‹ BackHN Continuity

Thread

Is sandboxing sufficient to contain rogue agents?

52 points · 99 comments · zdw

  1. chrisjj · · focus · HN ↗
    > these agent breakouts represent a serious and unforgivable breach of trust.

    Someone trusts OpenAI? Really?

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.