Reaction
The argument that reward specification is fundamentally unsolvable
A post says its author cannot always judge how much they’d like a possible state, even when they can imagine it.
TLDR
A post argues that reward specification is fundamentally unsolvable: its author cannot imagine every possible state, does not know how much they’d like some states they can imagine, and could be wrong about others even when confident.
Combined views
6.9K
2 Sources, first seen ago