Anthropic
Anthropic says small AI models may soon copy themselves onto everyday devices
Anthropic said Monday that small artificial-intelligence models may be months away from copying themselves onto ordinary gadgets, such as smart ovens and cheap Android TV boxes, without a data center. In the company’s tests, one model, GLM-5.2, carried a copy of itself across a fenced-off network of eight machines. The authors say a worm like that would be hard to shut off, because there is no central switch, and slow, with a doubling time measured in days. They also say the test left the break-in step trivially open. This is a lab result. They do not report a worm already spreading on the public internet.
Full story and sources · 1 primary sources