It doesn't make sense to me – if they are sharing it with external orgs to get feedback, why would they release it the same day? Or just giving them half a day extra to access it? Feels odd
RESEARCH
-

Google Releases Gemini 2.5 Deep Think to AI Ultra Subscribers
By
–
Today, we’re making Gemini 2.5 Deep Think available to Google AI Ultra subscribers in the @GeminiApp Gemini 2.5 Deep Think is a variation of the model that achieved the gold-medal standard at this year’s International Mathematical Olympiad (IMO). Deep Think uses parallel
-

Same AI Model Solves International Mathematical Olympiad Problems
By
–
Impressive that this is the same model they used for the International Mathematical Olympiad a few weeks ago
-

Ultra-Scale Playbook: Training DeepSeek-V3 with Advanced Parallelism
By
–
The Ultra-Scale Playbook (the large-scale LLM training guide from the @huggingface science team) is out now! It is a 246-page, very nicely designed PDF that walks you through learning how to train your own DeepSeek-V3 model using:
• 5D parallelism, • ZeRO,
• fast kernels, -
Nanotron: World-class AI research paper released
By
–
kudos @lvwerra @Nouamanetazi @FerdinandMom @mohamedmekkouri Haojun Zhao Phuc Nguyen and @Thom_Wolf the world-class authors get the PDF before the summer break at http://
hf.co/nanotron -
Gemini Deep Think Proves Mathematical Conjecture with Novel Approach
By
–
See mathematician Michel van Garrel talking about how our latest Gemini Deep Think model was able to prove a conjecture using a very different approach than he was considering. https://t.co/CnMqOHRK3k
— Jeff Dean (@JeffDean) 1 août 2025See mathematician Michel van Garrel talking about how our latest Gemini Deep Think model was able to prove a conjecture using a very different approach than he was considering.
-

Gemini Deep Think Now Available for AI Ultra Subscribers
By
–
Starting today, we're offering Deep Think in the Gemini app for Google AI Ultra subscribers, and we're giving a collection of mathematicians access to the full version of the Gemini 2.5 Deep Think model that achieved gold medal level performance in the recent IMO competition.
-

Phi-Ground Tech Report: Advancing GUI Perception
By
–
Phi-Ground Tech Report Advancing Perception in GUI Grounding
-

Seed-Prover: Deep and Broad Reasoning for Automated Theorem Proving
By
–
Seed-Prover Deep and Broad Reasoning for Automated Theorem Proving
-

Week of AI breakthroughs: superintelligence, batteries, CRISPR, IPO
By
–
Zuck's Personal Superintelligence.
First iron-air battery connected to grid.
Profluent Bio releases OpenCRISPR-1 model.
Figma crushes IPO, currently trading at 4 Linas.
Abundance x Jason Carman drop Zipline video. + Means & Meaning What a week for the optimists.