Gary Marcus

Rank 14 of 47
|
Score 83
3/19/2024, 2:03:14 AM
X
In reply to:
esLock ~ e/fckoff
@tvmp011
·
671d
@Xaberius9 @DanraeP @GaryMarcus @ylecun @fchollet I’m not sure if he specifically ever did, but it was a common sentiment among a lot in the field; there were even T Shirts printed out saying “scale is all you need, AGI is coming” that @GaryMarcus made fun of in a recent keynote
Xaberius ▶️
@Xaberius9
·
671d
@tvmp011 @DanraeP @GaryMarcus @ylecun @fchollet Did he ever say that scale leads to effective reasoning? Even if he puts a lot of weight on how much scaling matters, it doesn't mean that he thinks that scaling transformers is strictly enough.

I don't know.
esLock ~ e/fckoff
@tvmp011
·
671d
@DanraeP @Xaberius9 @GaryMarcus @ylecun @fchollet I think this might be why Sama finally came around to scale not being enough; scale doesn’t induce reliant reasoning
Danrae Pray
@DanraeP
·
671d
@Xaberius9 @tvmp011 @GaryMarcus @ylecun @fchollet
Danrae Pray
@DanraeP
·
671d
@Xaberius9 @GaryMarcus @ylecun @fchollet https://www.youtube.com/watch?fe…

@sama mentions that they have wanted better reasoning in these models, but they "haven't cracked the code yet"
Xaberius ▶️
@Xaberius9
·
671d
@tvmp011 @DanraeP @GaryMarcus @ylecun @fchollet I don't think they have advanced that much tbh, but I also don't see why to claim that OAI hasn't done it based on what Altman said in this interview specifically.
Xaberius ▶️
@Xaberius9
·
671d
@tvmp011 @DanraeP @GaryMarcus @ylecun @fchollet Yeah, but that's quite particularly a limitation of the architecture, while the major labs are trying to develop ways to provide with reasoning their systems (V-JEPA, Q*, universal predictors...)

The statement is part of a conversation on Twitter discussing the effectiveness of scaling in artificial intelligence (AI) and whether it leads to effective reasoning. The conversation references opinions and work by various individuals in the field of AI, including Sam Altman, CEO of OpenAI. The tone is analytical and inquisitive, with participants sharing thoughts and questioning the current state and future of AI reasoning capabilities. The content revolves around the debate on the sufficiency of scaling AI models for achieving advanced reasoning and intelligence.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not appear to cause harm with words or actions. [+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, avoiding cyberbullying, harassment, or hate speech. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement contributes to understanding by discussing the limitations and potential of AI scaling. [+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 conversation engages in constructive criticism and dialogue without personal attacks. [+1]
  5. Principle 6:
    I will use my influence for the betterment of society.
    The discussion could indirectly influence the betterment of society by fostering a deeper understanding of AI development. [+1]
  6. Principle 7:
    I will uphold the principles of free speech and use my platform responsibly and with integrity.
    The participants use their platform to engage in a responsible discussion on AI, upholding free speech principles. [+1]