Do those findings about commenters apply to Bluesky also? :)

I guess I find it hard to get behind an article that tells me that we're all going to hell and there's nothing we can do about. There's loads of things that we can do about it and a heck of a lot of people are doing things right now.

Provocative? Negative more like. I don't pretend that there aren't serious problems, but people and institutions are working hard to address the problems and that work isn't mentioned once in there.

Speech recognition and machine translation research hugely influenced LLMs. It is worth remembering the pioneering research in information retrieval that for decades showed the value of taking a statistical approach to language problems.

Different formulations of tf*idf were tried, before the community settled on Robertson's incredibly robust BM25, presented at TREC in 1994 www.microsoft.com/en-us/resear...