The Academic Field Pyramid - any point to encouraging broad but shallow AI risk engagement? — LessWrong