tech
March 2, 2026
No one has a good plan for how AI companies should work with the government
As OpenAI transitions from a wildly successful consumer startup into a piece of national security infrastructure, the company seems unequipped to manage its new responsibilities.
TL;DR
- OpenAI CEO Sam Altman fielded public questions on X about the company's decision to accept a Pentagon contract previously declined by Anthropic.
- The contract involves potential participation in mass surveillance and automated killing, activities Anthropic refused.
- Altman stated his belief in the democratic process and the role of elected leaders, but was surprised by the extent of public disagreement.
- The situation reflects OpenAI's transition into national security infrastructure and its apparent unpreparedness for the associated responsibilities.
- The Pentagon's blacklisting of Anthropic for contractual limitations and OpenAI's subsequent acquisition of the contract has generated significant blowback.
- OpenAI's engagement with the government has shifted from a consumer-focused approach to a more serious engagement due to AI's power and capital needs.
- U.S. Defense Secretary Pete Hegseth's plan to designate Anthropic as a supply chain risk poses a significant threat to the company.
- This move against Anthropic is seen as a chilling message to other vendors and could damage the broader tech industry.
- OpenAI faces internal pressure from employees regarding ethical lines and external pressure from right-wing media regarding political alliances.
- AI companies like OpenAI are increasingly forced into the same game as defense contractors like Palantir and Anduril, requiring them to pick sides politically.
- Tech investors in influential positions seem content with 'tribal logic,' where companies perceived as favoring one administration are penalized.
- Unlike traditional, slow-moving defense conglomerates, modern AI startups are faster but less prepared for long-term political shifts.
- There is a lack of clear planning for how AI companies should collaborate with the government.
Continue reading the original article