deepmind

Google Quantum AI demonstrated magic state cultivation on its Willow processor for the first time, cutting errors 40-fold and hitting 99.99% logical fidelity — a direct attack on the resource bottleneck that could consume 95% of a quantum machine.

Google DeepMind's new flagship reasoning model, Gemini 4 Argon, is globally available. The benchmark scores grab headlines, but the figure that changes what you can build is the output limit: 1M tokens, up from 64K. Here's how to read the launch.

Google DeepMind CEO Demis Hassabis wants frontier AI models to pass an independent safety exam — a FINRA-style standards body — before they reach US users. Here is the bet behind the ask, and the voluntary-to-mandatory catch inside it.