@jonny @joepie91 I think the technology will eventually get there but it's more an irritating situation than an apocalypse. Think Metasploit proxy chains, but with more reach. Big model writes exploits and tooling, medium models do some of the easy dev and manage topology, small models just babysit shell scripts and do some amount of filtering events and data to send upstream.
Oct 17, 2026, 3:25 AM@jonny @joepie91 That's the interesting part, at some point we cross the capability threshold where the worms can self-maintain without a "someone" in charge, no chain of command. No off switch and no automatic loss of viable hosts. Then some bright people will propose gene drive style eradication and things could get exciting. But most likely just more annoying.
Oct 18, 2026, 1:22 AM@mirth
@joepie91 that part is still citation needed for me. The critical question is sharding coordination, not model capability. And a shared model bootstrapping and maintaining coherent sharding for itself and its clones is a not trivial hard tradeoff between the capability of the model and the difficulty of coordinating the swarm. The plausible models here can't even install the bundled model which is like 4 shell commands, could glm size model autonomously shard over thousands of hosts and remain coherent? I'll believe it when I see it