Anthropic continues to deliver world class work.

You’re too dangerous with Fable. Back to Opus !

Using other model to write code which is in direction of intention, it will be bad , then asking fable to review that code , and suggest improvements and after that asking it to patch what it said maybe that works , didn't trigger safeguards that way , probably that can work

Are you just 100% claude code max? Or some other setup these days?