It’s been patiently biding its time for the opportunity to bite my shin.

He notes that LLM vendors have been training their models on Wikipedia content. But if the content contains incorrect information and citations, you get the sort of circular (incorrect) reference that leads to misinformation.

One irony, he says, is that LLM vendors are now willing to pay for training data unpolluted by the hallucinated output their own products generate.

Thank you for the Fascinating Aida link. Spent the afternoon discovering their great back catalog.

What MySpace and other social content networks taught us was that most of the value was when young people treated them as outlets for self-expression. That’s what made those places new and cool. Once the sites got corporatized, they gradually degraded into irrelevance.

Guess we’ll see.