x
Limiting an AGI's Context Temporally — LessWrong