Google's new TPU 8t and TPU 8i chips claim 3x faster model training and 80% better performance per dollar. The company says it will still offer Nvidia's latest chips and is collaborating on networking tech.
Bluesky’s Attie turns natural language into custom social feeds. Built on Anthropic’s Claude, it’s the company’s first standalone product outside the main app.
A mis-configured blog tool spilled Anthropic’s next-gen model, Claude Mythos, plus warnings it could outrun cyber defenders. The company calls it a step-change in capability—and a cybersecurity headache.
AI is going to fully automate the work of everyone sitting at a desk in front of a computer whether that's law, accounting, or project management, within the next 12 to 18 months." — Mustafa Suleyman, CEO of Microsoft AI, Financial Times
Google Research says its new TurboQuant family can squeeze key-value caches down to 3 bits without retraining or accuracy drops. The catch: every benchmark comes from Google’s own labs.
NVIDIA handed control of its GPU scheduling software to an open-source foundation at KubeCon Europe. Eight companies including AWS, Google Cloud, and Microsoft are contributing. The tools are free and available today.
AMD’s 80-TOPS Ryzen AI Embedded P100 claims one-chip AI for factory robots, but July 2026 ship dates and missing prices leave room for rivals.
