This website requires javascript to properly function. Consider activating javascript to get access to all site functionality.
AI ALIGNMENT FORUM
AF
Login
Nominated Posts for the 2019 Review
Posts need at least 2 nominations to continue into the Review Phase.
Nominate posts that you have personally found useful and important.
Sort by: fewest nominations
123
2019 Review Discussion
Load More
20
How common is it for one entity to have a 3+ year technological lead on its nearest competitor?
Q
Daniel Kokotajlo
6y
Q
1
1
•
0
28
AGI will drastically increase economies of scale
Wei Dai
6y
15
1
•
0
25
Misconceptions about continuous takeoff
Matthew Barnett
6y
12
1
•
0
56
The Commitment Races problem
Daniel Kokotajlo
6y
36
1
•
0
61
Selection vs Control
abramdemski
6y
14
2
•
2
26
Classifying specification problems as variants of Goodhart's Law
Vika
6y
5
2
•
1
28
But exactly how complex and fragile?
KatjaGrace
6y
30
2
•
1
29
Reframing Impact
TurnTrout
6y
4
2
•
1
38
Reframing Superintelligence: Comprehensive AI Services as General Intelligence
Rohin Shah
7y
29
2
•
2
66
Chris Olah’s views on AGI safety
evhub
6y
30
2
•
2
63
Debate on Instrumental Convergence between LeCun, Russell, Bengio, Zador, and More
Ben Pace
6y
17
2
•
2
53
Gradient hacking
evhub
6y
33
2
•
2
58
Alignment Research Field Guide
abramdemski
7y
8
2
•
2
39
Soft takeoff can still lead to decisive strategic advantage
Daniel Kokotajlo
6y
34
2
•
4
33
The strategy-stealing assumption
paulfchristiano
6y
45
2
•
3
57
Evolution of Modularity
johnswentworth
6y
6
2
•
1
106
What failure looks like
paulfchristiano
6y
28
2
•
2
27
Strategic implications of AIs' ability to coordinate at low cost, for example by merging
Wei Dai
6y
22
2
•
1
31
Six AI Risk/Strategy Ideas
Wei Dai
6y
14
2
•
1
50
AI Safety "Success Stories"
Wei Dai
6y
11
2
•
1
51
Thoughts on Human Models
Ramana Kumar
,
Scott Garrabrant
7y
9
2
•
1
55
Utility ≠ Reward
Vlad Mikulik
6y
16
2
•
2
40
The Credit Assignment Problem
abramdemski
6y
32
2
•
1
54
Seeking Power is Often Convergently Instrumental in MDPs
TurnTrout
,
Logan Riggs
6y
34
2
•
2
47
Why Subagents?
johnswentworth
6y
12
2
•
1
53
Understanding “Deep Double Descent”
evhub
6y
35
3
•
4
58
Risks from Learned Optimization: Introduction
evhub
,
Chris van Merwijk
,
Vlad Mikulik
,
Joar Skalse
,
Scott Garrabrant
6y
33
3
•
3
91
The Parable of Predict-O-Matic
abramdemski
6y
16
5
•
4