Two new Apache 2.0 projects achieve 97% better hardware efficiency for bursty bursty agentic workloads. Want to get involved?
Google unveils a new developer platform for building multi-agent workflows with Gemini-optimised harnesses.
With increasing NPU support, Google argues it has the platform that makes it trivial to roll your own agents for Android, iOS and even Raspberry Pi, via what used to be TensorFlow Lite.
DeepMind turned to a 2022 paper to speed up its open source Gemma 4 models.
"A few months ago running something this capable locally meant serious hardware and serious tradeoffs on quality..."
Anthropic turns off the compute taps for third-party tools and rushes to get more TPUs online.
Chrome has already seen half as many zero days this year as it did in 2025.
TurboQuant reduces the working memory needed for vector quantisation without sacrificing accuracy, offering a material reduction in inference costs.
Hackers "enumerated and accessed objects within S3 buckets, terminated production EC2 and RDS instances, and decrypted application keys."