Gary Marcus

Rank 19 of 47
|
Score 64
6/15/2025, 12:25:55 PM
X
In reply to:
Harshal G. Hayatnagarkar
@harshalhayat
·
26d
@skdh @GaryMarcus Claiming to solve big problems and yet making fundamental mistakes is the characteristics of not only the LLMs but also their hype-loving proponents.
Sabine Hossenfelder
@skdh
·
26d
@GaryMarcus Totally agree with that, I constantly have the problem with LLMs eg that they will chop together pieces from different sources, so suddenly in the middle of the supposed proof the definition of some key variable changes and the rest doesn't make any sense. As Tao says, it's not a
Gary Marcus
@GaryMarcus
·
26d
Terence Tao reporting on the same problem we have seen endlessly in other domains: LLM’s produce output that “looks” correct but is often deeply wrong and even stupid on careful inspection.

I know of no domain in which this is NOT the case.

And yet people seem surprised over
vitrupo
@vitrupo
·
26d
Terence Tao says today's AIs pass the eye test -- but fail miserably on the smell test.

They generate proofs that look flawless. But the mistakes are subtle, and strangely inhuman.

“There's a metaphorical mathematical smell.. it's not clear how to get AI to duplicate that.”

The statement is a brief, non-substantive reaction to a discussion about the limitations of large language models (LLMs). It uses an emoji to express agreement or emphasis but does not contribute to the public discourse on the topic.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not cause harm but also does not contribute constructively to the discussion.
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The use of an emoji does not promote understanding or empathy, as it lacks substantive content.