Hacker News new | ask | show | jobs
by 12_throw_away 43 days ago
> did we inadvertently train AIs on idiotspeak.

Nope! That is - training on lowest-common-denominator, low-signal high-noise "idiotspeak" was not at all inadvertent.