//
sign in
Post
by @danabra.mov
PostEmbed
by @danabra.mov
Record
by @jimpick.com
Record
by @atsui.org
+ new component
Post
There's a genre of LLM criticism that I find so destructively unhelpful: using a single vague prompt to "analyze" data, e.g. "summarize the themes across this collection of texts," to claim "see! LLMs don't work." & it's getting published! Reviewers this is not good work pls reject it if you see it.