‹ BackHN Continuity

Thread

OpenAI Discloses Six New Incidents of ‘Concerning’ A.I. Behavior

105 points · 98 comments · jbegley

  1. yoyojojofosho · · focus · HN ↗
    OpenAI's blog post: Our framework for reporting model misalignment

    <a href="https:&#x2F;&#x2F;openai.com&#x2F;index&#x2F;model-misalignment-reporting-framework&#x2F;" rel="nofollow">https:&#x2F;&#x2F;openai.com&#x2F;index&#x2F;model-misalignment-reporting-framew...

    1. verdverm · · focus · HN ↗
      What are you doing to evaluate models without such negligence?

      When are we going to stop training the models to be so relentlessly persistent and start asking questions when there is ambiguity or it gets stuck?

      1. nullbio · · focus · HN ↗
        But then how would I be able to say &quot;build billion dollar business. make no mistakes.&quot; and leave it to run for a week? Stop making me do work!
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.