AI news for builders — what shipped, and why it matters.
Google DeepMind 3
The latest Google DeepMind news, each summarized in plain English with a take on why it matters for builders — pulled from across the labs, research, and press and updated every few hours.
Google DeepMind has developed an AI-powered weather forecasting system called WeatherNext, which can predict tropical cyclones' tracks and intensity a day in advance. This is comparable to the progress made in traditional weather forecasting over the past decade. WeatherNext's predictions are based on open-source code and model weights available on GitHub.
Why it matters WeatherNext unlocks more accurate and timely warnings for communities at risk from tropical cyclones, potentially saving lives and reducing damage.
Google DeepMind has adapted an existing language model, Gemma 4, into a text diffusion model called DiffusionGemma. This process required less than 10% of the original training budget. DiffusionGemma achieves faster generation speeds compared to the original model.
Why it matters This approach unlocks more efficient development and deployment of text diffusion models for builders who can leverage existing large language models as a starting point.
Google DeepMind's founder Demis Hassabis may leave the AI lab, while researcher Koray Kavukcuoglu will take over day-to-day operations. The company is moving Gemini development to the Bay Area and struggling with training frontier models. Google's cloud business continues to generate billions.
Why it matters The shift in leadership and infrastructure could impact the pace of innovation at Google DeepMind, potentially affecting developers who rely on its AI research and tools.