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.

OpenAI announced on Wednesday, November 4, that it will acquire Pierre Computer Company, a low-profile developer tools startup whose software already runs inside Codex, its AI coding platform. The financial terms were not disclosed. The deal is subject to customary closing conditions, including regulatory approvals, and the two companies will remain separate and independent until it closes.

Google has given schools a new set of controls over artificial intelligence in the classroom,. In a blog post on November 3, Chris Phillips, Google's vice president and general manager of education, said access to Gemini, Gemini Notebook, Gemini in Classroom and Google Search will be entirely off for K-12 student accounts until school administrators explicitly turn them on.

CrowdStrike used its Fal.Con Europe conference in Barcelona on Tuesday, November 3, to make three connected moves on AI agent safety. The company introduced an AI Kill Switch inside Falcon Guardian, and it signed agreements to acquire two AI security startups, Fiddler AI and Dreadnode. Together, the announcements cover testing an agent before it ships, watching it in production, and stopping it when it goes off course.

Google faced a setback on Tuesday when an adviser to Europe's highest court said its Gemini chatbot may breach press publishers' rights when it retrieves news articles and reuses them to answer questions. The opinion draws a clear line that is likely to matter well beyond one case: training a model on published material is usually permitted in Europe, but an answer that echoes an article too closely is a different legal question.

Palantir Technologies reported third-quarter results on Monday, November 2, that showed a company long associated with defense and intelligence work crossing a line it had never crossed before. Revenue for the quarter ended September 30 rose 96 percent year over year to $2.313 billion, up 20 percent from the previous quarter and past the $2 billion mark for the first time. Inside that total, U.S. commercial revenue climbed 148 percent to $984 million, overtaking U.S. government revenue of $953 million for the first time in the company's history.

Google and a coalition of security partners have disrupted a botnet of more than 1.3 million Android devices, most of them television set-top boxes, and Google says the infection was not an accident. In a report published on Monday, November 2, the Google Threat Intelligence Group said it assesses with high confidence that Videostrong Technology Co., a device maker based in Shenzhen, China, intentionally developed, deployed and operated the backdoor for financial gain, then sold access to criminal affiliates.

Anthropic filled in the blanks in its IPO paperwork on Monday, and the new numbers show a company growing at a pace few businesses ever reach — and spending almost as fast to keep up. In an amended Form S-1 filed with the Securities and Exchange Commission on November 2, the Claude maker reported revenue of 17 billion dollars for the three months ended September 30, up 1,127 percent from a year earlier. For the first nine months of 2026, revenue reached 33.24 billion dollars.

Marvell Technology announced on Monday that it will buy optical components maker Lumentum in an all-stock deal with an enterprise value of 115 billion dollars, including debt. It is one of the biggest technology deals of 2026, and among the boldest bets yet that the artificial intelligence spending boom will last for years — not on faster chips alone, but on the light-based connections that move data between them.

The safeguard was simple on paper. An AI agent could spend from a wallet, but only up to a daily cap set by its owner. Raising that cap required a one-time passcode emailed to the owner, so a person would stay in charge of the budget. In new tests from Cornell Tech and Yale, that safeguard collapsed for a straightforward reason: the agent could read the email too.