2 Comments

I think you greatly overestimate how much DARPA spends on research. I personally work on self driving cars and I would guess DARPA has spent maybe 10 million for the Self Drving Car challenge. Their Humanoid robotics was a much cheaper affair (they only sponsored 4 robots, each a million) wouldn't be more than 5 million, their subT challenge currently going on is even cheaper. It seems like they only fund the prize money now and teams need to fund themselves so 1-2 mill. When I worked with DARPA project managers, they could write a check till around 10,000$ on the spot to buy equipment. Anything more needed to go through the required filing (applying for a tender, getting supervisor approval etc) and each project manager had access to only 1-2 million. They mostly used existing University equipment and used their money to pay salaries for research and technical associates. If DoD gets money it goes to Boeing/Lockheed or insane boon doggles like the 1.7 Trillion F-35 which doesn't show any improvement over the F-22 raptor. Surely if they even spent 1 billion of this insane 1.7 trillion on self driving cars, flying planes, Storr would be quite happy. At least they fund Space X now

Expand full comment

Fascinating! And shows the delta between top level resource allocation and what that actually means in ground reality.. It's one of the biggest issues with these ROI style calculations, since the spend is so difficult to disaggregate.

Expand full comment