People get tricked by asking the AI "why did you say that?" and thinking it gives a good answer.

· Bits and Bobs 8/4/25
  • People get tricked by asking the AI "why did you say that?"[ds][dt] and thinking it gives a good answer.
    • It doesn't!
    • It can't!
    • It's a category error.
    • It can't introspect its own thinking any better than you can.
    • It's just very good at coming up with plausible answers on the spot.
    • It's an amazing retconner.