Yikes.

Video summary

ColdFusion traces Claude’s role in military targeting, Anthropic’s refusal of surveillance terms, and OpenAI’s controversial Pentagon deal.

ColdFusion examines how AI moved from consumer chatbots into military operations, including the reported use of Anthropic’s Claude in the capture of Venezuela’s Nicolás Maduro and the Iran campaign. Palantir’s Maven Smart System combined Claude with data from 179 sources to rank targets; a Georgetown study says one artillery unit could do the work of 2,000 staff with a 20-person team. Anthropic had secured a $200 million Pentagon contract but rejected demands for “all lawful purposes,” insisting on safeguards against mass surveillance of Americans and autonomous lethal decisions without human oversight. The government then cut off its agencies’ access and labeled the company a supply-chain risk. Hours later, Sam Altman said OpenAI would take a Pentagon deal, prompting employee criticism and a QuitGPT backlash; organizers claimed 2.5 million people took action. Dagogo Altraide warns that AI-enabled surveillance could extend beyond war, citing data brokers, Palantir systems in Australian Coles stores, and Meta’s planned facial identification on Ray-Ban glasses. He urges stronger laws and limiting personal data exposure.

Chapters

1. 0:00AI in modern warfare: Claude used in Maduro capture as Anthropic rejects Pentagon demands
2. 1:08OpenAI enters the scene: Sam Altman offers ChatGPT maker for Pentagon deal
3. 2:08Military vs. consumer AI: Anthropic’s dedicated classified model has far more compute
4. 4:49Maven Smart System operation: Palantir and Claude process 179 sources for targeting
5. 7:47Anthropic’s refusal of terms: Amodei rejects Pentagon demand for unrestricted Claude use
6. 10:18OpenAI contract controversy: two-day Pentagon deal fuels Quit GPT backlash
7. 12:23Future risks and protection: data brokers, Coles surveillance, and Meta’s Ray-Ban name tag