This website requires javascript to properly function. Consider activating javascript to get access to all site functionality.
AI ALIGNMENT FORUM
AF
Login
Recommended Sequences
Embedded Agency
by
Abram Demski
Iterated Amplification
by
Paul Christiano
Value Learning
by
Rohin Shah
AI Alignment Posts
22
Introducing the AI Alignment Forum (FAQ)
Oliver Habryka
,
Ben
,
Raymond Arnold
,
Jim Babcock
7mo
0
2
[Question]
Would an option to publish to AF users only be a useful feature?
ricraz
4h
0
31
Offer of collaboration and/or mentorship
Vanessa Kosoy
4d
0
5
Training human models is an unsolved problem
Charlie Steiner
10d
0
3
Not Deceiving the Evaluator
michaelcohen
12d
26
10
[AN #55] Regulatory markets and international standards as a means of ensuring beneficial AI
Rohin Shah
16d
0
5
Oracles, sequence predictors, and self-confirming predictions
Stuart Armstrong
17d
0
9
Self-confirming predictions can be arbitrarily bad
Stuart Armstrong
17d
0
6
Nash equilibriums can be arbitrarily bad
Stuart Armstrong
19d
2
10
[AN #54] Boxing a finite-horizon AI system to keep it unambitious
Rohin Shah
22d
0
Recent Discussion