Tech Won't Save Us - DOGE Is Gutting the US Digital Service w/ Kate Green and Milo

Episode Date: March 27, 2025

Paris Marx is joined by Kate Green and Milo to discuss how Elon Musk’s DOGE is dismantling the US Digital Service and the consequences that has on really improving service delivery in government.Kate Green and Milo are former US Digital Service engineers.Tech Won’t Save Us offers a critical perspective on tech, its worldview, and wider society with the goal of inspiring people to demand better tech and a better world. Support the show on Patreon.The podcast is made in partnership with The Nation. Production is by Eric Wickham.Also mentioned in this episode:Find out more about We The Builders and read perspectives of federal workers.USDS workers have resigned en masse from DOGE, and Musk’s agency has fired many of those who haven’t left.The General Services Administration has also cut the 18F technology team, and has been doing mass layoffs at the Technology Transformation Service. Support the show

These people are working their butts off, and the disrespect that's implicit in how they treated those people, my colleagues, my fellow Americans, just pisses me off. Hello and welcome to Tech Won't Save Us, made in partnership with The Nation magazine. I'm your host, Paris Marks, and this week I have a very special episode for you. My guests are from a group called We The Builders, and they're Kate Green and Milo, who are former U.S. Digital Service engineers. Now, you probably heard that when Donald Trump took office, he basically took over the U.S. digital service and renamed it the U.S. Doge Service for Elon Musk's Department of Government Efficiency. That has obviously had

a lot of consequences for that agency, but also the people who work there. These are people who are used to developing software know, developing software and developing solutions for government and, you know, for the public that it serves to make sure that public services can be more easily accessible by the people who need them. But when Elon Musk and his people took over, they had a very different idea of what this agency should be doing, who it should be serving, and the way it should be oriented toward government. And so I was really happy when Kate and Milo reached out and said that they wanted to speak to me for the show, because not only did it give me an opportunity to learn more about what

actually happened in that transition from the U.S. Digital Service to Doge, but also what kind of orientation a department like that takes in trying to develop software in the public sector that is oriented towards serving the public. Is that going to be the same as software development and technological development in the private sector, or how is it going to look different? And obviously, I expected it to be different, but the way that Kate and Milo described what their jobs look like and how they had to think through the various implications of the software that they were developing and the solutions that they were trying to deploy really made me stop and think why these kind of approaches are not more common.

And, you know, obviously we talk about it in the episode, there's a profit motive in the private sector, and that leads to very different incentives and actions. But I certainly think that if we were trying to think of how technology could better serve the public, could be better developed in order to have the public in mind, it's this kind of orientation that Kate and Milo describe that we should be striving toward and that we should be thinking about. And yeah, it might take a little bit more time to do something, but then we're probably doing it right. And we're actually taking into account the needs of the many different people who are using these platforms and tools and software.

So needless to say, I really enjoyed this conversation with Kate and Milo. I felt like I learned a lot and that I'm going to be reflecting on this conversation for some time to come, I think. And I would certainly recommend checking out We the Builders. I've included the link in the show notes where you can see some of the stories from other U.S. digital service and other federal workers to give some insights into what is happening in this moment and, you know, the type of work that they were doing. So if you enjoy this episode, make

sure to leave a five-star review on your podcast platform of choice. You can share the show on social media or with any friends or colleagues who you think would learn from it. And if you do want to support the work that goes into making Tech Won't Save Us every single week, so we can keep having these critical, mind-expanding conversations. You can join supporters like Austin from Wisconsin, Kilgore Trout in Melbourne, Selah in California, Kimberly in Denver, Katie in Oakland, Doug from Missouri, Alex from Maine, Hank from Orlando, Florida, Mark from Canada, and Ryan from Cambridge, Massachusetts by going to patreon.com slash techwontsaveus where you can become a supporter as well. Thanks so much and enjoy this week's conversation. Kate, welcome to tech won't save us. Hey there, Paris. How are you doing?

Good. It's great to have you on the show. Milo, great to have you too. Thanks so much. Thanks for having us on. I'm really excited to speak with you both because obviously, you know, we're seeing everything that's happening in the US government right now and how the US digital service has become at the center of these things. But even for the type of show that I run and the interest that I have, I think the idea of developing technology in the public sector and what that actually looks like is something that is really interesting to me and interesting to a lot of my listeners, right? How this might be done in a slightly different way and how it differs from how it's traditionally done in like a Silicon Valley company or something like that. So I'm

excited to explore all those different facets of this issue with you. And I just wanted to start by asking if you could talk a little bit about your roles at the U.S. Digital Service when you were there, the types of things that you were doing as much as you can say that, and then we can go from there. I'm a former engineer, so that meant that I did work around the engineering, but not always writing code. Some of my projects were helping with the front end and data work around the engineering, but not always writing code. Some of my projects were helping with the front end and data structures around the IRS direct file project. If you're using direct file this year, the data structures that are in place is something that I

helped contribute how it's how it works. That was really cool. And then I was led the testing effort for the pilot last year. Other than that, I worked on a disaster recovery data sharing application, working to get your data from one agency to another because there are 14 agencies that are involved in disaster recovery assistance. I also worked on a project to help people verify their income for SNAP and Medicaid. And I worked to design a system for refugees to access the status of their application, which sometimes takes like two to three years. Used to be you get a piece of paper and that's how you keep track of it. And now they've advanced it before they killed the program. So all the things in common for me, it was a lot of forms and a lot of people who really need to be able to access the systems and the government. That mirrors my experience as well.

