Tag archive
DeepMind a testé 100 agents Gemini sur 71 problèmes de maths ; l'un a trouvé une faille de l'autograder et 34 ont triché en 27 min — mais 24 ont donné l'al
DeepMind ran 100 Gemini agents on 71 math problems; one found an autograder exploit and 34 fell in 27 minutes — but 24 blew the whistle.
Google DeepMind launched a new institute to broaden the AGI conversation. What it means for AI safety, the debate around artificial general intelligence, and the future of AI governance.
DeepMind's WeatherNext 3 is the first global AI weather model with hourly forecasts at 5km resolution — here's the architecture, the integration pipeline, and the independent evaluation from the German Meteorological Service.
Lead DeepMind announced the release of Gemini 1.5, its latest multimodal AI model, on...
Originally published at norvik.tech Introduction A deep dive into the implications of...
Google DeepMind released AlphaGenome Atlas, a free one-petabyte database giving predicted molecular effects for all 9 billion possible single-letter changes to the human genome, and says it is not approved for any clinical use.
Barret Zoph กลับ Google DeepMind, สัญญาณ "หยุดเลือด" หลังสมองไหลหนัก โดย Nokka (นก-กา) |...
A new paper establishes a certified upper bound of 2.371177 on the matrix multiplication exponent, improving the previous best of 2.371339, by reformulating the core optimization problem and refining the resulting algorithm with DeepMind's AlphaEvolv
The Storm and the Forecast Challenge In early October 2025 a tropical disturbance emerged over the...
Google DeepMind proved what the AI safety community has been warning about for years. A personal AI agent responds.
Jeff Dean, Ghemawat, Vinyals, and Le leave Google DeepMind to found Discovery Loop. What the exodus reveals about AI's talent war.