This is the fundamental paradox of AI: if it's actually helping you, there is no way to know when it is no longer helping you. Put another way: if you can supervise it effectively enough to catch its mistakes, you probably didn't need it in the first place.

NPnatureportfolio.bsky.socialSep 30, 2024 1:39pm

Scaling up and shaping up large language models increased their tendency to provide sensible yet incorrect answers at difficulty levels humans cannot supervise, highlighting the need for a shift in AI design towards reliability, according to a paper in Nature. go.nature.com/4eCAnis 🧪

Say it’s wrong per Bluebook and Chicago Manual

I believe it’s a Solicitor General thing - used throughout the public version of their style guide

Didn’t he used to wear a shaded visor

That seems like medical incompetence possibly deserving of a report somewhere

How much confidence? Because that’s a consequential risk to manage

UVA Law had a class “the legal vineyard” for transactional skills around a vineyard client from formation to operation to bankruptcy - as I recall like hundreds of course registration bids for like 12 spots (I never got in)

“Wine cheese and chocolate appreciation” I heard that’s a real seminar at Wake Forest

Your AI doctor is never going to diagnose what caused this cold

NO OFFICIAL STATE CODE, ANNOTATED OR NOT, IS COPYRIGHTABLE; AND SMALL DIFFERENCES IN HOW LEGISLATURES CONTRACT WITH LEGAL PUBLISHERS ARE NOT AN ESCAPE HATCH FROM GEORGIA V. PUBLIC.RESOURCE.ORG (SCOTUS 2020)