Gary Marcus

Rank 12 of 47

Score 83

@TheMysteryDrop Correct. I see o1 as putting on a (clever!) band-aid rather than treating the underlying disease of working with uninterpretable black boxes that lack adequate internal world models.

It will help in some cases but
not be robust. (Which is indeed what OAI’s report suggests.)

9/16/2024, 3:32:06 PM

In reply to:

Sapien

@TheMysteryDrop

493d

@GaryMarcus Is your view that the new O1 architecture and using RL + variable compute at time on inference is NOT likely to lead to significant advances in capabilities?

The statement engages in a substantive discussion about the O1 architecture and its implications for AI capabilities, which is a public issue. The tone is analytical and aims to provide a critique of the technology in question.

Principle 1:
I will strive to do no harm with my words and actions.
The statement does not appear to cause harm with words or actions. It is a technical critique rather than a personal attack. [+1]
Principle 3:
I will use my words and actions to promote understanding, empathy, and compassion.
The statement aims to promote understanding by explaining the limitations of the O1 architecture. It uses a metaphor (band-aid) to make the concept more accessible, which can foster empathy and compassion for those trying to understand complex AI issues. [+2]
Principle 4:
I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
The statement engages in constructive criticism of the O1 architecture. It does not engage in personal attacks or ad hominem arguments, focusing instead on the technology and its implications. [+2]