Gary Marcus

Rank 19 of 47
|
Score 64
1/6/2025, 11:43:23 PM
X
In reply to:
Alessandro Della Corte
@Ognifedefingo
·
185d
It's a bit like asking a dog to bark how much is 3 x 4 and stop him right after the 12th bark yelling: "nailed at the first try!".
Alessandro Della Corte
@Ognifedefingo
·
185d
LLMs stans often emphasize when it "nails it at the first try". If it's not "at the first try" for me the value is almost zero.
You can just as easily guide it to a correct solution when the first try is wrong, as you can mislead it towards nonsense when the first try is ok.

The statement is part of a conversation critiquing the reliability and interpretation of language models' outputs, using the analogy of Clever Hans, a horse that was thought to perform arithmetic but was actually responding to cues. The tone is critical and somewhat humorous, aiming to highlight the potential for misinterpretation of AI capabilities.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not appear to cause harm; it uses humor to make a point about AI interpretation.
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement uses humor to promote understanding of AI limitations, aligning with the principle of promoting understanding. [+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 statement engages in critique without personal attacks, focusing on the topic rather than individuals. [+1]