@ben_j_todd Shane Legg, DeepMind cofounder, in 2009: https://t.co/siTeqFXpLp
See https://t.co/O3lg2KGQgV for a prediction around "proto-AGI" that has largely happened.
(Also see https://t.co/MmD9aCKK5Y for thread w/ content mirrored)
We congratulate Levent Alpöge and Tristan Buckmaster on their remarkable mathematical work.
We (the researchers and the agents) did not see any of their work through any means until they released it publicly — in particular, no specific user data was accessed in order to solve this problem.
While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models.
However, our proofs differ significantly and even the precise results proved are different in the Euler case (forced vs. unforced).
(It is also lightly amusing to me how some of the firms most trying to use biology/health to help their public PR position also have the most bioweapons doomers.
…but this is all enough meandering, vague musings for one day)
Not saying it’ll definitely happen, but (as a non-biologist) it definitely _feels_ like a comparable difficulty level to getting say, cures for most cancer, extending lifespan, etc
As if my life isn't already absurd enough... the cruise ship with the hantavirus is the same one I went to Antarctica on in November 2023.
(It was a great trip with excellent crew and staff; unfortunate that's all happening.
My hypothesis about it would be penguin guano.)
Not naming names on Twitter cause this isn't a case where public mob attention is useful. [__] is the only firm to have done this out of the 7-or-so places I've been chatting with... 🙃
I'm at my 3rd or 4th schedule change for interviews with [__], this time after already sitting in an interview video call for 10 minutes.
Come on [__]. Get your logistics together. :)
@AI_richard …does the one with John Shulman also include lightly crass references to hyperfixations and development ages?
(The Leopold one was what I was thinking of, though need to get around to doing some long runs and getting through those too.)
@CJHandmer On a more pragmatic note, “receiving (wanted) messages” is likely very important for retention. I wouldn’t be surprised if adding friction to the messaging flow (even if it’s just an innocuous calculation of “is this message worth the price”) has some unintended consequences.
@CJHandmer Among other things — I’ve had folks reach out on LinkedIn “because who cares about the silly rules of these dating apps” when I very intentionally stopped responding; I’m currently dealing with a bit of a stalker, etc