OpenAI restricts GPT-5.6 under government pressure
OpenAI has limited the rollout of its GPT-5.6 model following a government request, sparking debate about regulatory influence over AI development. The company emphasized that such restrictions should not become the norm, while the New York Times criticized Microsoft for building what it calls a copyright-infringing supercomputer for OpenAI. Only a small group of trusted partners currently have access to GPT-5.6 in a limited preview.
The Trump administration has authorized Anthropic to deploy its Mythos cybersecurity AI model to over 100 US companies and government agencies. This marks a significant endorsement of AI-powered cybersecurity tools at the federal level. Anthropic had previously been required to restrict Mythos's availability pending government review.
Amazon Prime Day is underway with aggressive discounts across electronics and smart home categories. WIRED highlights the best deals worth spending on, including top picks under $100 and significant Apple product discounts. Digital Trends also rounds up last-minute smart home, laptop, and accessory deals that shoppers should not miss.
China has released its first connectivity standard for AI agents, introducing digital ID cards for AI systems to ensure interoperability and trust. Separately, Thailand's Saha Group is partnering with Japan's SoftBank to develop AI-powered business services. Japanese banks and tech firms are also collaborating on an AI chat tool for retail asset management, signaling a regional push toward standardized AI deployment.
AMD-powered PC deals are drawing attention during Prime Day, with steep discounts on Ryzen-based prebuilt systems and component bundles. Tom's Hardware highlights a Ryzen 7 9800X3D prebuilt with RX 9070 XT and 32GB DDR5 at $750 off, along with AM5 motherboard bundles that effectively include a free Samsung 990 Pro SSD. The 10th Anniversary Ryzen 7 5800X3D CPU has also come back in stock.