The more someone believes in the dangers of ai-alignment, the less faith they should have that it can be solved.