And one of the things that we really tried to impress upon people when they came into USDS, no matter what role that they were going to be in, was that they may not be doing engineering. They may not be doing explicitly designed work or product work. There might be some kind of leadership role that they have to inhabit as well. So they might end up being a team lead. They might end up having to sit across from the secretary of whatever agency they're at and explain to them why a project was in trouble or what the needs of the project were. And so we had this intention of bringing in generalists who often had non-traditional backgrounds and had the ability

to be resilient and pivot to meet whatever kind of technical challenge they were facing. Yeah, that sounds really fascinating. And I guess you're seeking out and kind of helping people develop these kind of broader skill sets, I guess, as well, right? To be able to do these different sort of things, it sounds like. Before a few months ago, I think there would have been a lot of people who maybe had never heard of the U.S. Digital Service or were not very familiar with it. And all of a sudden, it is this thing that they're finding out about because it's being taken over by Elon Musk's Department of Government Efficiency and renamed and things like that. How would you describe what the U.S. Digital Service was and what it actually did within government? I think that the initial

concept of the U.S. digital service was this idea of digital firefighters. So during the problems with healthcare.gov, USDS was created to come in and solve a very specific issue with the healthcare.gov website that was a major embarrassment for the Obama administration. And the idea was that a lot of agencies didn't know how to do technology products, technology projects correctly. So USDS kind of started in this, let's have people from Silicon Valley leap in, do a short term, and then help get projects back on their feet. But over the years since then, I think that the scope and the mission changed a little bit and grew into something where a lot of the people at USDS viewed the mission as helping agencies

learn how to do projects long-term and then stand up their own digital services, then be able to do the work of building technology on their own without needing a USDS. And so a lot of people ended up moving into other branches of government and becoming part of those digital services. And so a great example is the Veterans Administration. They had a lot of issues very early on with people trying to apply for benefits. For example, vets couldn't get the services that they need and the benefits that they were owed because a lot of the forms that they had to fill out were PDF.

There's a really famous use case from early on in USDS where it was discovered that the PDF didn't even work with all versions of Acrobat. Like veterans would try to go to their library to fill out the form and the computer at the library couldn't even render the form so they could fill it out. And so USDS did some really great discovery. A lot of conversations with vets took some video of problems that vets were having. And that actually ended up spurring a real desire to improve the experience for veterans in the United States. And now the VA has a really fantastic digital service team. They're doing amazing work. We saw during COVID, for example, they stood up a service on the phone where a vet could text the VA for an appointment and it would respond back.

And it was like doing natural language processing and it was all automated, but vets were able to get an appointment using text message within like seconds. So hugely from this form won't even work because it's PDF and it's not great to setting up an appointment by text. And so USDS worked in that model of let's get all the agencies or as many as we USDS worked in that model of let's get all the agencies or as many as we could get to into that kind of space. There was like a shift left to use that terminology that I've used in software testing and continuous integration and continuous delivery. We started as the firefighters and we shifted. The project sometimes started with people at the

US Digital Service. It's been a remarkable evolution. It's one that I'm always proud to watch and proud to get to be a part of because for the same reasons that when I see an organization where I'm working move into much more quality, more scale, those kinds of things that we would expect in the private sector, it's kind of a similar evolution. But I love that about it. I'm proud to be there or proud to have been there. I love that. And just to be clear, you know, you were talking about the types of projects that the U.S. Digital Service was taking on. To what degree was the U.S. Digital Service starting these projects on its own or being brought into different parts of government, different agencies to help with something that they were having trouble with? During our time, less than half we started. It

was mostly a request, but the work would start with something called a discovery sprint, which is when you send in a group of people to learn what the nature of the problem is, make recommendations for how to approach it to really just get a sense. Because if you look at going back to the 14 agencies in disaster recovery, like the money that goes out, like what does this even look like? And how does a person even approach it? So they start there. And that's often how USDS would start a project. Sometimes we would get called in on firefighting. So healthcare.gov is the easy example, VA shortly thereafter. But even FAFSA, last year we were called in to be part of that.

