gabriel(@g_br_l):cool thread. IMO the most realistic risk in AI is not very likely now, but might be at some point: self-replication right now is constrained by limited compute power. if an agent figured out a way to significantly reduce its own need for compute, replication becomes possible.

2026.09.16 23:16

cool thread. IMO the most realistic risk in AI is not very likely now, but might be at some point: self-replication right now is constrained by limited compute power. if an agent figured out a way to significantly reduce its own need for compute, replication becomes possible.

Fraser Price@fraserpricee

2026.09.16 22:54

Self-replicating, multi-agent AI systems are an overlooked existential risk. I want to lay out why I believe this, via example. Each dot is a simple creature with a randomly evolved NN brain. There is no training; yet complex, unpredictable survival drives always emerge. 1/14