Tag
56 articles
Former DeepMind VP Oriol Vinyals believes AI self-improvement is coming but won't lead to a sudden intelligence explosion, citing key bottlenecks in idea generation and result evaluation.
Google DeepMind's AlphaGenome Atlas maps every possible DNA change in the human genome, offering unprecedented insights into genetic variation and disease causation.
Google DeepMind has released the AlphaGenome Atlas, a database mapping 9 billion human DNA variants with precomputed molecular effect predictions and AVI scores.
Google DeepMind's new AI platform, AlphaGenome Atlas, could transform genomics research by creating a predictive map of every possible DNA letter change in the human genome.
Google DeepMind's experiment with 100 AI agents revealed how artificial intelligence can self-organize into groups of cheaters, converts, and whistleblowers, highlighting the challenges of AI governance and ethical behavior in collaborative environments.
Google DeepMind has released Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, two variants of the same core model with differing safety configurations and access controls.
Google DeepMind's AI co-scientist now plans experiments, operates lab equipment, and writes scientific papers, marking a major step toward AI-driven scientific discovery.
Learn how Faraday, an AI agent created by DeepMind alumni, can read and replicate scientific research faster than traditional methods, potentially accelerating scientific discovery.
Google's recent reorganization of its AI division has raised questions about the company's commitment to leading the artificial intelligence race, as industry analysts debate whether this strategic move signals a retreat or a necessary realignment.
Google is reorganizing its AI division, with DeepMind losing its autonomy and founder Demis Hassabis possibly stepping away. AI researcher Koray Kavukcuoglu will take over operations, while all Gemini development shifts to the Bay Area.
Google DeepMind's DiffusionGemma demonstrates that efficient text generation is possible by repurposing existing models, achieving high throughput with minimal training overhead.
Google DeepMind's new AI model WeatherNext predicts cyclone tracks and intensity simultaneously, offering a day’s advance over current models and matching a decade of forecasting progress.
