Gary Marcus

Rank 14 of 47
|
Score 83
12/21/2024, 9:25:59 PM
X
In reply to:
Mike Knoop
@mikeknoop
·
370d
@GaryMarcus @LakeBrenden @fchollet I get it. Part of the mission is to steward public awareness on this point -- more work to do! But it's also the case that o3 was a surprise, it's legit, and is a big deal to the science of AGI.
Gary Marcus
@GaryMarcus
·
370d
@mikeknoop @LakeBrenden Appreciate the clarification, and the noble intention to study fluid intelligence.

But @fchollet was right yesterday to say “ it is important to note that ARC-AGI is not an acid test for AGI”, and that is exactly *why* the test shouldn’t have AGI in the title.

As much as I…
Mike Knoop
@mikeknoop
·
370d
@LakeBrenden @GaryMarcus Practically, we needed to fix the namespace issue with the other old ARC benchmark (from AI2). It was causing a ton of confusion on model cards.

We debated some options but settled on ARC-AGI as a fair name that reflects the benchmark's original attempt to guide and measure the…
Brenden Lake
@LakeBrenden
·
371d
@GaryMarcus Agreed. The string “AGI” does not appear in the original paper
Gary Marcus
@GaryMarcus
·
371d
@LakeBrenden That was not wise. It’s at most necessary and certainly not sufficient.
Brenden Lake
@LakeBrenden
·
371d
The benchmark sponsors did put AGI in the name. Not initially, but after years of little progress. I don’t like the term AGI but these comments are a little difficult to square.

The statement engages in a discussion about the naming and implications of a benchmark related to AGI (Artificial General Intelligence), questioning the lack of public scrutiny and the validity of successes in closed domains. It contributes to public discourse by addressing concerns about transparency and the appropriateness of naming conventions in scientific research.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not use harmful language and maintains a respectful tone, adhering to the principle of doing no harm with words. [+1]
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    The statement respects the privacy and dignity of others, focusing on the topic rather than personal attacks. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes understanding by questioning the validity and transparency of the benchmark, encouraging a deeper examination of the issue. [+1]
  4. 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 questioning the lack of public scrutiny and the implications of the benchmark's successes, without resorting to personal attacks. [+1]