Gary Marcus

Rank 15 of 47
|
Score 76
6/16/2025, 10:21:47 PM
X
In reply to:
David Leiser دافيد דויד
@Professologue
·
30d
@GaryMarcus @rao2z
David Leiser دافيد דויד
@Professologue
·
30d
Structural understanding is not optional for General AI. It is the foundation. Without coherence undergirding performance, systems will keep producing fluent outputs — and unacceptable errors. *No structure without genesis. No genesis without structure*(Piaget)
David Leiser دافيد דויד
@Professologue
·
30d
Scaling alone won’t yield structured understanding. What’s needed isn’t more capacity, but a capacity of a different kind: the ability to reorganise one’s own representations, not just extend them.
David Leiser دافيد דויד
@Professologue
·
30d
As models grow more complex, interpretability has become a major concern. We now have attention maps, saliency methods, attribution graphs. These tools show how outputs were shaped — but not whether they reflect any coherent internal structure.
David Leiser دافيد דויד
@Professologue
·
30d
This isn’t an isolated case. As long as models remain indifferent to contradiction and unable to abstract from their own operations, there can be no internal pressure to resolve inconsistency — and no path toward structured understanding.
David Leiser دافيد דויד
@Professologue
·
30d
Verbal fluency often masks incoherence — contradictions can hide behind persuasive prose. But here, the incoherence is spatial. There’s nowhere to hide. The map reveals what the words often conceal: a lack of structured understanding.

The statement is part of a broader discussion on the development and understanding of General AI, specifically focusing on the need for structural understanding in AI models. It engages with public discourse by addressing a significant issue in AI development, which has implications for technology, ethics, and society.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does no harm and is part of a constructive dialogue on AI development. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes understanding by aligning with arguments that support the need for structural understanding in AI. [+1]
  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 engages in constructive dialogue by agreeing with and supporting previous arguments without personal attacks. [+1]
  4. Principle 6:
    I will use my influence for the betterment of society.
    The statement uses influence to contribute to the betterment of society by engaging in a discussion on improving AI systems. [+1]
  5. Principle 7:
    I will uphold the principles of free speech and use my platform responsibly and with integrity.
    The statement upholds free speech and uses the platform responsibly by contributing to a meaningful discussion on AI. [+1]