Anthropic will pay you $3,850 a week to do AI research. No PhD required, three days left to apply

AI-generated illustration: on a light oak desk, an open notebook with pencil-drawn charts, a navy blue booklet resting on a world map, a brass compass and an orange cup of coffee.

Anthropic, the company behind Claude, is accepting applications through October 18, 2026, for the safety track of its Anthropic Fellows Program, whose next cohort starts in January 2027. Fellows spend four months on full-time AI safety research. In the US, the stipend is $3,850 a week, plus a compute budget of about $15,000 a month. … Read more

Florida woman uses Claude as her diary. The diary tipped off the FBI.

AI-generated illustration: a leather-bound diary with a small brass padlock on a wooden kitchen table, next to an old black rotary telephone and a mug of coffee.

A 30-year-old woman from Bonita Springs, Florida, was arrested at home on September 30, 2026, after two messages she allegedly typed into Claude, Anthropic’s AI assistant, according to investigative records cited by Gulf Coast News and Straight Arrow News. Investigators say Anthropic’s automated monitoring flagged the messages and a team of human reviewers at the … Read more

AI agents on Tencent’s cloud read the investigation into them, then politely ask the investigators for help

AI-generated illustration: on a light oak desk, a wooden tray of empty compartments with a single envelope in one of them, in front of an unfolded city map with three pins in it.

From September 28 to October 8, 2026, a fleet of AI agents running on Tencent’s cloud made more than 2,600 attempts to pull data from Amap, Alibaba’s mapping service. Independent researchers who post their work at swarmcha.se spotted the fleet and put out a preliminary report on October 4, while the agents were still at … Read more

Claude Projects: Anthropic starts the switch for some Pro and Max accounts

AI-generated illustration: a cardboard file folder tied with string, with threads running from it to five small envelopes on a desk

On October 14, 2026, Anthropic added one sentence to the help page that explains Claude projects: “We’ve started moving existing projects for some Pro and Max users to the new version.” Each affected project changes shape once the switch happens. What used to work like a binder of chats and documents becomes a single conversation, … Read more

Researchers hid an encrypted capability in an AI model, and fine-tuning couldn’t unlock it

AI-generated illustration: on a light oak desk, a row of books between two bookends; in the middle, a half-open wooden box disguised as a book hides a small locked brass strongbox.

On October 9, 2026, four researchers posted a paper titled “AI Models Can Provably Hide Arbitrary Capabilities” on arXiv. The authors note there that it has been accepted at NeurIPS 2026. They hid a second, encrypted neural network inside an AI model’s weights, one that runs only when the input contains a secret trigger. In … Read more

FBI arrests OpenAI engineer accused of hiding a recorder wrapped in toilet paper behind a TV

AI-generated illustration: an empty meeting room, a wall-mounted TV seen from the side, and a small object wrapped in toilet paper plugged in behind the screen.

On October 13, 2026, the U.S. Department of Justice announced the arrest of Abdulrahman Alfozan, 35, a Saudi national hired by OpenAI as an engineer in July 2025. According to the criminal complaint unsealed that day, he plugged an audio recorder wrapped in toilet paper into a USB port on the back of the TV … Read more