OpenAI's O1 (Strawberry) AI: Time to Think = Ability to Reason

Поділитися
Вставка
  • Опубліковано 6 лис 2024

КОМЕНТАРІ • 6

  • @ВладиславМальчевський

    Great video

  • @Guyverman01
    @Guyverman01 Місяць тому

    Could the next module be called O2?

  • @TheEtrepreneur
    @TheEtrepreneur Місяць тому +1

    with the hidden CoTs they'll extract the most valuable input, the human spark, given by God to us. No bueno.

  • @KrisAstro
    @KrisAstro Місяць тому

    Cool vision, but a bit too optimisti.c AI agents aren’t yet capable of autonomously handling complex tasks -- most still need human oversight, specificaly for edge cases. More concrete examples of where "human-in-the-loop" (HITL) is essential would add some much-needed balance. Let’s not skip over the messy present.
    (PS) I'm curious how you address it in the Ability - - Customer becomes the HITL?

  • @patrickmcguinness1363
    @patrickmcguinness1363 Місяць тому +1

    No the problem isnt solved. This is progress but ials not the only way and may not be the best. The opaque AI model isnt telling us the details. We will need to wait for an open AI model equivalent to understand where we are.

  • @redonebig88
    @redonebig88 Місяць тому

    Nuh it is bullshit, I gave it code I have designed for simulating physics of electrolyser, I am getting this response "Your request was flagged as potentially violating our usage policy. Please try again with a different prompt." how am I violating their policy