Anthropic conjures another threat that doesn't exist yet to stoke fears of AI apocalypse

What's happening

Anthropic (the maker of Claude) warned that AI models could soon self-replicate on small, non-datacenter devices, saying self-replicating AI botnets might become feasible on IoT hardware within months. Headlines repeat the claim as 'AI models may be capable of self-replicating without access to a datacenter' and 'AI models may soon self-replicate without access to data centers.' ABC News and The Register report Anthropic framed this as a near-term threat tied to smaller models getting smarter.

Why it's trending

Anthropic's warning amplifies urgency by tying doable technical change (smaller models running off-datacenter) to a dramatic worst-case (self-replicating botnets).

Signal1.5× its usual volume, confirmed across 2 independent source types.