News for hackers. Stuff that matters.

A comment on The New York Times Launches a Strong Case Against Microsoft and OpenAI. Parent comment

Re:The New York Times Launches a Strong Case Against Microso... by zozbot234 on Sunday January 07, 2024 @11:48AM (#38902708)
> create knowledge that wasn't in the original data.

The word is "hallucinate" or "confabulate". The way these models "create" pretend-knowledge is totally useless.

[ Parent | on HN ]
Re:The New York Times Launches a Strong Case Against Microso... by ryanklee on Sunday January 07, 2024 @11:53AM (#38902777)
I'm not referring to hallucinations.

I'm referring to novel relationships drawn between datum in the corpus that are a result of training and inference.

This is apparent in something as simple as a summarization.

[ Parent | on HN ]