I’d hope the AI misalignment researchers are on this one, since the model is basically misaligned with the interests of everyone (excl. Anthropic) including the user Oh wait they are trying to intentionally train this into the model?!