AI is Now Being Used in War

TL;DR

The US military used Anthropic's custom Claude AI to strike 1,000 Iranian targets in 24 hours and capture Venezuela's Maduro.

Key Points

• 1.The Maven Smart System, built by Palantir and powered by a custom Claude model, pulled data from 179 sources including hacked Tehran traffic cameras to prioritize Iranian targets in real time.
• 2.Military AI is vastly more powerful than consumer AI — dedicated data centers give the military 5–6 orders of magnitude more compute per request, allowing one 20-person unit to do the work of 2,000 staff.
• 3.The Pentagon demanded Anthropic allow Claude for all lawful purposes, including mass surveillance of Americans using geolocation, browsing data, and financial records bought from data brokers — Anthropic refused and was labeled a "supply chain risk."
• 4.Hours after Anthropic was cut off, Sam Altman publicly offered OpenAI to replace them — a deal rushed in just 2 days — triggering mass backlash and an estimated 2.5 million users cancelling or boycotting ChatGPT.
• 5.Over 450 Google and OpenAI employees signed an open letter opposing their own leadership's decision, and a real-world protest erupted in San Francisco under the "Quit GPT" movement.
• 6.The surveillance threat extends beyond war: Meta's Ray-Ban glasses have an internal feature called "Name Tag" that identifies strangers in public, and Palantir has already installed surveillance systems in Australian supermarkets.