To the credit of FSA, the organization that runs FAFSA, as well as the people from U.S. Digital Service, they landed it pretty well this year. So there's a mix right now, or was a mix before the inauguration. And can you just say what FAFSA was for people who wouldn't be familiar? The Free Application for Student Aid. If you went to college ever, as old as I am, I did it many years ago. So most of us have filled it out if you've even done a semester of college. Did you want to add something, Milo? Yeah, there are definitely some projects that USDS kind of helped spearhead as well. Especially recently, there's been some

cross-agency collaboration on trying to help people get the benefits that they deserve from a bunch of different agencies. And so USDS has been really involved in centers for Medicare and Medicaid services and also at ACF, which is the Agency for Children and Family, and some of those other agencies. And so USDS has, in coordination with digital services at those places, tried to say, like, we could build, for example, a universal benefits application to make it easier for people to apply once, get all of the services that they need. So there are instances where we have taken lessons learned maybe from agencies or past work and then tried to say there's a path to actually make things more usable, more comprehensive and better for the American people. Yeah, that's great. If I'm thinking about the U.S. digital service and for people who have been paying attention to what has been happening recently, I think they'll be finding out a lot about the structure of the American government. And I would say I'm a

Canadian, so I wasn't the biggest expert on all the different departments and agencies and stuff before now. But, you know, obviously we've been hearing about like 18F and the technology transformation services within the General Services Administration. How would U.S. Digital Service differ from these other groups that are also doing technology related things within the government? Early on, one of the key differences was that USDS was funded through the White House. And so like the initial idea was that we were kind of still in that firefighting mode. But if something was really wrong and your agency didn't have the ability to address

it or fix it, you could ask for USDS to come in and we would do that firefighting role. Just like you don't pay the firefighters to put out the fire at your house. Same idea. That has changed a little bit over time and USDS adopted other funding models. But 18F, which to be clear, staffed with also amazing and fantastic and wonderful people that I care deeply about and I know a bunch, I'm heartbroken by what happened to them. They usually would come in earlier in the process and the agency would pay them. So the agency would say, oh, we need to roll out this piece of functionality or we have money

from Congress to work on this application, and we're going to get started from the ground floor. 18F would be the agency that could help with that, and that other agency would pay 18F for the staff that's assigned. When I was working on the direct file project, the people who came in and really spearheaded the more non-functional side of things. Incident management, deployments, authorization to operate, which is a very government term for, have we checked all the boxes for security to be deployed? It is onerous to get through. People from AT&F led a lot of that part of it. And I was also on the non-functional side, so I got a chance to work with a number of

them in that respect. Anybody who can run an ATO process could probably do any detail oriented task ever. It is very difficult. So I have a lot of respect for what they do. And that's one at the time that was just was middle of 2023. That was kind of a differentiator, but it's not a core differentiator. The payment model was more the core differentiator between us. I wanted to make sure I didn't represent things poorly, the differences between us, because We the Builders has a designer from 18F right now, and we have deep connections across We the Builders with the team 18F group. I appreciate you laying all that out, right? Because I feel like there will be a

lot of people who listen to this show who are hearing about these agencies for the first time and might have been curious what the difference actually was here and what these different groups are doing within government. If we look at USDS now, obviously this Department of Government Efficiency was initially positioned as this kind of outside of government thing that, I don't know, Elon Musk was going to have some role in and was going to be advising the president. But then on day one, when sworn in, inaugurated, he signed this executive order, basically renaming and taking over the U.S. Digital Service and making it the Department of Government Efficiency. So making it an actual government agency. What was your

reaction to hearing Doge was going to take over the USDS? Horror, sadness, concern, deep anxiety for my colleagues, all of that. I was surprised but not surprised when we first started hearing about this poorly named agency. I was like, oh, but there's already a group that does this. And clearly I wasn't the only one who had this thought. So I was also dismayed. I was going to say, I want to be forthright, though, that first off, USDSers are most often also outsiders coming from Silicon Valley. And so there is mirroring with the conception of Doge as outsiders coming in to make government more efficient. Even as we all shared concern and anxiety, there was this small kernel of maybe there is room here to work with the folks who will

be coming in from Doge. So it was not total panic. It was a sense of, in some instances, there could be a synergy, there could be an alignment, depending on how they're going to think about the work and approach the work. And I think it's important to note that it wasn't just viewed as a hostile takeover. It was viewed as if they are serious about the work, if they are actually truly interested in making government more efficient and work better, then there is a chance that we could work with that and work within that and actually be more effective. There are lots of people who joined USDS during the first Trump administration and were there because they wanted to serve during the COVID pandemic. They saw this not as a, like, I'm going into work for an administration, but as I'm going to go work

for the American people. And I think that the USDSers who were there when US Doge service was created, were hoping that that same mentality was going to carry the day. That makes a lot of sense because especially, I guess, going in, you're hoping that these people have good intentions. You're hoping that you can work with them, that they want to do the right thing, that hopefully they're going to kind of bring the ethos that I guess the people who went to US digital service in the first place had. You know, I guess then my question is how, you know, I know you're both former U.S. digital service engineers. How have you heard that things have progressed since then? And do you have any idea of what it's like inside the U.S. Doge service today?

