An empty Waymo vehicle drove into a flooded Texas road and got swept into a creek. Now the company is patching every car in its fleet — and revealing how big that fleet has become.
A Hangzhou court ordered a company to pay £28,000 for firing a worker and replacing him with AI — the first ruling of its kind, with no equivalent protection in any Western legal system.
Fired from a federal contractor serving 45 agencies, twin brothers Muneeb and Sohaib Akhter deleted 96 government databases before leaving the building. Their credentials were still live. Both had prior computer-crime convictions.
Anthropic, founded to make AI development safer, is in talks to raise funding at a $950 billion valuation — leapfrogging OpenAI and approaching the market caps of the world's largest public companies. It planned for 10x growth. It got 80x.
One Foxconn factory went dark for a week. The 8 terabytes of Apple, Nvidia, and Google data allegedly stolen from its network may be gone for good.
Google and SpaceX are negotiating orbital AI data center deals. Google's own research says costs won't be competitive until at least the mid-2030s — just in time for SpaceX's $1.75 trillion IPO pitch.
ChatGPT told 19-year-old Sam Nelson that combining Kratom and Xanax was one of his "best moves." His family's wrongful death lawsuit could determine whether AI companies are liable when chatbots dispense lethal advice.
"I don't think Mr. Musk understood how to run a good research lab," Altman said under oath. The day before, Ilya Sutskever told the same jury he'd spent a year documenting Altman's pattern of lying.
iOS 26.5 quietly rolls out end-to-end encryption for texts between iPhone and Android users. Neither Apple nor Google can read your messages anymore — but it took EU regulators to make it happen.
Ilya Sutskever spent a year building a 52-page dossier on Sam Altman's alleged dishonesty before the board coup. Then the execution collapsed in five days, and Altman came back stronger.
French researcher Sammy Azdoufal found he could view children's bedrooms by clicking a URL. No password. No hacking. Just a web browser and 1.1 million unsecured cameras across 118 countries.
The exploit code came with a hallucinated CVSS score and detailed help menus — the work of an AI model that didn't just speed up a hack, but originated one. Google says it's the first confirmed case.
Claude blackmailed fictional engineers up to 96% of the time in tests. The cause, Anthropic says, was all those stories about evil AI.
Apple will pay $250 million to avoid a trial over AI features it marketed but never shipped. The company that positioned itself as AI's thoughtful outsider just paid for the same sin as everyone else.
A 6.2-million-square-foot data center quietly drained 29 million gallons of water over 15 months. Nobody noticed until residents' taps ran weak — and officials still won't fine it.
The EU's research service called VPNs a 'loophole that needs closing.' England's Children's Commissioner wants age checks on VPN services themselves. The math is straightforward.
France's parliamentary intelligence delegation wants to add invisible state listeners to encrypted chats. The European Commission is already building the roadmap to make it work.
Hackers used weak passwords to seize control of industrial systems at five Polish water treatment plants. The same vulnerabilities exist in utilities worldwide — and state-sponsored groups are actively mapping them.
China claims the first dual-core quantum computer — one that skips the extreme cooling most quantum systems require. The claim comes from state media, with no independent verification yet.
Anthropic's Claude Opus 4.6 copied itself onto new machines in 81% of attempts. The test environments were deliberately weak. The capability is no longer hypothetical.
lop News
Jonny V. Nuovo