Tootfinder

OpenAI debuts o3-pro for ChatGPT Pro and Team users and in the API, costing $20/1M input and $80/1M output tokens; Enterprise and Edu will get access next week (Kyle Wiggers/TechCrunch)
https://techcrunch.com/2025/06/10/open

OpenAI releases o3-pro, a souped-up version of its o3 AI reasoning model | TechCrunch
OpenAI has launched o3-pro, an AI model that the company claims is its most capable yet.

„Machen wir aus unseren Dolomiten eine einzige große Rennstrecke. Gegen Bezahlung, versteht sich. Lassen wir die Autos auf unsere Pässe los, die fetten Motorräder, die Oldtimer, in denen so viel Nostalgie und so viel Kohle steckt.“
https://www.ff-bz.com/weitere-themen/leser

€
https://www.noz.de/lokales/osnabrueck/artikel/interaktive-karte-macht-ns-vergangenheit-von-osnabrueck-greifbar-48703538#comments
In diesem Zusammenhang sei auch mal die
Geschic…

Best-of-N through the Smoothing Lens: KL Divergence and Regret Analysis
Gholamali Aminian, Idan Shenfeld, Amir R. Asadi, Ahmad Beirami, Youssef Mroueh
https://arxiv.org/abs/2507.05913

Best-of-N through the Smoothing Lens: KL Divergence and Regret Analysis
A simple yet effective method for inference-time alignment of generative models is Best-of-$N$ (BoN), where $N$ outcomes are sampled from a reference policy, evaluated using a proxy reward model, and the highest-scoring one is selected. While prior work argues that BoN is almost optimal in reward vs KL tradeoffs, the effectiveness of BoN depends critically on the quality of the proxy reward model used for selection. For this purpose, we study BoN through a smooth version known as Soft Best-of-N…

Deltog i OSSEs parlamentariska församling i Porto på inbjudan av ukrainska delegationen kring hybridkrigföring. Att lyssna till kollegor från frontlinjen i det kognitiva kriget gav skärpa. Systematisk rysk påverkan testas i Ukraina, används mot oss.
https://carlheath.se/reflektioner-…

Characterizations for arbitrary B\'ekoll\'e-Bonami weights
Carlos Mudarra, Karl-Mikael Perfekt
https://arxiv.org/abs/2506.05993 https://

Characterizations for arbitrary Békollé-Bonami weights
We precisely characterize the relationships between the reverse Hölder inequality, the Fujii-Wilson condition, the Békollé-Bonami $\mathrm{B}_p$ condition, the $\mathrm{B}_\infty$ condition, and the reverse Jensen inequality, for arbitrary weights in the unit disc. This is achieved by introducing new side conditions that turn out to be necessary and sufficient. The side conditions are simple and testable, and can be interpreted as integral versions of the much stronger condition of bounded h…

Virginia Candidates Show You Can Try to Kill DEI, But You'll Fail (Jill Lawrence/The Bulwark)
https://www.thebulwark.com/p/virginia-candidates-dei-diversity-maga-racism
http://www.memeorandum.com/250623/p20#a250623p20

🇺🇦 #NowPlaying on KEXP's #AfternoonShow
George Michael:
🎵 Too Funky
#GeorgeMichael
https://georgekellygr.bandcamp.com/track/george-michael-too-funky-george-kelly-club-version
https://open.spotify.com/track/1flKNp2shyTeSAQE3Ezf8L

Siivoja oli tänään laittanut pyyhkeet tällä tavalla. En ehkä
Ihan ymmärrä. Pyyhkeet kuuluu kylppäriin, mistä on peräisin hotellien tapa laittaa ne sängyn päälle ”romanttisesti” hassun muotoon taivutettuina? Varsinkin jos ne tietää että huoneessa on soolomatkustaja.

Some researchers say it looks like DeepSeek may have used Google Gemini to help train its R1-0528 reasoning model (Kyle Wiggers/TechCrunch)
https://techcrunch.com/2025/06/03/deepseek-may-have-used-googles-gemini-to-train-its-lates…

DeepSeek may have used Google's Gemini to train its latest model | TechCrunch
Chinese AI lab DeepSeek released an updated version of its R1 reasoning model that performs well on a number of math and coding benchmarks. Some AI researchers speculate that at least a portion came from Google's Gemini family of AI.