Go Back
Choose this branch
You can't go any further
meritocratic
regular
democratic
hot
top
alive
29 posts
Corrigibility
Instrumental Convergence
Treacherous Turn
Programming
2017-2019 AI Alignment Prize
LessWrong Event Transcripts
Satisficer
Petrov Day
35 posts
AI Risk
183
Debate on Instrumental Convergence between LeCun, Russell, Bengio, Zador, and More
Ben Pace
3y
60
131
Let's See You Write That Corrigibility Tag
Eliezer Yudkowsky
6mo
67
123
Seeking Power is Often Convergently Instrumental in MDPs
TurnTrout
3y
38
101
Announcement: AI alignment prize round 3 winners and next round
cousin_it
4y
7
90
Announcement: AI alignment prize round 4 winners
cousin_it
3y
41
82
Environmental Structure Can Cause Instrumental Convergence
TurnTrout
1y
44
80
Satisficers Tend To Seek Power: Instrumental Convergence Via Retargetability
TurnTrout
1y
8
70
A Gym Gridworld Environment for the Treacherous Turn
Michaël Trazzi
4y
9
60
You can still fetch the coffee today if you're dead tomorrow
davidad
11d
15
55
Corrigibility
paulfchristiano
4y
7
51
Clarifying Power-Seeking and Instrumental Convergence
TurnTrout
3y
7
41
Can corrigibility be learned safely?
Wei_Dai
4y
115
40
Empowerment is (almost) All We Need
jacob_cannell
1mo
43
40
[Linkpost] Treacherous turns in the wild
Mark Xu
1y
6
462
AGI Ruin: A List of Lethalities
Eliezer Yudkowsky
6mo
653
243
Counterarguments to the basic AI x-risk case
KatjaGrace
2mo
122
147
Worlds Where Iterative Design Fails
johnswentworth
3mo
26
135
AGI ruin scenarios are likely (and disjunctive)
So8res
4mo
37
124
AI Safety "Success Stories"
Wei_Dai
3y
27
113
Niceness is unnatural
So8res
2mo
18
108
The Main Sources of AI Risk?
Daniel Kokotajlo
3y
25
107
What can the principal-agent literature tell us about AI risk?
Alexis Carlier
2y
31
98
AI will change the world, but won’t take it over by playing “3-dimensional chess”.
boazbarak
28d
86
93
The alignment problem from a deep learning perspective
Richard_Ngo
4mo
13
91
And the AI would have got away with it too, if...
Stuart_Armstrong
3y
7
83
What does it mean for an AGI to be 'safe'?
So8res
2mo
32
83
The strategy-stealing assumption
paulfchristiano
3y
46
79
Oversight Misses 100% of Thoughts The AI Does Not Think
johnswentworth
4mo
49