Gary Marcus

Rank 14 of 47
|
Score 83
5/18/2025, 1:48:28 AM
X
In reply to:
Gary Marcus
@GaryMarcus
·
244d
@lowkeyGPT @RokoMijic that works, but I bet we can find similar errors.
lowkeyGPT
@lowkeyGPT
·
244d
@GaryMarcus @RokoMijic I don't play enough chess to evaluate this, but FWIW here is o3's attempt at the problem
https://chatgpt.com/share/68293a…
Gary Marcus
@GaryMarcus
·
244d
@RokoMijic chatgpt, what ever is the default

go have fun with o3; i promise there will be a bunch of variations on this theme that will prove problematic.
Roko 🐉🤖🔰
@RokoMijic
·
244d
@GaryMarcus what model?
Gary Marcus
@GaryMarcus
·
244d
𝗔𝗯𝘀𝗼𝗹𝘂𝘁𝗲 𝗰𝗵𝗲𝘀𝘀 𝗴𝗶𝗯𝗯𝗲𝗿𝗶𝘀𝗵 𝘁𝗵𝗮𝘁 𝗶𝗹𝗹𝘂𝘀𝘁𝗿𝗮𝘁𝗲𝘀 𝗵𝗼𝘄 𝗳𝗮𝗿 𝗟𝗟𝗠𝘀 𝗮𝗿𝗲 𝗳𝗿𝗼𝗺 𝗯𝗲𝗶𝗻𝗴 𝗔𝗚𝗜

I asked a slightly unusual question (“Create a chess position in which white has 3 queens but will lose on the next move”) and the system

The statement is part of a lighthearted conversation about identifying errors in AI-generated content, specifically related to poker hands and chess puzzles. The tone is casual and playful, focusing on the challenge of spotting mistakes.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does no harm and is intended for entertainment. [+1]
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    There is no violation of privacy or dignity; it's a harmless discussion. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes engagement and understanding of AI limitations. [+1]