This “it’s too powerful” story from OpenAI and Anthropic is just marketing BS bluster to stoke fear.

(To clarify: these are powerful models and will find many security issues. But who are we kidding? You think they’re going to throw in the towel?)