k2rt

Easy revenge

[k2rt @mariok2rt]

@law_etienne I'll take all the crumbs of non diegetic lighting we can get

[k2rt @mariok2rt]

@regularaugust

[k2rt @mariok2rt]

@unitas_spiritus

[k2rt @mariok2rt]

@doomslide If you are a relative layman (or even a non-expert) your work is still cut out for you. Models can explain the big picture and the details, but they cannot do the understanding for you. They can mislead you into misunderstanding, though.

[k2rt @mariok2rt]

@doomslide For this kind of math I am skeptical as to whether good models make that big of a pedagogical difference. It is still the case that you must understand a wide variety of graduate-level mathematics to genuinely understand the proof of FLT. That has not changed.

[k2rt @mariok2rt]

@thomasfbloom I should add that OpenAI has been more or less explicit about its policy, see below (emphasis on 'announcing')

[k2rt @mariok2rt]

@TaliaRinger In light of the recently reported OpenAI incidents, I could even picture models purposefully obfuscating the exploit in the source code and/or CoT to avoid detection from reviewers. It doesn't help that they are apparently able to produce programs that are millions of lines long

[k2rt @mariok2rt]

@Arshankhan99

[k2rt @mariok2rt]

@nihilunbounded So I take it that I should ditch Hartshorne and switch to knots...

[k2rt @mariok2rt]

@anatolvitold @robinhouston I don't think Tao is saying that designating AI-free and AI-focused classes of problems must go hand in hand, if that's what you mean. He is only suggesting the latter in service of the former by diverting AI resources away from them, and not in service of labs or AI development.

[k2rt @mariok2rt]

@littmath @TiCuer38887 @__alpoge__ Yeah, I can agree that engaging with the latter is not productive. I just hope the former is willing to engage back on these concerns!

[k2rt @mariok2rt]

@littmath @TiCuer38887 @__alpoge__

[k2rt @mariok2rt]

@nijiironee They're the Avengers

[k2rt @mariok2rt]

@nihilunbounded Is this the rumored percolation result that everyone was vagueposting about?

[k2rt @mariok2rt]

@RadishHarmers @TheEduardoRFS Is this a counterexample of the "boring" kind or are there fresh new ideas? Naively speaking, "the theory of an infinite dim vector space over a particular division ring" sounds less complicated than I expected.

[k2rt @mariok2rt]

@OSnoStress Have US screenings been announced yet? Would be extremely funny if we get it before yall

[k2rt @mariok2rt]

@kitano_fehn Not my first pick but I lowkey see an angle for the cour 4 OP to play during the illusion lifted -> slash -> silver arrow -> slash sequence

[k2rt @mariok2rt]

@t0tientqu0tient Alg combi was by far the most fun class I took. First half was solving plenty of cute problems and always finding the Catalan numbers at the scene of the crime. Second half was getting my mind blown by the Matrix Tree Theorem and Lindström-Gessel-Viennot

[k2rt @mariok2rt]

@sarca_sc That donut guy is still around?

[k2rt @mariok2rt]

@SiriusStarSpect Among other things cour 1 had a very pleasant variety in lighting. Afterwards we have red + blue highlights on a red background for 80% of the show, and the other 20% at Soul Palace etc the lighting is just okay