Marc Andreessen’s influential essay, It’s Time to Build, is sort of a fusion of techno-optimism, YIMBY-inflected arguments, and the despair associated with the early months of 2020. In it, he lambasts America’s inability to ‘build’, by which he apparently means primarily physical things — at the time test swabs, ventilators, and vaccines were top of mind. Andreessen blames many agents for this: regulatory capture, neoliberal offshoring, political partisanship, culture in general.
Of course, this essay created backlash. Many expressed derision at a Silicon Valley software founder turned venture capitalist making blustering on about the physical world. It’s definitely possible to read an undercurrent of software-couched condescension in the piece.
But what’s striking to me is that Andreessen never needed to venture outside of software to make his point — in fact, given his background in software I’m surprised he didn’t take this approach. He does touch on this near the beginning (and unfortunately doesn’t build on it):
In the U.S., we don’t even have the ability to get federal bailout money to the people and businesses that need it. Tens of millions of laid off workers and their families, and many millions of small businesses, are in serious trouble *right now*, and we have no direct method to transfer them money without potentially disastrous delays. A government that collects money from all its citizens and businesses each year has never built a system to distribute money to us when it’s needed most.
This is definitely a software failure! The inability to process the early-pandemic surge of unemployment claims was famously caused by outdated, fragile computer systems used by the government agencies responsible for disbursing funds.
I like the software angle because it simplifies a lot of things (and I guess it puts it more in my wheelhouse too). With building transit many will argue that America sucks at building transit because we don’t have any expertise in it, and we’re failing to hire foreigners to show us how it’s done. Fair enough. But software? We have a world-leading private sector in software. And nobody is blaming neoliberalism, partisanship, or culture for holding it back.
The culprit here is government capacity: the ability of the state to do the things it’s decided it wants to do. I don’t think there’s an arch-enemy to blame here. In the case of software, building good software is just hard and the government isn’t good at it.
Make Bureaucracy Great Again
I’ve never seen someone campaigning on improving the quality of bureaucracy. It’s kind of quaint to think about — “if you elect me all of the government websites will have fewer bugs and the line at the DMV will be half as long.” But, really, the whole point of government is to offer useful services to people. It feels a little sleepy, but I think this would be a great thing for Mayors and Governors to campaign on.
Do Mayors and Governors have the ability to fix the bureaucracy? I’m optimistic! In Recoding America (highly recommend), Jennifer Pahlka discusses some of the ways that government fails to build good software. One of the key issues she identifies is culture — essentially, civil servants exist in a risk-culture where orders come from the top and are followed exactly. If we could instead encourage a culture where civil servants up and down the chain felt empowered to report what wasn’t working and to make changes with autonomy, we might well have better government services. These culture changes could come from the top too, from elected executives trying to make improvements.
These changes aren’t easy per-se, but it’s also not the case that bureaucratic mediocracy is inevitable. In reality this is well within the scope of achievable political change. This is something we can demand of our elected representatives, and that they can deliver for us. So let’s do it!