Gary Marcus

Rank 15 of 47

Score 89

@kenneth0stanley I don’t think you can learn to understand the visual world from text alone. but congenitally blind people learn a ton about the world without visual input, with much richer representations than I would ascribe to LLMs.

Landau & Gleitman had a terrific book about this in the…

10/5/2024, 11:36:27 PM

In reply to:

Kenneth Stanley

@kenneth0stanley

360d

@GaryMarcus Do you think text alone can be sufficient? Or it needs to consume multimodal data?

Gary Marcus

@GaryMarcus

360d

@kenneth0stanley I am sure there is. Brilliant humans have trained on a tiny fraction of the internet.

Kenneth Stanley

@kenneth0stanley

360d

Though useful and powerful, an LLM today trained on virtually all the text on the internet is still far from human-level. Do you think there exists *any* architecture and algorithm that when trained on this same corpus would surpass humans? Or is the data simply inadequate?

The statement is part of a public discourse on the capabilities and limitations of language models (LLMs) in understanding the world, particularly in comparison to human cognition. The tone is analytical and seeks to contribute to an ongoing discussion about AI and human learning.

Principle 3:
I will use my words and actions to promote understanding, empathy, and compassion.
The statement promotes understanding by comparing the learning processes of congenitally blind people and LLMs, suggesting a nuanced view of how knowledge can be acquired without visual input. [+2]
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 responding to a question with a thoughtful perspective, without resorting to personal attacks or dismissive language. [+2]