Gary Marcus

Rank 20 of 47
|
Score 73
Vlad Haltakov
@haltakov
·
30d
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.
5/10/2025, 1:45:00 AM
X
In reply to:
Daniel Litt
@littmath
·
29d
@GaryMarcus @Wasgo @polynoamial I was out here pointing out it was BS.
Gary Marcus
@GaryMarcus
·
29d
@littmath @Wasgo @polynoamial Where were you when 1000s of influencers a day were chanting “exponential”?
Daniel Litt
@littmath
·
29d
@GaryMarcus @Wasgo @polynoamial This is all consistent with logarithmic growth, which is very slow!
Gary Marcus
@GaryMarcus
·
29d
@littmath @Wasgo @polynoamial I seriously doubt it at this point. All the labs have leaked that they aren’t getting the results they thought they would, and none of the new models cures the core problems that I have harped for decades.

Largish improvements now are from new techniques and focused data
Daniel Litt
@littmath
·
29d
@GaryMarcus @Wasgo @polynoamial Ok, I don’t think OpenAI’s naming scheme is really important here. I was trying to get to the bottom of the claim about scaling laws. As I understand it Altman etc. claim “intelligence scales logarithmically with compute.” Are you saying this is false? If so, why?
Gary Marcus
@GaryMarcus
·
29d
@littmath @Wasgo @polynoamial Some people are saying “[exponential] scaling continues, and it is just that benchmarks have saturated”, and I think this is a bad take.

My take is that exponentially increasing data and compute failed to solve the core problems (hallucinations and boneheaded errors) and nothing

The conversation involves a discussion on AI growth patterns, debating between exponential and logarithmic scaling. It touches on public discourse about AI development and its implications.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The discussion is technical and does not appear to cause harm. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The conversation promotes understanding by discussing different perspectives on AI growth. [+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 dialogue is constructive, focusing on technical aspects rather than personal attacks. [+1]