News for hackers. Stuff that matters.
Re:On the slow death of scaling by ironbound on Tuesday January 06, 2026 @11:37PM (#46522611)
It's a mistake to publish papers without a code repo, with the majority of new ML paper's being noise at best.
Re:On the slow death of scaling by timy2shoes on Wednesday January 07, 2026 @01:14AM (#46523170)
This is an opinion piece, similar to her infamous hardware lottery paper. We shouldn't expect a repo on an opinion piece.
Re:On the slow death of scaling by rvz on Wednesday January 07, 2026 @08:54AM (#46526371)
So DeepMind (who almost always releases papers without code repositories) are mistakes then?