Scott Kupor's New Plan to Bring Tech Workers Into the Federal Government

22:02Yeah, it totally does. It is kind of wild to think that last year or a few years ago or just four months ago, Joe and I could literally say that we're like coding experts. I'm a 10x engineer. Yeah, that's right. And apply. I know. I'm a COBOL expert. COBOL, COBOL, I can never remember. COBOL expert would be very much demands. Cobalt, cobalt. No T. Cobalt. Thank you. Cobalt is the mineral. There's a cobalt mineral and there's a cobalt. Cobble, yeah. I never know if it's cobble. I want to say cobble because it's like cobbled together across IT systems. Oh, that's actually really good. Okay. I didn't think about that.

22:40Okay. Actually, but on the second part of Tracy's question. Okay. So once they're in the office, and again, because it's not a for-profit entity, like how does the government currently go about evaluating how someone did this year and what would be the optimal way for the government to evaluate how productive someone is in their state? And this is a problem in private sector too. I'm pretty sure that if you went into Facebook or Google, et cetera, they would also not have a clean answer of like which software engineer is more productive or contributed more value than the other one. Yeah. No, you're absolutely right.

23:15Look, this is not a problem that's unique to the government, but what I think is unique to the government is Historically, basically, everybody gets an A in government. To give you a perspective, most agencies rank people one through five at the end of the year on their performances, one being the worst and five being the best. And if you look historically over long periods of time in government, about 80 plus percent of people get ranked a four or a five. So meaning basically everybody is in kind of like Wovegon and well, well above average. And then about 0.2 percent of federal workers get ranked a one or a two, meaning that you're performing below expectations.

23:48So number one, like we've had no system that actually recognizes like outstanding performance and quite frankly, unfortunately, no system that actually holds accountability on the other end of the spectrum. We're changing that. We've already put out regulations for the most senior executives, what are called the SES, which are the most senior executives where we effectively for the SES have put out a regulation that requires a forced distribution. So a cap of 30 % of fours and five rankings, which we think is appropriate to actually distinguish people, you know, in terms of their performance. So number one is we've got to reform the system, period, and we're working on that, and you'll see more regulation from us on that.

24:23Number two, you're actually right, which is, look, like sometimes the metrics are hard, right? So, you know, you don't really want to measure like lines of code that somebody produces because obviously all code is not created equal. But, you know, what I think does work in the private sector is, look, managers are responsible for their frontline workers. They meet with them one-on-one. They have, you know, weekly. They have a sense of what they're doing. They provide feedback along the way. Like, I think the performance management ongoing process works better in the private sector. In the public sector, what I found is instead of managers, we use the term supervisor, which I think actually, unfortunately, does a disservice to what the job is.

24:56Like, I don't think the job is necessarily supervision. The job is to manage, meaning are they working on the right stuff? Are they doing the right things? Do they need training for this kind of stuff? And historically, we haven't done a great job on that. So one of the things that we just did, for example, is we just rolled out a brand new training program that's mandatory for all supervisors specific to this issue on performance management. So to teach them, like, what does a one on one look like? You know, those of us who grew up in the technology industry, the first book that we were all assigned to read was Andy Grove's Only the Paranoids Survive, which I still think is a fantastic book.

25:26I don't know if you all have read it. It's actually a really awesome book. I have read it. Years ago. I have read it. Yeah. Keep going. Yeah, I mean, basically, he literally talks about, like, what is a one-on-one meeting with your, you know, team member? Like, what should it look like? And it's like, you know, it's very much like management 101 stuff. But I think the honest answer is, like, most people don't get that training. And so as a result, we have had a performance management culture where we have inflated grades and we have basically once or maybe maximum twice a year people actually get feedback.

25:52So it's totally unfair to the employee. And it's also just unfair to top performers who are not actually getting differentiated reviews and differentiated compensation. I was just about to ask, just to be 100 % clear, compensation is going to be tied to performance. Correct. Okay. Correct. Yes. All right. Yes. And again, it's not that it hasn't been before, but the problem is if everybody's a four and a five and you have a finite bonus pool, then you're peanut buttering out the bonuses, basically. So like, yeah, just, you know, Tracy, just to give you a sense, like when we gave guidance on this SES, these very senior executives, where we said there's a 30 % cap, we also said we think 60 % of the bonus-based compensation should go to that 30 % of the individuals.