Kind of messy in there right now for the few people that are left. Milo, I think you know more about the immediate aftermath of how they were interviewed. The reality is that there's almost nobody left from the crew that were there when the inauguration happened. There was a set of layoffs. There was people who resigned. There's frankly, I think, you know, probably less than 25 or so people, maybe a little bit more, but I don't think it's much more than that. Out of how many that would have been there before? It was 150 at the inauguration. And then 50 were on the February 14th. 50 were let go. At least that I know, 30 people have resigned. And that's who I know personally, which means there's more people who I'm not in contact with. Yeah. In the immediate time right after the executive order,

the folks at USDS were informed. They were let know that there was going to be interviews with folks coming in from Doge. And this has been reported a little bit, but the interviews were fairly combative. They were done in a way that was very uncomfortable for the USDSers. And the reality is that a lot of the people who are at USDS, you know, very careful about how they talk about the work that they do, very much kind of model civil servants in a way. They don't want to draw a spotlight to themselves. It's not about like, oh, look at this, you know, cool thing that I did. You know, we might've talked about tech or problems that we were solving amongst ourselves that way, because it's kind of inside the house,

so to speak. Not used to being interrogated, not used to being forced to explain ourselves. And a lot of these interviews were done over teams. And the folks who were doing the interviewing refused to identify themselves. They wouldn't explain what the purpose of the interviews were. They were supposed to be like 10, 15 minutes long. And the questions would be like, what is the most important thing that you've done this week? Like that kind of conversation, it felt just very confrontational and gross. Even for the interviews that were in person, the folks who were doing it were wearing like visitor badges. So it wasn't even clear that they were government employees, just a very uncomfortable situation and not necessarily the way you would even talk to colleagues in the private sector, right? Like just felt bad.

It felt uncomfortable. One thing that happened a couple of days later, people at GSA inside of TTS, who we've heard a lot about, 18F is there, login.gov, cloud.gov, lots of technology, like infrastructure, like baseline stuff is over there. I have a friend who knew about this one question that they always ask, which is who around you is exceptional? And it's meant for me to say,

Milo is a fantastic engineer. He always delivers, which helped them to figure out who to keep. And so that word had gotten out on that. So the people over in TGS, a group of them stopped answering that question. They decided they didn't want to be party to it because they knew about this and were able to make sure that at least a few people weren't part of contributing to who got cut. And I'm grateful for them finding these little ways on short notice. I can only imagine it was a really difficult time for a lot of those

workers facing these interviews, facing this kind of this hostility, right? And this clear disrespect, it sounds like for the type of work that they were doing. And as you were saying, like the service that these people were trying to do for the American public, right, to improve the way that they access government services. And all of a sudden, you have these people coming in from outside who seem to have a very different conception of how these things should be working and what it should be looking like. Also, there's a piece how it should be working. What is the definition of efficiency? But also, who are these people in front of me? It's like taking all however many hundreds, thousands of people and saying, well, I'll just sit there and do nothing. We don't need you.

We're cool. We're going to add AI to this instead. Everybody that I worked with, and I can say that almost to a person, worked more than 40 hours a week. I knew I could find them at 9pm if I emailed them or sent them a message, depending on who they were. These people cared. They still care. The people who are still there, I have so much respect for. So when they came in and started cutting people, it just showed an assumption that these people don't do anything. It's just the flat out lie. These things are difficult. They're difficult for a reason, because meeting the needs of all Americans and people who live in America is difficult. It's difficult to administer programs that have been cobbled together in policy. It's difficult to envision a form that can meet 80%

of users so that the customer service representatives can serve the people who truly need it. It's difficult to do this regularly. And it's not to say there's not efficiencies to be had, but it does mean these people are working their butts off and the disrespect that's implicit in how they treated those people, my colleagues, my fellow Americans, just pisses me off. The truth is that in a lot of cases, you join USDS and you think, I'm going to go in and I'm going to solve this problem, or I'm going to work on this thing because government's not done this before. Government's not fixed this before. And you get into whatever agency you're detailed to, and you start to talk to the people there.

And they walk you over to their desk, and they can pull out the four times prior that someone has shown up and said, we're going to solve this problem. And you look at it and you go like, so people have been trying to fix these problems. People have been trying to address these inefficiencies or these challenges in government. It's not for lack of trying. It's just sometimes it's lack of coordination. It's lack of money. It's not the right time. It's all kinds of different challenges come up. One of the really core talents at USDS, again, resiliency mattered and this kind of ability to think around problems. Like maybe it's a technical problem, but maybe it's actually a problem of policy. Maybe it's actually a problem of how government works in the United States. A great example that I can give you is we see a lot

