Gary Marcus

Rank 14 of 47

Score 83

@tdietterich agreed, and nice point. but also… it’s in assuming that you need large scale GPUs that he goes wrong. they aren’t actually all that great at reasoning. and certainly aren’t remotely as efficient as human brains.

i do agree that (depending on the nature of the task)

11/15/2025, 5:10:04 PM

In reply to:

Thomas G. Dietterich

@tdietterich

44d

@GaryMarcus I think he has a point independent of whether LLM technology is involved. Intelligent systems must respond to unfolding situations with agility. That means the responses can’t be precomputed — real time reasoning is needed.

Gary Marcus

@GaryMarcus

44d

The logic here is exactly backwards.

Jensen isn’t giving an account of why there is no bubble - he’s explaining why the whole GenAI approach is doomed.

See my reply for why.

Rohan Paul

@rohanpaul_ai

44d

Jensen Huang explains the difference between AI and traditional Software and why there's no bubble.

Traditional software was pre-compiled, meaning it was built once, stored, and then executed repeatedly with little computation. It didn’t need constant high-power processing once

The statement engages in a technical discussion about the efficiency and reasoning capabilities of GPUs compared to human brains, in the context of AI development. It critiques the assumption that large-scale GPUs are necessary for reasoning, suggesting they are not as efficient as human brains. The tone is analytical and aims to contribute to the debate on AI technology.

Principle 1:
I will strive to do no harm with my words and actions.
The statement does no harm and contributes constructively to the discussion. [+1]
Principle 3:
I will use my words and actions to promote understanding, empathy, and compassion.
It promotes understanding by providing a perspective on the limitations of GPUs in AI reasoning. [+1]
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.
Engages in constructive criticism without personal attacks, focusing on the technical aspects. [+1]