x
Improving The Akrasia Hypothesis — AI Alignment Forum