AI Could Help Produce Weapons, Anthropic CEO Dario Amodei Says

I feel that a lot of companies developing closed AI models are trying to overhype "AI Dangers" to reduce competition from open-sourced or permissively licensed LLM projects.

Rogue actors are a dangerous possibility, but even more certainly AI will be used by militaries in an official capacity to control weapons, strategy and tactics.

The CEO of Palantir just published an open letter calling for a Manhattan Project for superintelligent AI weapons. https://archive.is/KSOv4

It seems to me that the plan is: the West and East each create their own SkyNet connected to their military arsenals, have them fight it out, then if there are any humans left alive who are still able to control the AIs, those are the "winners".

I mean that's not an explicit plan yet but without another way to resolve geopolitical conflicts, it's the path we are on.