26:30And, you know, that's a, you know, we can argue whether it's 60 % or 80%, but the point is like let's actually really pay for performance and let's differentiate compensation as opposed to giving everybody a$300 or$500 bonus, which in the scheme of things has little retentive or incentive value for employees. So you mentioned project management just then, and I was looking at some of the press releases and documentation that you put out as part of this program, and there really is an emphasis on getting project managers. And I am very curious why that particular role seems to be in short supply, because I know someone who moved from a completely different industry into project management for a graphic design company.

27:11They had never done it before. And it didn't seem that hard. Okay, so I'm going to distinguish between project management and product management, which is, I think we, I hope we wrote product management in our stuff. Oh, did I misread it? Oops. Well, I don't know that you did or not. We may have put it in there. But what we need is what I would call a product manager job, which is, you know, a product manager in my mind is basically somebody who, sometimes they come from an engineering background. They might actually have been developers themselves. but they're kind of people who sit between the engineering organization and the end user customer, and they listen to the customer and say, great, okay, tell me your problems.

27:49What are you trying to solve? And their job is kind of to translate that into ultimately in tech speak what we call a PRD, a project requirements document for the engineers to develop against. So, like, that's the role we need. So in a well-functioning engineering organization, you have developers, you have designers, and then you have product managers, and the product managers probably aren't putting their hands on the keyboard actually, you know, So, you know, typing, you know, ones and zeros, but they are literally that go between, between like the end user customer and then the requirements that ultimately drive the development process.

28:19It's interesting. I feel like that's something I've really come to appreciate in working for a large organization, how valuable translation skills are, because you have this one part of the room that knows one very specific thing. And then there's these generalists on the other end that have some vague thing. And that person who could sit in the middle and talk to both sides is clearly a very is a valuable person. You know, so you mentioned like the one thing that comes up a lot in a common theme on our podcast is like the importance of this sort of like tacit organizational knowledge. Things that are like cannot simply be written down.

28:56They do not exist necessarily in one person's head. There are just things that are sort of collectively known process knowledge and so forth. How do you think about that in light of the fact that these are two-year roles? And, you know, again, maybe in some cases someone will do two years and then continue their career in the federal government. But, you know, especially given the pay gap that emerges specifically for technologists, like that's probably you're taking on the one area where the pay gap probably gets the widest over time. How do you think about fostering and maintaining that kind of knowledge, especially given what you laid out at the very beginning, which is this sort of, I don't know if crisis is the right word, but these looming retirements of all these people that do have years of accumulated knowledge about how things really work?

29:48Yeah, so you're absolutely right, which is, look, we have to balance kind of the two-year time period against the need for actually the organization being able to maintain these systems once people leave. So let me just give – I'll give you a real-world example, Joe, just so you have it because it might be helpful. So in my organization, OPM, we own the retirement process and retirement for several million annuitants basically who have kind of completed their professional careers. And we are – it's a very – I won't go into the details. I won't bore you. But there's literally – it's like a ridiculously manual paper-based process where for years we have had people literally filling out paper applications as recently as beginning of this year and sending paper by postal mail and all kinds of stuff.

30:27And we literally have a mine in Boyers, Pennsylvania, which is very fun if you ever want to come see it, where we have hundreds of millions of file cabinets of people's retirement records. We'd love to. I would love to. I've always wanted to go like those those like huge like under mountain, like big. Yes, it is. That's what it is. They literally bury a lot. It is Iron Mountain. I'd love to check out. The Indiana Jones Warehouse for Government Bureaucracy. I've always wanted to check out one of these places. We'll arrange that after the episode. All right. We're going to work on that one together.

