yababa_y

I used a similar methodology for https://github.com/emberian/don feel free to dig thru anything in there I think all the notes are in the repo itself... of course it helps having a game that was explicitly designed around networked deterministic lockstep. forces the agent to learn a lot of lessons from reality instead of my typing :)

Comment · 1 month ago

in the PDF metadata we find the proper and appropriate title of this work:

    Nearly SNFS-Speed Signature Forgery Sans Factoring N (NSNFSSSFSFN)

Comment · 1 month ago

I will admit that when I heard they only had forced blowup I went back to sleep (but I've never been more that cursorily interested in analysis).

Comment · 4 months ago

I want Opus to be only marginally better, but I do mostly research engineering and its ability to not fuck up my projects is absent. Every time my credits lapse I let kimi and composer2.5 have some play and it’s basically just an excuse for me to keep playing computer because when the oai/ant credits refresh I always need to spend hours recovering from the other models either misconceptions or boneheaded eng practices. Even when I only let it touch my web games…

Comment · 4 months ago

separately trained experts can surpass performance in their activated regime and DOES result in a smarter model, the Claude system cards talk about this and eg there is https://openreview.net/forum?id=iydmH9boLb to read...

Comment · 5 months ago

There's only so much to do and coordination costs (already burdensome) become overwhelming.

Comment · 7 months ago

i've been missing this knowledge! thank you for the recommendation.

Comment · 8 months ago

local laws forbidding facial recognition tech have never been wiser

Comment · 9 months ago

I was somewhat surprised to learn that phi is _merely_ (1 + √5)/2, I didn't have a good conception of what it was at all but I didn't think it was algebraic.

Comment · 1 year ago

well, your loss then. clearly your work steps aren’t big enough to benefit from SoA LLM