A convenience feature in Claude Code, Anthropic's AI coding agent, has left real passwords and API keys sitting in public view on GitHub. Research published on November 5 by the security firm Apiiro found that when a developer approves a command with the option "Yes, and don't ask again," Claude Code saves that command word for word as a permission rule in a file called .claude/settings.local.json, inside the project folder. If the command contained a secret, the secret is saved with it, in plain text.

Dave Wehner, Meta's chief strategy officer and one of the longest-serving finance leaders in big technology, is leaving the company after 14 years to become chief financial officer of Fireworks, the artificial intelligence inference startup. Wehner announced the move on X on Friday, and CNBC reported the appointment on Saturday. The move is one of the most high-profile executive departures from Meta's leadership team this year.

Google is taking direct control of a Wyoming data center project it once declined to use as a customer. Crusoe, the Denver-based developer that started the campus in southeastern Wyoming, confirmed on Friday that it has sold the unfinished project to Google, a day after Google announced a planned 17 billion dollar investment to build its first data center in the state.

OpenAI has closed the first 11 billion dollars of its latest funding round, at a valuation of 1.25 trillion dollars, Bloomberg News reported on Friday, citing people familiar with the matter. Reuters carried the report on Friday evening, November 6. The money has been wired, according to the reporting, but the round itself is not finished, and OpenAI declined to comment.

A research system built by Physical Superintelligence, a laboratory known as PSI, has produced a proof of the Aaronson-Ambainis conjecture, a central open problem in quantum computing that had stood since 2008. PSI announced the result on Friday, saying that its internally developed system, running instances of a non-public Anthropic model on PSI's own infrastructure, had a complete proof during the third day of the run. The proof was then accepted by Lean, a proof assistant that checks every logical step by computer.

The artificial intelligence trade suffered its sharpest reversal of the year on Thursday, November 5, as investors reassessed what this week’s midterm elections mean for the data center buildout that has powered the boom. The Nasdaq Composite fell 5.52 percent to close at 24,315, its worst session since April 2025, and finished firmly in correction territory, down as much as 14 percent from its all-time high reached only a month ago. Reuters estimated that some 2 trillion dollars was wiped off the Nasdaq’s market value in a single day.

Anthropic is now improving its AI models at about twice the pace it managed before late 2025, and most of that acceleration appears to come from the company using AI to help build AI. That is the central finding of an independent assessment published on November 5 by METR, a nonprofit that evaluates frontier AI systems, after its staff spent eleven weeks examining Anthropic's internal research process.

Meta has acquired Conductor, a young artificial intelligence startup whose software lets developers run several coding agents at once, each working on a separate part of a project in its own workspace. Meta confirmed the acquisition to TechCrunch but declined to comment further, and the financial terms were not disclosed. The deal was first reported by Axios, and Conductor announced it on its own blog on Wednesday.

The artificial intelligence buildout met its clearest political test yet on Tuesday, and voters were not persuaded. In 41 local ballot measures across 10 states on November 3, voters backed restrictions on data centers or rejected proposals that would have helped them get built in 40 cases, according to an analysis by Ballotpedia and results compiled by CNBC. The results, still unofficial, covered bans, moratoriums, zoning changes and requirements for public votes on large projects.

OpenAI has published the results of a striking security experiment: an autonomous pipeline powered by an upcoming cyber model scanned the Linux kernel, proved hundreds of vulnerabilities were real, built working exploits for them, and drafted a patch for each one. In a single mid-October run, the system produced 948 distinct local privilege escalation vulnerabilities, the kind of flaw that lets an ordinary logged-in user gain root access, the highest level of control on a Linux machine.