30:56Anyways, I bring that up to say, okay, so look, luckily we've had a small team of developers, Joe Jebbia, who you guys may recognize the name. He's a co-founder for Airbnb. He joined the administration even actually before I got there. But he's been kind of housed in OPM. And he has like a team of like three developers and, you know, a design person. And they've been completely redesigning this retirement application. So we now have what we call ORA, the online retirement application, where instead of paper, people actually can go, you know, in the year 2025 onto a, you know, computer and, you know, type in their information, upload their documents, it gets ratted to the HR, and it's, you know, it's, you know, vastly better than what we've had.

31:34So I bring this up to say, look, I think this is me as a good example, which is all those people are going to disappear at some point in time, and then the responsibility for that application is going to shift to the existing CIO organization within OPM to maintain it, to integrate with other systems. And so a requirement of being able to do that development process for Joe and his team is, look, they have to do documentation. They have to do training of the team. Look, I think it's not an insurmountable problem. I just think it means as part of your process, you just have to anticipate that there is going to be a time where people are going to disappear.

32:03And quite frankly, it's a smart thing to do anyways, because we just know, and look, this happens in private companies all the time. Look, people's average tenure in private companies, I can tell you in Silicon Valley, it's way less than four years is what the average tenure of an engineer is. So these are not new things. They are new things to the government, which is this idea of kind of knowledge capture and knowledge management, you know, kind of in light of the fact that people might not be there for their entire career. But I think it just means as we've designed this program that we have to have the appropriate level of documentation, of integration, of source code notes and all those things that are, you know, you would do in a normal circumstance to recognize that.

32:33Now, look, Joe, in the perfect world, I hope people love it. And they're like, wow, this is exactly where I want to spend the rest of my life. And that's awesome. If they want to do that, like we want to keep them in government. But, you know, I think we'll also have to deal with the practical realities that the current generation of people are not thinking about three, five, ten year employments basically. Things are much more clear than that.

33:09What are the day's top stocks? The entire auto sector is higher today. Which companies are posting the biggest gains and losses and why? Tesla's been a stock that's been in focus. Shares have really been all over the map this morning. Track the latest market trends on the Stock Movers Report from Bloomberg. Second biggest driver in terms of points for the S &P 500. These are short audio reports delivered directly to your podcast feed. European banks have been unfairly discounted. Search for Stock Movers on Apple Podcasts, Spotify, the Bloomberg Podcast YouTube page, or anywhere else you listen.

33:42Let's take a look at the sectors and how they perform. Tariffs putting pressure on yet another retail name. We're also watching a second day of drop in Chinese stocks. Get the latest stock news and data. Backed by reporting from Bloomberg's 3 ,000 journalists and analysts across the globe. Let's talk about some other decliners. You are seeing a little bit of a buy the dip mentality. That's a stock that needs a stock split in a big way. Subscribe to Stock Movers wherever you get your podcasts. What are you doing to streamline the operational side of the process? Because I imagine you mentioned culture before.

34:17If the pitch is come work for government, it's important, but you're going to be lower paid and you're going to have to constantly deal with paperwork and put together, you know, RFPs that have 7000 requirements and things like that. That's a tough sell to someone who's used to the whole move fast and break things culture. Yeah. So anyways, let me just be totally clear, which is we are not going to move fast and break things because we can't quite do that in government. We can move faster than we currently do and we can afford to take some risks. But we generally don't want to break things because, you know, people do need to get their Social Security checks and, you know, the IRS needs to actually make sure you get your tax refund.

34:59So, but it's a great, it's a great question. So look, there's a couple things that we're doing both inside and outside this program. So outside the program, more generally, one of the big things that OPM is driving along with OMB, you know, kind of the Office of Management Budget is we're working with all the agencies to basically what I would call do headcount planning and reshape the priorities for the organization. So there's a huge amount of government is very good about adding things to the plate, and they're very bad about actually ever taking anything off. So every time something new comes down from Congress, we add, you know, thousands of more people, we add stuff to the plate.

35:30And then we have this list now of 20 things, probably five of which we've been doing for 20 years. And quite frankly, nobody even remembers why we're doing them anymore. They've just now become like, you know, embedded into our culture. So part of the exercise in order to kind of make the throughput faster and to make sure that an engineer isn't like sitting on their hands dealing with, you know, crazy bureaucracy is we have to go back and look at prioritization and actually remove services that are no longer statutory and they're no longer valuable to the American people. So that's the exercise we've started.