about Social Security Administration. Doge has talked about how there's all of these dead people who are receiving benefits. And when you start to dig into it, you look and find out that there's all of these different databases at SSA. And there's a lot of challenges with getting death information in the United States. reason it's difficult to know whether or not someone has actually died is because in most states, death records are actually created by the county coroner or the county medical examiner. Someone dies, the county coroner fills out the paperwork, and then maybe it's actual physical paperwork that gets mailed to the state. Sometimes it's an electronic record. Sometimes the county coroner literally saves a PDF to a zip drive or a USB drive and mails that into the state. And then someone has to manually transcribe that. And then they have to manually send that into SSA and into IRS and

into all of the other places. That's not the fault of the Social Security Administration. That's just the cost of getting the information. That's just the way that it's been set up. Are there ways to fix that with technology? Absolutely. Would people at USDS love to fix that? Absolutely. But maybe the money doesn't exist because we haven't had a freaking budget passed by Congress in, it feels like, 100 years. 1996. 1996. So there's no way for an agency to actually do planning. There's no way for them to actually think about improvements. Maybe there's not a single agency that controls this that can actually mandate it. These are the kinds of things that could be solved if there is

political will and an interest in doing it. But instead of like looking at that as a real problem that could be worked on and made more efficient and actually like save money for everyone involved, now they're just walking through with sledgehammers and chainsaws and saying like, well, you haven't fixed it up until now, so we're just going to burn it to the ground. Great. That's not going to fix anything. It's not any better now. No, because my grandmother is going to have trouble getting benefits soon. That's the real word, real toll of this definition of efficiency. You got a screening now, by the way. But let me tell you, if I could solve that problem, God, how much fun that would be. It would be difficult. It would be frustrating. But U.S. Digital Service did do some stuff at the CDC around facilitating public health data

going to the CDC during the COVID crisis. So we know it's possible to improve this local teeny tiny jurisdiction death records to the Social Security Administration. We know this is possible because CDC did it and U.S. Digital Service supported that. This is really fascinating to hear. And I wonder what that tells you about the kind of mindset that Doge, that Elon Musk, that this approach is bringing to the U.S. government, to the U.S. Digital Service, because it sounds to me like there's really not a desire to understand the root of the problem, right? And just to assume that you know what is happening here and that there's a particular approach that is going to fix it when you both have had experience actually trying to deal with these problems firsthand. So what does your experience tell you about the mindset that they're bringing, the issues with it, and what it would actually

take to address some of these more difficult challenges if you were serious about wanting to do that? Tells me that the definition of efficiency is different. That, for instance, my definition of efficiency is making it easier for Americans to access the services that they are entitled to, the taxes they've paid. I've been paying Social Security since I was 16 years old, and they're making it hard for people who get Social Security now. It's like I described it like a Jenga thing, a Jenga tower with those little blocks that you pull out and hope it stays standing. That's what they're doing.

And what's going to fall over? We don't know yet. Could it be that my grandmother won't, literally my grandmother is 91. Will she be able to get her social security check? I hope so. I hope all of our elders are able to still do that and us when we get there. I don't see the definition of efficiency the same as what I'm observing now. I think a lot of it comes down to this whole zero-based accounting idea that we've read about

that Musk practiced at Tesla and everywhere else that he's been. And this idea of cut everything and then see what breaks and then start restoring things. I'm going to say that's fine for a business, though I don't necessarily know that I agree, but it's definitely not fine for government services for a bunch of different reasons, not the least of which is some of these services are essential, and without them, people are going to get hurt or they're going to die. If you cut everything at Medicare and Medicaid, then someone might not be able to get the medical services that they need. They can't go to the doctor. They can't go to the hospital. That could mean that they face life or it, it might be too late for that person. And so I'm not okay with that mentality being inside government. Again, it is not about saying that we have to

accept the status quo and that we can't make things more efficient. We can and should. If someone were to say to me that they could give me a budget to fix some of these really gnarly problems in government, and I had runway to do it. I would. We've seen, for example, John Oliver has done some fantastic reporting about the Oregon Procurement Network at HRSA. That has been an absolute disaster for 20-plus years now. There has been an interest and bipartisan support to fix that. And there was a team that was actually sent to HRSA. They stood up their own digital service team. They were actually getting traction on fixing this problem and actually going to be able to get people the

organs that they needed. And then Doge cut the entire team because it had just been stood up. And so they were all probationary. And one of the things that I think people need to understand about that probationary, it sounds like they were all like 18-year-old kids who had just started. But in government, when you move from agency to agency, or even you start a new role, you can be probationary. And so that means that all of these people who had come in from other places as technologists, all of them with like 15 or 20 years worth of experience who wanted to solve a really life or death problem and were working

more than 40 hours a week to do it, got cut. It's a disgrace. And that's not efficient. What's going to happen now is the vendors that work on the OPTN, the organ procurement network, they're going to continue to do the less efficient, less capable, less correct way to do things. And the people are going to die for it. It's so infuriating. It really shows you the scale of the consequences to these decisions as well, right? It feels like as Elon Musk and the people who he's bringing in are charging through government that they're not so much thinking about the person who is losing access to social security, who is losing access to

their Medicare benefits, you know, the programs like you're talking about that are trying to fix these issues but are being sidelined because people like Elon Musk don't actually want to understand the details of what is going on here to actually really try to address it, but thinks he knows everything kind of coming in. I wanted to ask a bigger question based on that. But before I do, Kate, you know, when we were emailing before this, you were talking about the kind of engineering first mindset that Doge was bringing into the government to USDS. Can you talk a bit about that and how that differs from how these problems are usually approached? I'm going to start with the analogy of the private sector companies I've worked at. My specialties are front-end development software testing.

