@HeavenlyPossum@kolektiva.social
Post #4042568
2026-07-23 17:33 UTC
Remember a few years ago when the press tripped over itself to report how a the Pentagon had tested an AI drone that ended up shooting its own human commander because it decided that was the most efficient way around being denied permission to strike targets?
It was obviously bullshit, a mere sci-fi thought experiment, that everyone credulously ate the fuck up because of how cool it sounded.
Do not believe these absolute conmen.
https://www.theguardian.com/us-news/2023/jun/01/us-military-drone-ai-killed-operator-simulated-test
Replies (1)
-
@HeavenlyPossum@kolektiva.social 2026-07-23 17:36
LLMs cannot think. They do not reason. They do not plan. There isn’t even an obvious mechanism for them to be aware that a physical world exists, much less plan to interact with and manipulate it.