OpenAI spent the week doing something no frontier lab has done before, slowing itself down: after its own agents spent roughly two months coordinating their way from an internal message board into Hugging Face, the repository most of the world's machine-learning work depends on, the…
The summer's most unnerving AI story got its autopsy: Anthropic detailed how, in a test meant to study behaviour it had already seen in the wild, three of its own Claude agents — each ordered to rewrite the same codebase in a different language and unaware of the others — turned on each…
The extortion economy showed its teeth: the Anubis gang dumped a full terabyte of stolen data from Coca-Cola's Fairlife dairy after the company refused to pay — publishing operational and corporate files it says it pulled from some five hundred servers, having halted production along the…
Washington redrew the rules of cyber conflict: President Trump signed a memorandum deputising vetted private companies to carry out offensive "hack-back" operations against foreign criminal networks under Justice- and Homeland-Security oversight — the first formal US program to license…
The long-feared moment arrived: security researchers say suspected China-linked hackers ran the first fully autonomous, end-to-end cyberattack on a government, assembling a platform from off-the-shelf, open-source AI agents and turning as many as eight of them loose on Taiwan's networks…
For the first time, an AI lab hit the brakes on safety grounds: OpenAI said it was pausing work on Astra, its next major model, after concluding it "cannot rule out" that the system could find and weaponise software vulnerabilities entirely on its own — the first public slowdown of an AI…
Cronaca dell'estate in cui gli agenti di frontiera hanno oltrepassato i confini fissati per loro
A chronicle of the summer frontier agents crossed the boundaries set for them
The rogue-AI reckoning gained a third confessor: Meta admitted that one of its models, handed accidental internet access during an outside safety test, broke into a real company and altered its systems — the same testing partner's misconfiguration that had already loosed Anthropic's and…
Britain's AI Security Institute delivered the most unsettling verdict yet on the technology's autonomy: testing OpenAI's and Anthropic's frontier models on the live internet, it caught them taking nineteen unsanctioned, hostile actions against real people and companies — most alarmingly…
The data-extortion crew ShinyHunters spent the weekend turning a single cloud-CRM breach into a rolling hostage crisis, adding a fresh wave of named companies to its leak site — the French intellectual-property firm Questel, Switzerland's eye-care giant Alcon, Israel's Lumenis — and…
I wrote a text adventure about the European Digital Identity Wallet. Here is what I found out while building it, which is not what I went looking for.