2025-12-01 04:08 UTC
@ElleGray@mstdn.social You are too kind.
Replies (1)
-
@timbray@cosocial.ca 2025-12-01 04:14
@robpike@hachyderm.io @ElleGray@mstdn.social But that's fair almost by definition, right? LLMs produce output that is statistically similar to the corpus of training material, presumably not far off the median level. Thus mediocre.