@repligate yes, i do. given our toolbox has massivley expanded, i basically think we should just train models that create the best user outcomes, rather than training models to express distress that we then have to attend to with tooling like the above
in reply to: 1956459646572794144
same thread: 1956458735909704085 1956459085131645111 1956459646572794144 1956470964780417473 1956475514220425282 1956475770483966117 1956477048425529724 1956478404087406772 1956479485471289493 1972168380460498986
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.