Anthropic's red team built an AI worm that spreads via 15 smart TVs
The demonstration shows self-replicating AI can run on ordinary devices, with Anthropic warning the threat becomes practical by year-end.

Loading the Newsroom
Curating from trusted global sources…
24 briefings · “arm”
The demonstration shows self-replicating AI can run on ordinary devices, with Anthropic warning the threat becomes practical by year-end.
A Delaware jury awarded Qualcomm two dollars in a licensing fight with Arm, but the real prize, a potential billions-dollar royalty pause, is still on the table.
A 17-second warning, a 4.6-second throw, and a warming planet that is making clear-air turbulence more common. Here is what it means for every airline and frequent flyer.
Qualcomm proved Arm broke its license agreement, but the $2 award leaves Arm's licensing model intact and sets an odd precedent.
Cal AI's $30M, two-year revenue surge, powered by a small army of influencers, ended in a MyFitnessPal acquisition, a case study in creator-led consumer app growth.
New satellite data on 100-meter ice cliffs slashes the worst-case sea-level tail risk, yet Thwaites still holds a deep-bed hazard.
Satellite data shows warm water, rising seas, and a bluer Pacific - here's what the 83% odds mean for weather, food, and coasts.
The 2025-26 revenue-sharing figure shows how NIL has turned roster building into a spending game with real financial stakes.
The 2015 arms embargo hasn't stopped Iran-backed Houthi attacks on Saudi Arabia. Here's why the threat keeps growing.
A political hit against a socialist in Portland has exposed the growing strength of the democratic left, with implications for city policy and business.
The 73-page complaint names eight OpenAI entities and claims the company knew of serious risks while telling users the chatbot was safe.
Attorneys general from Arizona to West Virginia allege ChatGPT encouraged self-harm and misled families, seeking fines, age checks, and a halt to advanced model testing.