Gary Marcus

Rank 10 of 47
|
Score 106
3/29/2024, 10:30:26 PM
X
In reply to:
moon island
@islandinthemoon
·
486d
@GaryMarcus @s_batzoglou OP was contrasting modest accomplishments of pure sym ai with successes of DL

You may disagree, but many people are impressed with what LLMs can already do and see promise in continued exploration. That's different from demanding blind fealty or closing other research doors.
Gary Marcus
@GaryMarcus
·
486d
@islandinthemoon @s_batzoglou 1. i am not arguing for pure symbolic approaches – nor have I ever, even once.
2. blind fealty to LLMs in conjunction with enormous $ and repeated attempts to stifle legitimate criticism has been greatly harmful to the development of other approaches; unprecedented in my…
moon island
@islandinthemoon
·
486d
@GaryMarcus @s_batzoglou I should also say I agree with you that neuro-symbolic approaches are a critical path forward, but I don't think one makes a plausible case for this by crapping all over what everyone can see are major successes, especially compared to pure sym AI past results
moon island
@islandinthemoon
·
486d
@GaryMarcus @s_batzoglou protein folding is based on an attention-based neural network system, not symbolic AI: https://www.youtube.com/watch?si…

It's already clear that AI (including LLM AI) will be a massive productivity boost to software engineers

Creating public-domain image datasets is a trivial problem
Gary Marcus
@GaryMarcus
·
486d
@islandinthemoon @s_batzoglou protein folding isn’t genAI and is neurosymbolic.

code quality and security has arguably gone down.

image synthesis is fine if you don’t care about copyright violation.
moon island
@islandinthemoon
·
486d
@GaryMarcus @s_batzoglou success like:
code-writing assistance
text-writing assistance
image-creation assistance
sound/music assistance
to say nothing of protein folding, etc

Sym AI has nothing remotely close in any of those domains

The statement is part of a conversation on Twitter discussing the merits and accomplishments of different AI approaches, specifically contrasting symbolic AI with deep learning and neurosymbolic AI. The tone is informative and corrective, as the user is clarifying their own contribution to the conversation regarding AI. The intent is to contribute to an ongoing debate about AI methodologies, which is a topic of public interest and relevance to the tech community and society at large.

  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 clarifying the user's position on AI approaches. [+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 providing a clarification without personal attacks. [+1]
  5. Principle 6:
    I will use my influence for the betterment of society.
    The statement uses influence to contribute to a technical discussion, which could indirectly benefit society by promoting better understanding of AI. [+1]
  6. Principle 7:
    I will uphold the principles of free speech and use my platform responsibly and with integrity.
    The statement upholds the principles of free speech by participating in a public debate responsibly. [+1]