The case for Doing Something Else (if Alignment is doomed) — AI Alignment Forum