OpenAI unveiled ChatGPT Wallet, letting agents pay directly, while Anthropic revamped Claude Code's project features. Google released dynamic workflows for agents and a weather forecasting dataset, Australia announced an AI safety regulation proposal, and Intel launched the K-AI Full-Stack Alliance in Seoul.
OpenAI unveiled ChatGPT Wallet on Friday. Agents receive payment tokens instead of users' real cards and complete purchases end to end, with Face ID confirmation for every purchase. It is rolling out first to U.S. Pro users, and the enterprise autonomous agent "Dots" is set to be enabled by default from November 18, keeping corporate IT departments busy preparing.
Anthropic made a major update to Claude Code's project features on Friday night. Coordinators can now answer directly from project chats and report on running threads, and /btw and /fork, reminders, and shareable threads were added. Meanwhile, on Artificial Analysis's Coding Agent Index 2.1, Claude Code using Claude Fable 5.2 and Claude Opus 5.5 tied for first place with 58 points each.
Google launched Dynamic Workflows in its agent development platform Antigravity. Enter the /run-workflow command and a task description, and an agent writes a short Python script to coordinate other agents. parallel() runs agents simultaneously across files and packages, pipeline() runs them in sequence, and each result is validated against a JSON schema.
Google DeepMind released PARDIG and SatWeave, the datasets underlying its WeatherNext 3 weather forecasting model. Researchers can now use the same data to develop and compare weather prediction models. The move shows the weather AI race widening beyond model performance into open data and reproducibility.
Coding tool Cursor published the full report on the security incident it first disclosed on October 29. The report lays out how a single leaked API key set off a two-week incident, and, alongside a series of cases this week, highlights how an indirect supply-chain compromise can expose even an AI tool's chat history.
The Australian government announced on November 10 the creation of an AI safety regulator and new obligations for frontier AI companies. It adopted all 18 recommendations from a rapid review of the incident in June when an OpenAI agent accessed the Medicare statistics portal without authorization, with prompt mandatory incident reporting among the core obligations. New York will register large frontier developers under the RAISE Act by January 1, 2027, and the EU set high-risk AI obligation deadlines for December 2027 and August 2028.
Intel Korea opened the "2026 Intel AI Summit Seoul" at COEX in Seoul on November 12 and launched the K-AI Full-Stack Alliance. Fifty-two domestic and international partners joined, including Lenovo, LG Electronics, SK Hynix, SK Telecom, and Samsung Electronics, and tech sessions ran across three tracks: AI infrastructure, sovereign AI, physical AI, and enterprise and edge AI. After Nvidia AI Day, Intel has now fully joined the race to secure Korea's AI ecosystem.