RSS

Stuart_Armstrong

Karma: 21,149 (LW), 336 (AF)
Page 1

And the AI would have got away with it too, if...

Stuart_Armstrong
22 May 2019 21:35 UTC
71 points
4 comments1 min readLW link

Re­search Agenda v0.9: Syn­the­sis­ing a hu­man’s prefer­ences into a util­ity function

Stuart_Armstrong
17 Jun 2019 17:46 UTC
49 points
10 comments32 min readLW link

Self-con­firm­ing pre­dic­tions can be ar­bi­trar­ily bad

Stuart_Armstrong
3 May 2019 11:34 UTC
43 points
11 comments5 min readLW link

Defeat­ing Good­hart and the “clos­est un­blocked strat­egy” problem

Stuart_Armstrong
3 Apr 2019 14:46 UTC
39 points
15 comments6 min readLW link

Nash equil­ibriums can be ar­bi­trar­ily bad

Stuart_Armstrong
1 May 2019 14:58 UTC
35 points
24 comments2 min readLW link

By de­fault, avoid am­bigu­ous dis­tant situations

Stuart_Armstrong
21 May 2019 14:48 UTC
31 points
15 comments3 min readLW link

For the past, in some ways only, we are moral degenerates

Stuart_Armstrong
7 Jun 2019 15:57 UTC
29 points
14 comments2 min readLW link

Be­ing wrong in ethics

Stuart_Armstrong
29 Mar 2019 11:28 UTC
22 points
0 comments3 min readLW link