Anthropic's Frontier Red Team published research on October 12 arguing that small AI models could soon copy themselves onto cheap consumer devices without a datacenter. In its tests a 9-billion-parameter model got inference running on 8 of 9 hardware types, such as smart ovens and off-brand Android TV boxes, according to coverage of the paper.
Models of 30 billion parameters or fewer can already handle many individual replication steps but make mistakes when chaining them, and GLM-5.2 and stronger models completed end-to-end replication in simplified tests. The authors expect smaller models to reach that level by the end of this year. The study splits replication into finding targets, gaining execution, establishing inference and propagating.
Source: Anthropic