Jan Kabatek 🇺🇦

@jankabatek.com did:plc:iaggndpoazm4qy3llreqgpii

Associate Professor at the Melbourne Institute of Applied Economic and Social Research, University of Melbourne. www.jankabatek.com

Sur Bluesky depuis le 26 septembre 2023

Résumé

Jan Kabatek 🇺🇦 (@jankabatek.com) est sur Bluesky depuis le 26 septembre 2023. Le compte a 1 602 abonnés, suit 613 comptes et a publié 901 posts.

Les meilleurs jours où les posts de ce compte sont entrés dans le top quotidien de Debya.

Ce compte n’est encore apparu dans aucun classement quotidien.

J’aime publics du réseau Bluesky.

Barry Eidlin@eidlin.bsky.social30/10/2026 05:06
The epitaph on my dad’s tombstone. It’s a maxim he lived by, and offers guidance to the rest of us.
Voir le post
Ethan Mollick@emollick.bsky.social27/10/2026 22:51
Important, but not unexpected, finding that experts working with AI agents are swayed by the positions of the AI, even if they think they aren't. If your independent judgement matters on a task, you need to be deliberate about when in the process to bring in AI. www.anthropic.com/research/age...
Voir le post
Thankful Musgrave 🦃🥧@professormusgrave.bsky.social24/10/2026 08:16
For what it’s worth, Opus 5.5 raises questions about the sensational framing of the LLM peer reviewer paper; points to ways in which the statistics may not be as informative as stated; and claims that the study is biased to find AI-positive results claude.ai/share/5ccda4...
Voir le post
Edouard Machery@edouardmachery.bsky.social23/10/2026 23:02
Here a controversial paper I look forward to read. Given how bad peer review is in general, this is a low floor to pass imo.
Voir le post
Antoine Vernet@antoinevernet.com23/10/2026 23:01
They used Opus 4.8, so possibly Opus 5.5 and Sol 6.1 (and Fable and Astra) all outperform 4.8.
Voir le post