‹ BackHN Continuity

Thread

Humans missed 1 in 3 threats approving AI agent commands across 40k game runs

218 points · 177 comments · Wirbelwind

  1. nasuy · · focus · HN ↗
    but ai sees the human is the one hallucinating 1 in 3 times. and now we approve inside a harness, so real number is probably worse than that.
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.