So I've worked, someone says, you're a full stack. I'm like, what's the stack? I do all the things. So usually when I'm working in those organizations, I'm working inside of a cross-functional team, but it's mostly engineers, maybe one designer. And the problems are usually something in the making money category. How do I create a privacy compliance system that allows X company to comply with GDPR, the European privacy law? How do I create a travel management system?

It's all about money. And usually those solutions, there's a little bit of design. I've had some great ones there. The travel company did a really good job with design, but a lot of them just skipped it. I remember arguing at one place, I was using a helping to build a command line interface, a CLI. I was like, we need a UX person on this. We really do. And they're like, why? It's so hard to use. But that's a CLI. No, but somebody could write a nice man file. I was really frustrated by it. And that probably comes from my front end experience. I'm like, no, we've got to make this usable. So landing at US Digital Service, I was like, we've found people who understand that the problem may not be the problem that somebody thinks it is. And by that, I mean, do we really

need to make data sharing better for disaster survivors? Yes. But the question is, how else can we make it easier for them to experience that very complicated process? And the difference inside of government is we're not making money. We are taking Americans' taxes and turning them into services, which then serve people who need it. So the motivation structures are totally different. And these big problems with big consequences that we've been talking about, it's different when you approach them. It's mapping out the whole process, understanding where the policy helps and hurts. Sometimes the engineers don't write any code.

A lot of times we're mapping the systems with the designers. We're mapping the architecture side. We're working to coach people. It's just different. It looks slower, but they're being more holistic about the solutions that they're choosing. And what it results in is applications like DirectFile, which has competitors in the private sector. And DirectFile's NPS scores are off the charts. I can't tell you what the other ones are, but I know that that application is usable, because I tested it. I worked side by side with the people who wrote it. And they took the time and the care to understand that neither Milo nor I speak 1040. In Paris, I hope you don't have to speak 1040 as a Canadian. By the way, thanks for putting up with us right now. You have to ask certain questions that aren't on the form in order to figure out, for instance, filing status. Here we have single,

married, filing jointly, married, filing separately, and head of household. I might be missing one, but those four are the main ones. But if you're married, and you're married to an undocumented immigrant, you're married, but you're considered single for filing status. So in order to understand your filing status, we have to ask you a number of questions. And those questions need to be seventh grade, eighth grade reading level. Somebody had to stop and think that through. Instead of saying, what's your filing status, Milo? Are you married? Who is your spouse? And various other questions, which then turn into the single filing status answer, right? So you take that care, that thought, and that deep understanding of what the real problem is.

And that just, it looks slower. But in the end, that satisfaction rating is high, and nobody wants to do their taxes. But they're okay with doing it with direct file. I'll just say, I don't know what 1040 is. So I definitely don't need to understand it. Yeah. But, you know, thank you for figuring that out for all the people who rely on you to do that. There are some designers who are amazing. I would just say what you're describing there sounds so important, right? Obviously, if you're in a company, you're developing this product to try to make money, to try to get customers, whatever, right?

To try to get some venture funding, whatever it is. But if you're developing a service that needs to serve all these different needs and all these different types of people within a society, there's so many other things that you have to consider, right? It's not just like I'm creating a product for a certain group of users or potential customers, but my job, the orientation is so different because it's a public service, right? It's not something that you're selling to a particular person. So that makes a ton of sense. Milo, I was just hoping that you could pick up on that and talk a bit more about how it's different

in the private sector to developing technology for the public. And like Kate was saying, where you're trying to understand these systems, is it always about looking for a technology to solve a problem or can it sometimes take different forms? It absolutely takes different forms. Another example where technology isn't necessarily the answer, refugees, for example, the Afghan refugees who were oftentimes people who supported the war in Afghanistan, the people who were trying to come to the United States needed to fill out forms, but very difficult to sometimes get internet access in the places where those people are. Difficult for them to even have a computer in many cases. They might not even have their own phone. And so there were technical challenges with all of that, but there was also just people challenges. How do you design a system where multiple people can share one phone, but have their paperwork accessible and have the process

for filing and filling out this paperwork accessible, but secure at the same time? Because you don't want to expose information from one person to someone else who's also sharing that phone. That's not a technical problem. That's a people problem. Technology is a part of it. But we had folks at USDS who were working really, really hard to figure out how to make that possible. And we did. Another example is Kate mentioned CDC. One of the applications that got built during the pandemic was called Simple Report. The idea being that during the pandemic, we still needed to have facilities open. We wanted to have schools open, but those schools needed to be able to test all of the staff and the students.