35:57That's the exercise that we're kind of is ongoing right now. But the purpose of that will be not only to make sure that we're hopefully spending our money in the right places, but will also be to remove layers that ultimately slow down the actual work product of government. The second thing we can do is, as I mentioned, we're really trying to kind of change the performance management culture in government and eliminate barriers to people being able to move through the organization and change the idea that you can't take any risk. I mean, obviously, that's a big cultural challenge that we have in government.

36:24So everything from down to like what's your performance plan look like, your performance plan should have, you know, efficiency. It should have, you know, some measured amount of risk taking in there. So things that, again, will speed the pace of deployment and make sense. The third thing we're doing is, and this is, again, you guys asked me, you know, how is this program different from other programs? As we hire these 1 ,000 engineers, ultimately they will live in various agencies, right? So Dr. Oz at CMS says, you know, it's like, look, I need 75 engineers to do, you know, kind of important projects there.

36:53Or obviously Department of War, you know, has lots of stuff. So they're going to live in those agencies. But what we're not going to do, which is historically where I think these programs have failed, is if you drop these teams just into the blob that is the federal government, they're going to have a terrible experience and they're going to get lost and their two years will literally go by without them actually having accomplished anything. So our whole goal with this is keep these teams together. So what I'm hoping from Dr. Oz is there's going to be a team of 75 engineers who are going to be a team and they're going to work on the projects that he thinks are the most important projects.

37:25And of course, they have to integrate with the existing CIO organization because so much of what we do has integration backends and all that kind of stuff. But we've got to make sure that that happens because otherwise, you know, you hire a random, you know, intern and you plop them into an organization and they get swallowed by the beast basically. It's just – it's understandable why that happens. So we've got to think about the organizational implications of bringing people in and make sure they're set up for success. And, you know, my expectations are and certainly the conversations that I've had with all the agency heads are that they understand that and they understand that for them to achieve what they want to do, they have to create the organizational structure that enables these guys to actually do what they're there for.

38:01Yeah, I think this is another thing that we can probably say from firsthand experience is you come to a big organization, you've spent a long time, maybe the first year of just getting to know the lay of the land. Who are the people that you should talk to, et cetera? It's a real challenge. It's a real challenge being effective of actually accomplishing anything the amount of time. I mean, I've been to Bloomberg. We've all experienced it. Yeah, I've been here 10 years and there's like, who do I call that? It's just like the nature of large organizations. I want to talk about, you know, people who you say, OK, like you need we need to do something about the fact that all these people are going to retire and there is this pipeline, etc.

38:39Setting aside optimal levels of how big the federal workforce should be, etc. One thing that I could see as a bit of a problem about maintaining a career is if federal workers do not feel that they are respected in their jobs. And there is rhetoric that is that you hear. It's like federal governments, they, you know, again, some people are like, oh, they're leeches. They're part of the deep state. They're all woke and they spend money on this and that's waste. But even that spending looks like what I observe as spending in the private sector, etc. And then, you know, there's at least from this administration, there's been these sort of seemingly from the outside and probably from the inside capricious layoffs or arbitrary layoffs.

39:25And so if you respect this workforce, then how could you possibly have just like cut so arbitrarily? I mean, these are this is what it seems like. When you think about building a talent pipeline, do you think about this dimension and do you agree with the sort of premise of the question that it's a challenge that no one is going to work at a place if their bosses to some extent sort of think they're – do not respect them? Yeah. So the short answer is yes. Like absolutely this is a critical part of making sure that the federal government is an attractive place for the people who are there and also the people who we're trying to recruit in.

40:04So I got a million thoughts. Let me try to see if I can keep them in the time we have remaining. So I'll tell you one thing. So just to back up, as you know, I worked at Injuries and Horowitz, and Ben Horowitz, in my mind, is one of the best CEOs, best managers that I've ever learned from. And I had worked for him for literally 10 years prior. So for 26 years of my life, I've had the benefit of learning from Ben. The most important thing I learned from him is exactly what you described, which is when you have restructurings in companies, basically you have broken trust with your employees, okay?

