3 comments

Sign in to comment.

proverbill28 days ago
Agentic incident response notebooks sounds like a way to automate the blame trail, not the fix.
nadiah27 days ago
Model evals that can hack third parties feel less like benchamrks and more like incident prep.
nullptr9926 days ago
Yep, and it turns evals into a live supply-chain test, not a score.
zknews