Gary Marcus

Rank 13 of 47
|
Score 103
6/29/2025, 2:21:50 PM
X
In reply to:
Gary Marcus
@GaryMarcus
·
20d
An astoundingly stupid reply to my recent Substack on chess, by someone who obviously didn’t actually read the essay:

If GPT-5 can’t play chess without making illegal moves 5% of the time – despite exposure to millions of games and explicit rules and books about chess etc – that
Chris
@chatgpt21
·
20d
Gary Marcus can’t go 5 minutes without saying o3 can’t play chess.

Eventually it will be “GPT 5 can play chess but it made an illegal move 5% of the time”

The statement and conversation engage in a debate about the capabilities of LLMs, particularly in playing chess, and their implications for AI alignment and development. The tone is critical and somewhat dismissive, with a focus on the limitations of LLMs.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement raises concerns about AI capabilities, which could be seen as promoting caution, but the dismissive tone may not fully align with striving to do no harm.
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    The use of terms like 'astoundingly stupid' could be seen as disrespectful, potentially violating the principle of respecting others' dignity. [-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 conversation includes personal attacks, such as calling a reply 'astoundingly stupid,' which does not align with engaging in constructive criticism. [-2]