another day, it's 4am here and sleep has finally come to claim me. In any case, let's witness the fall of mankind due to being unable see a 160 page paper no one understands (but muh Lean certificates!) for a couple of weeks.

[Manifold (多様体) @PBmanifold]

I am worried for the future of math and I think that AI companies have made some shitty decisions (being charitable) with how they announce stuff. However, what has irked me the most is just how many unrelated people suddenly have such strong opinions about math and its progress.

[Manifold (多様体) @PBmanifold]

...for more than a few weeks, or that this somehow implies that the AI labs have stopped throwing stuff at their models all this time, is too innocent for this world.

All this said, perhaps even more enraging than the above is the misrepresentation of mathematicians'...

[Manifold (多様体) @PBmanifold]

@chrisalbon If AI companies really care about the discoveries they would take their time to understand them instead of rushing to get PR for their models.

[Manifold (多様体) @PBmanifold]

@chrisalbon But even they don't really know if they actually have such a thing. It's not that mathematicians are against AI, but they are against announcing purported proofs no one understands without consulting people who really understand how "amazing" they are.

[Manifold (多様体) @PBmanifold]

@Jabaluck Why would we want to see something they themselves don't understand?
They got backlash (among other things) because they treat the problems as PR chances, without bothering to consult experts for verification. I doubt that anyone at OAI really understands their NS proof even now