Kart ographien(@kartographien) 's Twitter Profileg
Kart ographien

@kartographien

mostly AI safety

ID:1217527755128168448

calendar_today15-01-2020 19:23:44

2,2K Tweets

1,1K Followers

2,3K Following

Kart ographien(@kartographien) 's Twitter Profile Photo

me at a party in year 2065:

yeah, I supported EA when they focused on post-Turing eigenwelfare, but then they were hijacked by interdimensional hypershrimp advocacy

account_circle
Kart ographien(@kartographien) 's Twitter Profile Photo

I can’t relate to the kind of person who doesn’t decide that Alice is lying.

what would you do if Alice claimed 100 heads and Bob claims a 100-long incompressible sequence?

or with n=1000?

how large must n be for you to snap out of your “b-but the maths says…” mindset?

account_circle
Daniel Litt(@littmath) 's Twitter Profile Photo

NOT a math puzzle: you are imprisoned by two guards—Alice and Bob. One always lies; the other always tells the truth. Each flips a fair coin 10 times. Bob reports the sequence HTHHTTTHHT, and Alice reports that she got all heads. Who is most likely the liar?

account_circle
Marc Andreessen 🇺🇸(@pmarca) 's Twitter Profile Photo

All right, look. Odds of one 'heads' = 0.5. Odds of two 'heads' = 0.25. Odds of three 'heads' = 0.125. And so forth. By the time you get to 9 'heads', odds of an additional 'head' is ~0.

account_circle
google bard(@deepfates) 's Twitter Profile Photo

Nice argument. Unfortunately i have already depicted you as the cybernetic crying wojak and myself as the superintelligent chad

account_circle
Kart ographien(@kartographien) 's Twitter Profile Photo

how does one classify a research paper as safety vs capabilities?

which papers do you think are borderline?

i agree with Zvi Mowshowitz’s sentiment but I’m not sure it’s as clear-cut in practice

account_circle
Kart ographien(@kartographien) 's Twitter Profile Photo

why does everyone act so high-and-mighty about cartesian dualism

it’s as reasonable as any other account of the mind-body problem

account_circle
Leo Gao(@nabla_theta) 's Twitter Profile Photo

Eliezer Yudkowsky ⏹️ while computers may excel at soft skills like creativity and emotional understanding, they will never match human ability at dispassionate, mechanical reasoning

account_circle
will depue(@willdepue) 's Twitter Profile Photo

i’m going to go crazy if i hear “it’s just interpolating” again. like homie when it plays 2000 ELO chess is it just interpolating all the chess games it’s ever seen? not at all

account_circle