Gary Marcus

Rank 13 of 47
|
Score 83
5/17/2025, 11:53:53 PM
X
In reply to:
Daniel Litt
@littmath
·
240d
@GaryMarcus @colin_fraser @MelMitchell1 @rao2z Ah, sorry I misunderstood. FWIW I think the article you linked is a reasonable example of my complaint—you link to a lot of papers (e.g. the 20x20 multiplication task) which show a mix of successes and failures in OOD tasks and pull out only the failures.
Gary Marcus
@GaryMarcus
·
240d
@littmath @colin_fraser @MelMitchell1 @rao2z NOOOO!

I did not ask you why MY writing is unpersuasive.

I asked you to look at large body of work and you blew it off and personalized it.
Daniel Litt
@littmath
·
240d
@GaryMarcus @colin_fraser @MelMitchell1 @rao2z As for “be better,” you asked why I find your writing on this unpersuasive. It’s because when you talk about subjects that I know a lot about, including my own commentary, your writing doesn’t match my understanding of the situation. Not sure how else I can convey my complaint.
Daniel Litt
@littmath
·
240d
@GaryMarcus @colin_fraser @MelMitchell1 @rao2z I replied about USAMO here. I just don’t understand your argument; why is poor performance on USAMO indicative of non-intelligence if they perform well on other OOD tasks?
Daniel Litt
@littmath
·
240d
@GaryMarcus @colin_fraser @MelMitchell1 @rao2z They’re hard proof questions, which are indeed difficult for frontier models. There’s nothing to explain.
Gary Marcus
@GaryMarcus
·
240d
@littmath @colin_fraser @MelMitchell1 @rao2z Dude. Aside from that being ad hominem and a bit insulting, it fails to address the USAMO result and the work by multiple authors. as such it is a pretty weak reply.

be better.
Daniel Litt
@littmath
·
240d
@GaryMarcus @colin_fraser @MelMitchell1 @rao2z I’m in a couple of things you’ve written, but the only things reported therein are my negative comments on AI capabilities. I also found your reporting of FrontierMath to be incomplete based on my own experience with it. These experiences have affected how I read your writing.

The statement is part of a broader discussion on the reliability of out-of-distribution (OOD) tasks in AI, which is a substantive engagement with public issues related to AI capabilities and research. The tone is assertive, aiming to clarify a point about the reliability of OOD tasks.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not cause harm directly but is part of a contentious debate, which can sometimes lead to misunderstandings or perceived slights.
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement does not actively promote understanding, empathy, or compassion, as it is more focused on asserting a point rather than fostering dialogue.
  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 is part of a dialogue with differing opinions but does not engage in personal attacks or ad hominem arguments. It maintains a focus on the topic rather than the individuals involved. [+1]