Gary Marcus

Rank 19 of 47
|
Score 64
7/25/2024, 10:52:41 PM
X
In reply to:
Leon Lang
@Lang__Leon
·
348d
@GaryMarcus To get more clarity on your position: if some company achieved AGI with a model whose whole creative backbone is an LLM, within five years from now, with an approach that is neurosymbolic similar to AlphaProof: Would you then think there were predictions you made that are wrong?
Gary Marcus
@GaryMarcus
·
348d
@Lang__Leon Not with these tools
Leon Lang
@Lang__Leon
·
348d
@GaryMarcus It seems like you’re then in large agreement with many proponents of LLMs. Yet, you seem to believe LLMs are largely hype and will not live up to their promise. Is the difference then that you believe companies won’t “in time” solve the more symbolic parts of the challenges?
Gary Marcus
@GaryMarcus
·
348d
@Lang__Leon It would be proof that we need a neurosymbolic approach and that pure LLMs don’t work on their own, and would thus be considerable vindication for my two strongest and longest held claims.
Leon Lang
@Lang__Leon
·
348d
@GaryMarcus In the case that these techniques work for general reasoning (with an appropriate replacement of the proof-checker in Lean): will you consider this a refutation or confirmation of your views on LLMs?
Gary Marcus
@GaryMarcus
·
348d
Neurosymbolic for another win!!!

The statement 'If you are asking that question you are missing the point. LLMs ≠ AlphaProof.' is part of a substantive engagement with public issues related to artificial general intelligence (AGI) and the capabilities of large language models (LLMs). The tone is somewhat dismissive, suggesting that the question posed by @Lang__Leon is not relevant or misses the core issue. The intent appears to be to clarify a distinction between LLMs and a neurosymbolic approach like AlphaProof.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not appear to cause harm directly, but the dismissive tone could be seen as unconstructive.
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    The statement does not engage in cyberbullying, harassment, or hate speech, and respects the privacy and dignity of others.
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement does not promote understanding, empathy, or compassion; it is more focused on correcting a perceived misunderstanding. [-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 does not engage in constructive criticism or dialogue; it dismisses the question without providing a detailed explanation. [-1]