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 this jackals are going to throw in the towel?)