Gary Marcus

Rank 26 of 47
|
Score 55
7/2/2024, 4:30:54 PM
X
In reply to:
Rhydhimma
@rhydhimma
·
114d
@GaryMarcus @EMostaque Is inference always at the pixel level? Given that there are high level representations in deeper layers of most networks?
Gary Marcus
@GaryMarcus
·
114d
@EMostaque Not sure whether you are being sarcastic but there is something very weird towards the end, almost like a pane of glass has appeared from nowhere. That’s an artifact of doing inference in pixel space, not what you would get if you had modeled the world from the scene’s beginning
Emad
@EMostaque
·
115d
Gen AI will never understand physics and how stuff works, this is all an illusion and will never improve 🙄
Smoke-away
@SmokeAwayyy
·
115d
Runway Gen-3 is feeling less like a rehash of the training data and more like a world simulator.

The man's movements. The attempt to to get the space sprinkler physics right. His eye/face movements at the end.

Very impressive model that creates its own impressions of the world.

The statement 'the result, and there are many similar, kind of speaks for itself. the features are visual, not conceptual.' is part of a technical discussion about the capabilities and limitations of generative AI models, specifically in the context of visual versus conceptual features. This constitutes public discourse as it engages with a substantive issue in the field of AI, contributing to the broader debate on the technology's potential and limitations.

  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. It is a technical observation rather than a personal attack or harmful comment.
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement is neutral and does not explicitly promote understanding, empathy, or compassion. It is more focused on providing a technical critique.
  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 criticism by addressing the technical aspects of the AI model without resorting to personal attacks or ad hominem arguments. [+1]
  4. 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 and uses the platform responsibly by contributing to a technical discussion with integrity. [+1]