Value Alignment

Human Compatible

The Policy: Coherent Extrapolated Volition and the Paradox of Perfect Alignment

Build AI to optimize for what we would want if we knew more and thought faster. Beautiful in theory. What if we don't actually want what our better selves would want?