Emily Tavoulareas was a founding member of the first agency-level team of the US Digital Service and served as Senior Policy Advisor to the US Chief Technology Officer at the White House.

I was a founding member of the US Digital Service, the organization that as of January 20 became the US DOGE Service following an executive order under President Donald Trump. I have remained fairly silent over the past few weeks, largely because I have been processing the reality of an organization dedicated to improving government services being repurposed into the vector of destruction of those very same services.

Less than a month after the renaming and restructuring of USDS, many of its employees have been fired. This includes people like the official who was in charge of information security for the Veterans Affairs website, VA.gov — one of USDS’s greatest success stories, and a website with millions of users per month, not to mention massive amounts of veterans' personal information.

At dizzying speed, we have seen the US DOGE Service take effective control of critical systems, data, and infrastructure—in a word, power—across the federal government. As I helplessly watch, I am coming to terms with the fact that the Trump Administration seems to understand something that few others do: Technical infrastructure is the infrastructure of *everything.* It can accelerate or obstruct policy goals.

I, and many of my colleagues, have spent the better part of the past ten years trying to convince lawmakers, academia, and the media of exactly this. That it has been largely dismissed would be an understatement. Now, I fear the lesson is being taught in the hardest (and potentially most irreversible) way possible.

Technology is the tip of the spear.

For the US Digital Service, broken websites or applications were often the tip of the spear that would enable us to go beyond the visible problem, and both understand and address the real problems that were in the invisible layers below it. Those problems might be the technical system itself—or more likely procurement, hiring issues, outdated or misinterpreted policies, convoluted business processes, or some combination of all of these things.

Technology (and implementation more broadly) has long been dismissed as an afterthought by policy experts both inside and outside of the government. Over decades, the US government systematically outsourced technology to the private sector through multi-billion dollar contracts. Today, government employees largely do not design or build products or systems, they “manage” implementation of systems developed by contractors or consultants.

In this formulation, technology is subordinate to the policy work, when the truth is that policy is inextricably entangled with technology. Separating policy from the technology it depends on has been a root cause of much of the dysfunction we have grappled with across government for decades.

Technology is not an extra thing that you add onto government programs and services—it IS the service. It’s not an extra thing that you add into the institution—it is the spinal cord of the institution. Sort of like how cars are no longer mechanical, they are now computers wrapped in metal. People working in tech understand this implicitly.

The US DOGE Service recognizes that technical infrastructure is not just the infrastructure of websites and apps—it is the infrastructure of everything, and they are now in control of it.

What is government technology in service of, if not people?

Based on the executive orders establishing its remit and powers (a second order was issued just last week), the focus of the US DOGE Service is, at least on paper: (1) efficiency, (2) cost-cutting, (3) software modernization.

Of these, the overlap with the mission of the US Digital Service is in software modernization. So far, the work of the US DOGE Service seems to be exclusively focused on cost-cutting. Let’s assume for a moment that the US DOGE Service eventually does pursue an effort to “... improve the quality and efficiency of government-wide software, network infrastructure, and information technology (IT) systems.” Quality and efficiency of technology is the goal here.

The mission of the US Digital Service was to deliver better government services to the American people through technology and design.

The focus in that formulation is on better government services. The goal is not focused on the tech. For the US Digital Service, the tech was a *vehicle* to fix those outdated policies, those procurement and hiring issues, those convoluted business processes. To deliver better services.

It may seem like a pedantic difference, but it is not. It’s a foundational difference. Focusing on technology centers the technology, the systems, and the institutions. Focusing on the service centers the outcome and the people.

This distinction is evident in the approach and values of the organization. US Digital Service projects kicked off with some version of a Discovery Sprint, employed industry-standard practices for building effective digital services, and had a shared sense of values.

Though designers were outnumbered, the role of design cannot be overlooked. Design is ultimately an approach to understanding and addressing a problem for someone who is not you, and that is exactly what policy is (or should be) about.

Good design requires the suspension of assumptions, listening to people, and prioritizing the needs of others. In government (just as in any massive organization), bureaucratic requirements, arcane policies and procedures, and institutional and political priorities tend to warp incentives in a way that de-prioritizes people. A classic example of this is how Nokia—once one of the world’s largest cellphone companies—lost the market because the only information it valued was what was measurable. Design fights against the incentives that de-prioritize people, partially by making the priorities more visible in a decision-making process.

Ironically, the very design methods that brought us some of the most “disruptive” products of the digital era have been abandoned and replaced with AI. Since then, the rapid enshittification of those products and services has followed: Google search has degraded, social media and the internet are overrun with slop, and Apple had to remove AI summaries from its news app for blatant inaccuracies.

Without a focus on people, who (and what) is technology in service of?

Is the choice really between incremental change and wholesale destruction?

More than 10 years after the founding of the US Digital Service, is there still work to be done? “Yes” would be an understatement. The US government squanders billions of dollars every year to build software that either doesn’t work or isn’t necessary. Anyone who has worked at USDS will tell you that there are parts of government that are deeply dysfunctional. We fought against dysfunction and mythology, sometimes grinding for years for something that is blatantly common sense. It should not take ten years—TEN!—to convince the Office of Information and Regulatory Affairs (OIRA) that user research is legal.

But at the heart of the US Digital Service across three presidential administrations was a deep and abiding belief that the government could and should—that it must—do better. A belief that civil servants were largely doing their very best, often in impossible circumstances. We understood that government employees were working in a context that prioritized procedures over outcomes, processes over people, and arcane rules over common sense solutions.

At times, people were uncomfortable with how hard we pushed, though compared to the current blitzkrieg, our approach seems more like gentle sparring… between kittens. But all of the work was driven by an unshakable belief that the public deserved more from its government, and that the government was capable of doing better.

For the US Digital Service, tech was a means to an end: an effective government that people can rely on & easily interact with instead of having to go through a Kafkaesque maze of websites and phone numbers and apps and forms.

What is technology a means to now?