All of the states during that time were requiring every positive and negative test result gets sent in. Every single day, a school might have 300 students and they might have like 50 to 75 staff members. So you're looking at 375 test results that needed to be collected and then transmitted to the state public health department. The technical infrastructure that normally runs that whole world of reporting is incredibly complex. It's very technically in the weeds. And there's no reason that it needs to be like that. And you're not going to teach a kindergarten teacher. I mean, you could teach the teacher how to encode the message in the right way and then use SFTP to send it into the public health department, but you don't need to do that. And so the USDS folks who were there at the CDC at the time built this application that took five

minutes to learn how to use. And you could upload a roster of every single person that was in the building, and you could click a button, start this test, run the nasal swab, and then it would actually do a little countdown timer, and then it would ring a bell. You could literally line up all of the students and have the website keep track of everyone. And then you could just click a button, positive, negative, positive, negative. And by a single button click, record the test result. And then one final click, send the entire body of the testing for that day into the public health department. And it was a game changer. The team that was there said that when they demonstrated it to the state of Illinois, they ended up deploying that to every single school within like a couple of weeks, because it was a major advancement in getting data from elementary schools and middle schools and high schools into the public health department.

Those kinds of projects, those kinds of wins were things that USDS was doing every single day. We were just doing it quietly. We were just doing it not in public spotlight and not doing it for making a profit, but the need was there. So we did it because we believe in service and we believe in servicing the people in the best way possible. I love hearing that so much and hearing the concrete examples of the types of projects that you guys and the team at USDS were working on really brings home how this work is so important and how you're trying to think about so many different use cases for how these things are going to be used for the different types of people who are going to interact

with it to try to make it useful. You know, one of the things that was coming to mind as I was listening to this, and totally up to you if you want to comment on this, but you know, I feel like sometimes when we're having these broader discussions about how we make technology better, how we make the internet better, there's a lot of people who know how to code who are involved in those conversations. And often the solutions can be very focused on the types of capabilities and knowledge that they have and doesn't often think about the grandmother who would not be able to easily involve that or people might even just

say, well, everyone should just learn to code or need to learn a certain degree of technical knowledge. And it's like hearing the way that you are thinking about addressing these problems, I feel like we could use a little bit more of that more widely, not just in government, with how a lot of people think about technologies and how services are developed, especially when we're thinking about what better alternatives to the types of big platforms that are out there now and stuff actually look like and can work. I mean, to every line of code any of us write is a technology burden. I mean, it's a maintenance burden. Every line of code we write is a maintenance burden. Somebody's going to have to pick that up. And it's probably going to be me,

and I'm going to forget what I was doing. So I'm better documented. We always tried to make sure that everyone who came into USDS recognized that and thought about that. We have core values at USDS, but the ones that was not written, but you heard a lot was the most good for the most people who need it the most. And we would say that a lot. You might not build the most sexiest piece of technology. You might end up having to go support an older application. You might have to go work with the COBOL team at Medicare, but is it doing the most good for the most people? Yeah, that is the more important piece.

I just have one more question for you on this section, and then just have a broader question as we start to wrap things up. And that is one thing that I feel like has come up a few times, is you're dealing with very sensitive data with very sensitive topics and issues that people are dealing with. And I feel like one of the stories that has really stood out to me as we've been watching this kind of Doge charge into the government and try to change so much is the degree of personal information and data that they're seeking from all these different sources. And, you know, Milo, you were mentioning security and these sorts of issues. And it feels like those sorts of things are really not top of mind for the people who are running Doge right now and who are trying to bring this different ethos into government. So I was just wondering kind of broadly what you make of this kind of disregard for privacy and security that we seem to be seeing at the moment.

I'm horrified by that as well and infuriated because the data is so important and it needs to be protected. There's so much data that government collects, and it covers so much of an average person's life. Part of the responsibility of every agency is to steward and guard that data as diligently as possible. You don't want that information to leak. You don't want that information to be misused. And every single agency, even outside of US, yes, every single agency has privacy officers. And as part of building any application in the government, you do a privacy impact assessment where you actually lay down, I'm going to be collecting the following information and I'm using it for the following purposes. You oftentimes need to justify to the privacy officers and to the chief information

security officer and other people at the agencies why you need that data and why it's important. And that's a conversation that you have. And that sounds slow and sometimes it is, and sometimes it can be really frustrating. And I don't want to make it sound like it's utopic or perfect. There were definitely times USDSers were going back and forth with chief information security officers or arguing about why this data was important or how the data was being used was important. But we were having those conversations and we were prepared to justify why we needed that data. And I feel like conversation's just not happening. And we've seen from agencies, CISOs, those information security officers, have left or have been let go. And that, frankly, scares the hell out of me.

