Gary Marcus

Rank 21 of 47
|
Score 55
5/10/2025, 1:58:25 AM
X
In reply to:
Daniel Litt
@littmath
·
29d
@GaryMarcus @Wasgo @polynoamial Most of these are claims about adoption of AI, which I agree will be slow. The only one that’s both precise enough to adjudicate and about capabilities is about video models and physics, where I’d bet against you.
Gary Marcus
@GaryMarcus
·
29d
@littmath @Wasgo @polynoamial Which claims that I made in this are likely to be wrong?

https://garymarcus.substack.com/…
Daniel Litt
@littmath
·
29d
@GaryMarcus @Wasgo @polynoamial I’ve put money on the line that specific strong claims about AI capabilities are overstated (see e.g. here: https://manifold.markets/TamayBe…). But I also think you’re substantially understating them.
Daniel Litt
@littmath
·
29d
@GaryMarcus @Wasgo @polynoamial I think it’s a stupid graph. Not sure why you’re asking me to comment on dumb ai hype when I’m discussing a specific (if slightly imprecise) claim about scaling.
Gary Marcus
@GaryMarcus
·
29d
@littmath @Wasgo @polynoamial So where do you fall here?
Vlad Haltakov
@haltakov
·
29d
Unpopular opinion - we are not seeing exponential growth in AI, but rather something more like an S-curve.

While an S-curve has places that look exponential it doesn't go on forever, but saturates. This is what we see with foundational models already.
Daniel Litt
@littmath
·
29d
@GaryMarcus @Wasgo @polynoamial I was out here pointing out it was BS.

The statement engages in a discussion about AI adoption and performance, particularly in business contexts. It suggests that while there was initial interest in GPT-4, its performance did not meet expectations, especially where errors were costly. This reflects on the broader discourse about AI capabilities and adoption rates.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not cause harm and provides a factual observation about AI adoption and performance. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    It contributes to understanding by discussing real-world implications of AI performance. [+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 dialogue by addressing specific issues with AI performance without personal attacks. [+1]