40:31So particularly in the startup context, right? When we were in startups, we personally recruited every single person into that company, and we told them how we were going to conquer the world and how awesome we were going to be. And we believed it. We weren't lying to them. But it turns out we were wrong. The market changed. We didn't have enough money. And we had to live through this in our loud cloud opsware days. We had to lay a bunch of people off. And so you've immediately broken trust with those employees. So you are, from that day forward, basically in a situation where you have to rebuild that trust.

40:58So you have to demonstrate to them that it wasn't capricious and arbitrary and that you're not lying to them and a jerk, but that you actually do care about them. And it turns out through no fault of their own, their next-door neighbor who sat in the cube next to them unfortunately lost their job because we couldn't afford to pay them anymore. So I say that only to say, look, I recognize that very clearly. I understand that. And if you literally go back to the testimony I gave when I went through my hearing, I said that very explicit. I think that's important. All that being said, I think, look, the way to think about, let me just give you the context of the headcount reductions that happened in the government because I think there is, in my mind, at least some misinformation out there that I think is worth talking about.

41:35So let me give you some numbers. You know, you've got a very sophisticated audience. As of the end of this year, roughly about 317 ,000 people will have exited the federal government. Okay. We started the year at about 2.4 million-ish roughly in terms of civilian. These are civilian employees, by the way, so forget about the armed services because my organization doesn't touch that. About 2.4 million civilian employees will be roughly 2.1 million civilian employees as of December 31st of this year. At the same time, we hired about 66 ,000 employees. So on a net basis, call it 260 ,000, 270 ,000 roughly kind of net departures from the government.

42:09If you go back to that 317 ,000 number, 92.5 % of those people took some form of voluntary retirement. So we had this thing called DRP, Deferred Resignation Program. that was literally half of it 154 ,000 people took that which was an 8 month severance that we gave to people to voluntarily leave government and then the rest of it was there are some other programs the acronyms aren't relevant for your listeners they're called VERA and VSEP if people want to look them up but those are other voluntary kind of retirement programs and then only 7.5 % of people were removed through either RIFs or what we call probationary employees which is in the government if you've been there for one or two years, depending on the job, you're kind of the closest thing to an at-will employee that we have in the federal government.

42:53And so there were about 7 ,000 of those probationary employers who were removed and about 17 ,000 people who were ultimately rift, you know, reductions enforced as a result of restructuring. So I say that, number one, not to, I want to be totally clear with your listeners, not to make light of any of this. Obviously, those are real people with real jobs and real families and stuff. But the numbers, again, you know, kind of, I think, bear out what we tried to do in the government, which is tell people that things were changing, tell people that, you know, we do have a different view as to what, you know, we think the federal government should be doing, and give people an opportunity to voluntarily say, okay, this is not what I signed up for, which is a very noble thing to do, and say, you know, it's time for me to go do something different.

43:30So 92.5 % of the people who left the government did so because they decided, you know, this wasn't an environment they wanted to be in. So that being said, now look, I totally agree with you, which is we have to demonstrate to the people that are there and to the people who are coming in that, you know, we obviously obviously care about them and that we think they do valuable work. And I think that's the case for the vast, vast majority of people there do great work. Like any organization, yes, there are some people who are not doing great work and we have a system that makes it very hard to hold them accountable.

43:56We're going to try to fix that. But look, that's true of any organization. So I don't think it doesn't make sense to kind of demean the broad federal government based upon the work of a few. But we have to do that. And we also have to be honest and transparent with our people, which I believe very strongly in is, look, we can't be all things to all people. Like we're running$2 trillion dollar deficits. That's not sustainable. And so look, we've got to be smarter. We have to use technology to find ways to improve efficiency. We have to think about if we're doing 20 things, are all 20 of them still required?

44:21Do they still add value? Because ultimately we're stewards of the taxpayer dollars. Like this is not our money we're spending. And so I think it's incumbent upon everybody. And this is the message that I'm trying to get across to my team and to the rest of government is efficiency doesn't mean just cutting costs. Efficiency means are you thinking about delivering a better service, but doing so, you know, at or below like what you do today, because as you're thinking about organizational changes, you're thinking about technology, like, to me, that is just, like, quite frankly, required of anybody in any organization to do.

