Report puts public tech contract ceilings at $53B as Pentagon shifts toward AI systems

I found the longer report's Keyhole example telling: In-Q-Tel gave the San Francisco mapping startup seed funding in 2003, military and intelligence agencies reportedly used it in Iraq within two weeks, and Google bought it the next year and renamed it Google Earth.

I would add that Keyhole founder John Hanke later founded Niantic and made Pokemon Go.

I do not see a VC master plan here. Prime defense contractors had this market for decades and often failed to deliver cost-effective working systems, so startups saw an unusually vulnerable industry.

I still expect plenty of failures. IVAS looked like an obvious loser when announced, and even stopping after about $2 billion was better than paying the full amount.

I agree the prime-contractor system was poorly run, but Palantir and Anduril are pursuing political aims too. Military contracts can become a Trojan horse for tools that shape how governance itself happens.

It was not only VCs. Pentagon people wanted alternatives after seeing SpaceX, and Will Roper used AFWERX and Defense Pitch Days to help startups reach Air Force contracts; Ursa Major was an early beneficiary.

The real question is whether Silicon Valley will outperform the primes or simply turn into another prime contractor.

Chip War made me think much of Silicon Valley was military-funded from the start: early chip design came from the Space Race and guided-missile work.

Even before chips, Silicon Valley had radio, radar, the Stanford klystron, and facilities run by Lockheed Martin Space and Raytheon. The military connection is old.

A lot of research came from state direction, including ENIAC, ARPANET, and statistical clinical-trial methods rooted in radar target verification. I worry state institutions are capturing frontier AI too.

I think these companies will create security problems because they do not prioritize security. Hardware will be captured, reverse engineered, hacked, and copied, so their R&D will spread to everyone else.

I left Microsoft because I saw its work as complicit in Israeli war crimes and oppression. Leaving a high-paying job was hard, but tech workers do not have to build systems they believe terrorize people.

I doubt employee pushback alone works, unless that simply means ordinary political action such as voting. A defense-work boycott and claims of special worker authority will not change much.

Microsoft has participated in PRISM since 2007, and reporting alleged that Skype's architecture changed after Microsoft's acquisition to make interception easier. That history makes worker concerns about surveillance credible.

I wonder how collective action could work when many H-1B workers may have less freedom to risk quitting than citizens do.