"coordinating workflows and recovering from failures without human intervention" [without human intervention]
The explicit claim of autonomous failure recovery and workflow coordination without human intervention directly supports the signal's claim that end-to-end targeting autonomy is being engineered out of the human-in-the-loop model.
Reasoning from this article
The article presents two parallel developments — a classified US AI targeting system and a publicly documented Chinese equivalent — as evidence that autonomous military AI is no longer a theoretical concern but an active capability race. The Chinese system's public disclosure is framed as a transparency move, but structurally it also functions as a capability signal to adversaries. The pattern generalizes: as LLM agent architectures mature commercially, military actors on multiple sides are integrating them into targeting pipelines, eroding the norm of human authorization for lethal force.