x
This website requires javascript to properly function. Consider activating javascript to get access to all site functionality.
AI ALIGNMENT FORUM
AF
Login
Subagents and impact measures — AI Alignment Forum
Subagents and impact measures
Looking at the ways that subagents can get around impact measures.
18
Subagents and impact measures, full and fully illustrated
Stuart_Armstrong
6y
2
28
Appendix: how a subagent could get powerful
Stuart_Armstrong
6y
10
7
Appendix: mathematics of indexical impact measures
Stuart_Armstrong
6y
0
18
Dynamic inconsistency of the inaction and initial state baseline
Stuart_Armstrong
5y
4
8
Why is the impact penalty time-inconsistent?
Stuart_Armstrong
5y
0