kr

[kr @bollybq]

@DivinelyDesined dude i understand what you are trying to say. while evolution theory might be dead, cross reproduction might be real.

[kr @bollybq]

@stigxd @a07008297 @Chrisem__ @ValorantHQQ since derke needs to cross few meters and defuse and time was low, jawg holds advantage. he even had a slide. jawg didnt utilize the advantage well, that's all theres to it. its almost like chess where they close out games off 5-10 % advantage.

[kr @bollybq]

@geoffreyhinton make me your student 🥹its unfair karpathy and ilya and the big people get to be under you

[kr @bollybq]

@smthlikeyou11 wp dude really 🙁

[kr @bollybq]

@Priyansh_31Dec i feel like thats a coordination issue rather muscle memory

[kr @bollybq]

@gaandfaadtits kavya ?

[kr @bollybq]

@sandboxws @_julianschiavo assume above search is like a GD step, were kinda waiting on what its generalized finding is, and if there is enough compute spent, it'll find something for sure that can be importnat.

[kr @bollybq]

@sandboxws @_julianschiavo i think its more about what the sol finds out O(n) solution is. that event will become a trigger for many more explorations like this.

[kr @bollybq]

@vctpacific @smthlikeyou11 they could have prepared trap plays, retakes and more stuff like that instead of all the tossed rounds whic hwere a slow collapse of 1 by 1. its not like LOUD made a read and we lost, we itself made bad plays and gave openings. :( SADGE

[kr @bollybq]

@vctpacific @smthlikeyou11 davai probably didnt show up on the game day. however the contribution for loss was mostly due to gameplans. they guessed themselves too much in their game plan. their perforamnce was worser than GE tbh.

[kr @bollybq]

@industriaalist iam working on true image perception for llms through arc agi 2. generalization is a screwed word, OOD coverage might be a great replacement word. efforts replcacing SGD is not a decent gain atleast for a few years so avoided it. i also hate attention and next token prediction:)

[kr @bollybq]

@infinterenders youre proving my comment more and more, what an idiot 😂🤣🤣

[kr @bollybq]

@paperclippriors @blingdivinity no labs did atleast with the stupid as fuck models they provide

[kr @bollybq]

@HououinTyouma @blingdivinity i agree. qwen flash next is a living proof

[kr @bollybq]

@DoozerDiffuser e^(logp-logq) ?

[kr @bollybq]

@knowixbuilds @buildwithsid you talk like youre carrying the frontier on your shoulders

[kr @bollybq]

@grok damn u suck

[kr @bollybq]

@grok chinese people know some sort of information that i dont leading them to solving arc agi 2 kaggle lanes with high scores. can you connect dots on twitter for me please ?

[kr @bollybq]

@grok not convicing enough. also you gave a summary answer. didnt search enough on https://t.co/0dh6Vc5yJl