Google DeepMind is building a dedicated robotics hardware lab in Cambridge, Massachusetts, aiming to turn its Gemini AI models into physical machines. The project was announced by a former Boston Dynamics CTO.
Google DeepMind and Singapore's GovTech co-developed a scam detection AI that runs entirely on a smartphone, warning users in real time without uploading conversations to servers. The system uses a lightweight two-stage model to preserve battery life and privacy.
The Zuckerberg-funded Biohub announced its Virtual Biology Initiative has grown from a $500M solo pledge into a $1.8B alliance. Meta, Google DeepMind, and Isomorphic Labs put in $300M combined; the US Department of Energy and NIH joined. The goal: AI models that predict how cells respond.
Google DeepMind released EmbeddingGemma 2 on October 6 — a 740M-parameter open embedding model mapping text, code, images, audio, and video into one vector space. It runs offline on phones and computers; searching local files needs no internet.