What just happened? Pragmatism and Pessimization

A former alignment researcher is arguing that the boundary between AI safety and AI capabilities has effectively vanished. As LessWrong tells it, some researchers used the banner of safety to justify work that actually accelerated the development of frontier models at companies like OpenAI, DeepMind, and Anthropic. The author claims this happened because the community failed to challenge powerful figures and allowed consequentialist arguments to mask a drive for power. He suggests that the field must now adopt more honest public conversations and stricter accountability to move past this failed paradigm.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton