OpenAI briefs government on GPT-6

Google Blog

Google launched three new Gemini Flash models—3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber—optimized for cost-efficient, high-performance AI agents with up to 65% token reduction and specialized cybersecurity capabilities. The releases signal Google's strategic focus on making production agentic workflows more economically viable while beginning pre-training on Gemini 4.

Bloomberg

OpenAI is preparing to launch GPT-6 and securing government buy-in through briefings with Trump administration officials and Congress on its capabilities and safety measures. This signals both regulatory engagement on frontier AI and the approach of the next generation model release.

X

Claude Cowork now enables users to teach the AI new skills by recording screen demonstrations with voice narration, which Claude then converts into repeatable automated tasks. This feature is available on paid subscription tiers through the desktop app.

TechCrunch

The U.S. is threatening sanctions against Chinese AI companies over alleged intellectual property theft in open source models, marking a potential escalation beyond chip export controls as Chinese models like Kimi K3 gain competitive ground. Industry experts dispute that model distillation constitutes theft and argue China's progress is driven by strong research capabilities rather than copying.

BBC

Google is building a custom AI chip (Frozen v2) arriving in 2028 that could be 6-10x more efficient for Gemini models, signaling tech giants' shift toward proprietary silicon to control AI infrastructure costs.