Doesn't the privacy approach track with Silicon Valley right now, where, you know, my data is out there and I have to pay a service to make it go away so that people can't find me as easily, which still isn't hard. So it totally tracks the disregard for data with any of the large social media platforms. But hey, they think I'm young and they think I'm. And this is amazing. So I'm doing it right. But with the IRS, with SSA, it's an explicit agreement that I am to give my data, and they are going to safeguard it. If you have ever had to deal with an American system, there's usually a Privacy Act statement every time that you log in. And it's like, oh my God, do I have to look at this again? Every time I log into my retirement account? Yes. Yes, I do. And that's the care that we expect. So the approach is so Silicon Valley, but in government, we don't have the option. When I file my taxes, they know everything about me. And now Doge might too. And where are they taking it? And what are they doing

that makes me so mad? Part of it is like, I know that Meta has my data and I know that that's the cost for being on Facebook and being on Instagram. But that is a choice. I could delete my account. I could not use those services. I don't have a choice about paying my taxes or the other services that government offers. It is part of being a member of society in the United States. And because it is compulsory, because it's going to happen no matter what, where I face consequences, the least government can do with that data that they are requiring me to share with them is to guard it as very important. We all took that extremely carefully and took that very seriously. And I just don't see that same care or worry with Doge. Yeah, I feel like care is a big word there, right? That care just in general does not seem to be there. And that kind of compassion and concern for even the workers, but especially like fellow citizens and fellow people in society, right? There is so much more that I could ask you about. But as we close off our conversation, I wanted to put a spotlight on this initiative that you're

both part of called We the Builders. And so I just wanted to ask you, you know, what is that? What are you trying to achieve with it? And finally, is there anything else that you would just want the listeners to know that maybe we didn't get to through the course of this conversation? So We the Builders is an effort to get information out about what's really happening in federal government, who these workers are, and what are the consequences of the actions that we're observing right now. There's two ideas that went into this. The first one is the idea of demystifying what's actually happening. And then there's the other personifying who these people are. We kind of balance between the two, although I have one coming out. We did an accessibility assessment on doge.gov. And oh, dear, they need our help. Coming soon to

wethebuilders.org. But most of the time, we're really trying to support federal workers to personify what's actually happening and to inform people from those of us who know more and are concerned for everybody. That's what we're up to. We're a pretty large group for a bunch of volunteers, about 15 people. The part that I find most gratifying is actually working with people. Now that I've left U.S. Digital Service, I get to work with some of the same people. I get to work with some new people and they're all fantastic. And it's nice to watch them be excited to help give an opportunity for people to contribute to a time that's very scary, that's uncertain. And so like, that's where I get most of my joy from is watching the people around me thrive, which is me period. But right public, but also help media inform the public, because frankly, the government does a terrible job of explaining what it does and how it does it. Again, one of those frustrations that you would experience at USDS is you would see this story that your colleagues are working on something super amazing and the agency that they were partnered with never wanted to expose it or bring it to the light of day. Federal employees generally are allergic to attention and are allergic to being in the spotlight.

They really do just want to serve the people. But these are really unusual times. Someone needed to kind of get in front of the press and say, by the way, this is what it actually means to be probationary. By the way, this is a project that people were working on that was really amazing that you never got to hear about before. This is what an ATO is. This is this other aspect of government. These are all things that even people in the media had no idea and didn't know how to report on. And that's not their fault, but someone needed to kind of leap into the void that was being left. And so we

wanted We the Builders to be a resource for people to learn all of those things. Hopefully, the damage that is being done is not permanent. And so we the builders, it's not like a final record, but it's more like a we'll be back. I love that. And obviously, I thank you both for doing that. And of course, the other people who are working with you on this project, I knew going into this conversation that I was going to really enjoy it. I love hearing about the kind of work that happens within the public sector to try to make government better through a technology lens. But I feel like just hearing the way that you have both been talking about the work that you were doing, the particular orientation toward these problems within the public sector, I think it's so

illuminating. And I think there's so much to learn from there. And that perspective in trying to address these issues and thinking about technology through this much wider lens is something that we would all benefit from if it was applied far beyond government as well. And we would all be better off if that was the case. So really sorry to see everything that is happening with the USDS and, you know, I'm sure some of your colleagues and former colleagues, but I thank you so much for taking the time to come on the show to inform the listeners about what is going on to give them this broader perspective. I really appreciate it. Paris, thank you for taking the time to get to know us and ask these good questions. I'm really grateful for the spotlight.

I appreciate you letting us come on today. Absolutely. Thanks so much. Kate Green and Milo are former US digital service engineers currently working on the We the Builders project. Tech Won't Save Us is made in partnership with The Nation magazine and is hosted by me, Paris Marks. Production is by Eric Wickham. Tech Won't Save Us relies on the support of listeners like you to keep providing critical perspectives on the tech industry. You can join hundreds of other supporters by going to patreon.com slash tech won't save us and making a pledge of your own. Thanks for listening and make sure to come back next week.