I think the West has created something incredible. We’ve had a period of unprecedented peace and unprecedented human rights. It’s not perfect, but it’s better than anyone else has ever had. So the US and its allies are tasked with protecting that…
I’m terrified over the years of a weird dynamic developing in the West. All the big defense work unapologetically goes to Northrop Grumman, Lockheed, Raytheon, those guys — they’re very good at making fighter jets, submarines, aircraft carriers, and iterative stuff. They’re not good at artificial intelligence, or computer vision, or sensor fusion. It’s just not what their talent is.
On the other side you have the big tech companies, and they have all the talent in this space — big tech and all the other little companies doomed to be absorbed by big tech. But those people refuse to work on military problems, partly for ideological reasons, partly because it makes it hard to play in the commercial space, and partly because they’re afraid of tech media which is by-and-large anti-military. So who is going to work on these problems?
It’s worth noting, if you look at the entire defense landscape, in the last 30 years there’s only two unicorns that have emerged. Two private companies with valuations over $1 billion that got that way by doing business in the defense space. Now, what do both of those business have in common. They were founded by two billionaires who got into the space because of ideological reasons.
There’s this great joke. How do you become a millionaire in the defense business? You start out as a billionaire in defense work!
That was Palmer Luckey interviewed on Collision, “Building an AI arsenal.” HT: Richard Wahidi. I think it’s interesting Luckey described fighter jets and subs as primarily “iterative stuff” when they are made in such small quantities, and the most of the value generated isn’t in bending materials but in the product design and supporting business process. These intangible attributes of major weapon systems outweigh the assembly-line reproducible aspects.
Here’s a bit more:
China is very good at innovation in their technology community. If you look at their top defense companies, it’s not those that have been around for 50 or 60 years, it’s newer companies. They are very good at fostering innovation and pipelining stuff from the consumer products sector to their military sector.
From my count, 5 of the 8 Chinese defense firms with $5 billion in sales were formed in 1999 or after. Two others were formed in the 1980s. I’m sure most of these firms formed out of pre-existing firms, but still, Chinese defense firms are generally much younger. And as Ben Horowitz said, it isn’t whether a firm is big or small that contributes to innovation so much as whether a firm is old or young.
A lot of this just comes down to comparative wages – software engineers who could have jobs at the big tech companies can demand *huge* compensation (see https://www.levels.fyi for numbers), and neither the government nor its contractors are willing to match that compensation.
Mechanical, electrical, industrial, etc. engineers don’t have that outside option, so wages at government contractors are still pretty attractive for them.
I know multiple engineers I went to school with who would *love* to do defense work, but making the switch would cost them literally hundreds of thousands of dollars per year in foregone compensation. So they’re all working at big tech companies instead.
So the government gets the output of some of the best and brightest engineers that make physical things, but isn’t attracting or retaining the talent it would need for best-in-class machine learning systems.
Yes. Armen Alchian and Kenneth Arrow argued that govt shouldn’t hesitate to pay top dollar for top S&T talent. Some industry firms stated back in the 1950s that they were at a disadvantage to govt because workers there felt they served a higher purpose and were able to make big decisions. That is no more. Perhaps all the govt needs to attract top talent — even if it is just those like Palmer Luckey who got rich in the private sector and want to try something big and challenging — is to give managers more authority. Currently, govt employees are largely working to the baseline plans set by others, either before them or from in the bureaucracy. Give them some self-determination!
More self-determination would be huge. Another critical factor is an efficient way to weed out low performers. It’s very demoralizing to work extremely hard to hit ambitious goals if you can see your cubemate doing literally nothing all year and getting the same paycheck. Managers are currently nearly powerless to address this.
Significantly more money in exchange for significantly less job protection would be a good trade.