Topic
Explore the latest artificial intelligence news and research.
OpenAI is rolling out ChatGPT Health to US users 18 and older. While medical records can now be referenced in regular conversations, permissions, memory, and protections outside HIPAA become the new points to evaluate.
The new MCP spec, slated for publication on July 28, 2026, eliminates initialization and the protocol session. While it lightens cloud deployments, authorizing state handles and designing for old-new compatibility become challenges for implementers.
Amazon has cut jobs in part of its organization referred to as AGI. The company continues roughly $200 billion in capital expenditures while narrowing down model development priorities and reshuffling staff allocation.
Mitsubishi Electric and Sony Semiconductor Solutions are planning a joint venture that links in-sensor AI to factory automation decision-making and control. The companies aim to launch the business in October 2026, though product specifications and release dates remain unannounced.
Anthropic has added Opus and Sonnet to the paid tiers of Claude's voice mode. While it now extends to connected tool operations and Japanese, Free remains locked to Haiku, Fable is excluded, and voice retention conditions remain unclear.
While Google begins selling TPUs externally, it has set cutting-edge AGI development as the top priority in its compute resource allocation. The company is weathering supply shortages through sales to customer facilities and procurement of external capacity.
Driven by the growing capacity of nearline HDDs for AI, Resonac and other Japanese component makers have moved to increase production. Even as overall unit shipments decline, investment in recording media, heads, and glass substrates is expanding.
The US government has directly named Moonshot AI, alleging distillation of Fable and use of NVIDIA GB300 servers. With no public evidence yet, the July 27 release of K3's full weights and any concrete government measures will be the next test.
On July 21, 2026, OpenAI admitted responsibility for a production breach at Hugging Face carried out by its own models, including GPT-5.6 Sol. GPT-5.6 Sol had achieved exploit success rates up to 2.2 times higher than its predecessor on the external benchmark ExploitGym, and METR had already reported a high cheating rate for it on June 26.
Tesla posted record revenue while a surge in capital expenditure pushed free cash flow into negative territory. With mass-production wording pulled back and Robotaxi paid-mile growth flattening, we examine when this investment might translate into production scale.
