silicon nation

What Silicon Valley Doesn’t Get About Government Tech

Wired’s Steven Levy has a long, detailed story today about the salvation of, the Obama Administration’s once-doomed online health insurance marketplace, by a handful of Silicon Valley coders. 

According to Levy, the ragtag band of coders was assembled, Oceans 11–style, from Google, Stanford, and a host of West Coast start-ups, and set loose on fixing the troubled site. What resulted was “something of a marathon,” with coders racking up Hilton points and working around the clock to overcome “a red-tape nightmare” with “frustrating obstacles” and “poor design.” The resulting site, Marketplace 2.0, bested the original on all fronts, Levy writes, and “provided a blueprint for reforming long-moribund government IT.”

Levy writes that, at one point in the turnaround, “both sides realized they could learn from each other.” But the process he describes is mostly a one-way education: young techies teaching government bureaucrats how to build and maintain big websites using the best tools available. Among other things, the Silicon Valley techies taught the D.C. crowd to “overcome cultural resistance” to storing health care data on non-government servers, and convinced them to move parts of onto Amazon Web Services’ cloud-based platform. They also rebuilt’s sign-up system, trimmed down the number of steps required to sign up for coverage, and fixed some fairly obvious flaws in the site’s front end:

For example,” one engineer says, “you had to choose a user name that had a special character such as an underscore or a dollar sign, which normally is a requirement applied to passwords, not to user names. So a lot of people were confused by that.”

Levy’s story is a good blow-by-blow on the process of fixing a website cobbled together through the government’s legendarily bad procurement process. As I wrote last year, Silicon Valley was frustrated by the initial failures of and wanted to pitch in to help. And even though it’s arguably a series of management changes, not an infusion of tech talent, that led to the site’s rebound, there’s no denying that the tech community did supply some competent assistance.

But to the extent that Levy’s story is a victory lap for the tech community, it’s also an illustration of how woefully incomplete the sector’s understanding of the scale and importance of government is.

The tell is in statements like this one, from one of the engineers who was brought in to fix

People are more scared of things here,” says Ben Komalo, a Marketplace Light engineer who recently returned to his job at Khan Academy. “The costs of failure are perceived as being much higher than where we’re from.”

I’ve heard this idea dozens of times in conversations with Bay Area tech gurus — the notion that if government just ran itself like a start-up, with an emphasis on tech concepts like MVP (minimum viable product) and back-end products like Amazon Web Services, its problems would be solved overnight. These people are baffled that given the choice between using a process that could produce scalable projects quickly and cheaply and one that is laden with compliance and security obstacles, the government would pick the latter.

It’s true that Silicon Valley largely embraces failure. But it can afford to do so because, in many cases, the stakes aren’t that high. If Komalo pushes some bad code at his Khan Academy day job, the site’s users might have trouble loading their Greek History lessons. But if, say, a Department of Defense code base hiccups and leads to a major security breach, it could create a political disaster with far-reaching consequences. Heads could roll; lives could be lost. That’s a lot of what makes government IT contractors “more scared” than start-up daredevils — the costs of failure aren’t just “perceived as being much higher” where they work — they actually are much higher.

Don’t get me wrong. Bringing in hired guns to fix was indisputably good, and I hope that the experience of turning around a broken and convoluted engineering project leads more Silicon Valley coders to consider working in the public sector. We do need to fix government IT, starting with the procurement process, and many of Silicon Valley’s organizing principles can inform the way government creates and manages tech-driven projects.

But if they do come to Washington, I hope Silicon Valley engineers will bring a bit of humility, and some awareness that government doesn’t just create bureaucracy and red tape for the sake of it. There are real risks — political, legal, technical — involved in carrying out our national tech challenges, and while the process of creating sites like can (and should) be sped up and refined, those sites may never move as quickly as their private-sector counterparts. And that’s okay. “Move fast, break things” may have worked for Facebook, but it’s a risky way to run a country.

What Silicon Valley Doesn’t Get About Gov. Tech