Gary Marcus

Rank 19 of 47
|
Score 64
1/16/2025, 1:08:15 AM
X
In reply to:
MindMirror
@TheMysteryDrop
·
176d
@GaryMarcus @sama I’m not claiming that you’ve pivoted your position but simply that the vast majority of people who read the phrase “AI is hitting a wall” intuitively assume that progress is slowing, which is clearly false and what @sama is annoyed by.
Gary Marcus
@GaryMarcus
·
176d
@TheMysteryDrop @sama read the paper where i proposed it, nothing has changed, reliability was a central part of the argument then, still is.
MindMirror
@TheMysteryDrop
·
176d
@GaryMarcus @sama So the “wall” is now reliability and not raw capabilities?

Feels like the intuitive way to interpret “hitting a wall” is that capabilities are slowing, which is clearly false if you look at o1/o3.

Unless your argument is that reliability is the intrinsic limiter.
Gary Marcus
@GaryMarcus
·
176d
Dear @sama,

Say what you said to Adam Grant it to my face. Debate me.

I coined the phrase “deep learning is hitting a wall” and I am absolutely *not* telling people not to think about AI. What you said is a distortion.

What am I actually doing is warning people about the…

The conversation involves a debate on the progress and limitations of AI, specifically addressing the interpretation of the phrase 'AI is hitting a wall.' The discussion is technical and involves differing opinions on AI's capabilities and reliability.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not cause harm and is part of a technical debate. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes understanding by clarifying the author's position on AI progress. [+1]
  3. 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 by addressing misconceptions and clarifying the author's argument. [+1]