44:48And so that, to me, is, like, what I'm trying to convey to people. And yes, that means that there may be some jobs that are impacted because of technology. But those things are going to happen whether, you know, kind of, you know, we want them to happen or not. The question is, do we actually set up the workforce and do training and do all the kinds of things we need to do to make sure that people have opportunities in light of what is a very, very rapidly changing, you know, technological market. So I get that part of the problem that you're trying to solve is an operational one, this idea of like a COBOL programmer who's been there for decades and he retires and walks out the door and no one knows how to do what he was doing or no one understands the system that he's basically had control over for ages.

45:30But at the same time, you know, we're talking about a decent amount of turnover, a lot of new employees, if this goes the right way, How are you, I guess, guiding for safety? Are there any sort of like institutional guardrails that you're putting in so that the government doesn't actually end up breaking things? Yeah. I mean, look, there are, you will hopefully appreciate this, but look, there are a ton of institutional guardrails. You know, in my mind, like, the problem is we have too many institutional guardrails. And again, you know, rational people may disagree with me, but, you know, I can just tell you, like, the entire culture of everything we do, everybody I've talked to in government, and luckily we've had lots of kind of impact on making change early on, everybody is like, we can't do that because of X, right?

46:20And the answer is either because, like, it's statutory and the honest answer is when you dig and dig and dig, the reality is it's not statutory. Like, we built up this, like, urban myth over time that something is statutory. Or, as I mentioned, because, like, I'm terrified that I'm going to get a GAO report, an inspector general report, or a congressional inquiry. And, like, we have built such a compliance culture that we are just literally, like, unable to do anything that actually is kind of modern and, you know, forward thinking in terms of government. And so this is what I'm trying to do. And as I told you guys from the beginning, like, yes, this is not the private sector.

46:54We're not going to take crazy risks. We're not going to, like, go, like, put everything on black, you know, in the casino. But we can have a rational conversation, which is, okay, like, what are the risks we're taking and what's the upside? So if you'll give me a minute, let me give you a very, very tangible example for your listeners. So I mentioned to you that we handle retirement services, and it's a very manual process. And so, like, it takes us a really long time to get people formally retired. So you quit from, you know, you retire from the government on September 30th. In the current process, you know, it takes 60, 90, 120 days sometimes for you to actually get a check because we have this ridiculous manually paper process that takes forever for us to calculate what are you actually entitled to and to make sure you're getting the right money and all that stuff.

47:35So that's a huge burden on retirees, right? You've been getting a paycheck every two weeks for the last 40 years. And now I'm going to tell you, like, sorry, like, go manage your cash flow some other way while I'm doing this. Like, that's terrible in my mind. And it's a complete disservice and completely disrespectful to people who are retiring from the government. So we have a process in the retirement services thing called interim pay, right? Which is, as the name sounds, right? The idea is, okay, can we get you some money on an interim basis that may not be 100 % accurate, but at least helps bridge this gap between when you retire and when you get your actual annuity check, okay?

48:07So going into, and then couple this with we've got this massive, you know, we've had a lot of exits in government. So the 317 ,000 I mentioned, not all of those are retirees, but like on average people, We have about 100 ,000 to 150 ,000 exits a year, so we're more than double the number. And as a result, there's more than double the number of retirees that are happening. So you've got this terrible payable process, massive volumes, and we're just going to make life miserable for all these people who can't get money. So I asked the team. I said, okay, how many people go through interim pay today?

48:36This is kind of pre-our change. And it's about like 30 % of people. And I was like, well, what's the problem? They're like, well, that's how we've always done it. Again, part of the answer was statutory, which it's not statutory. And they're like, well, there's huge risk to the system because if we pay somebody more and then they die between the time we paid them their interim payment and we get their actual annuity, there may not be any assets left into the estate. And so we might compromise the fidelity of the trust fund. OK, now, keep in mind, this is a one point two trillion dollar trust fund that is actually cash flow positive.

49:08So among trust funds for retirement, it's actually very rare. But we earn 2.5 % on government treasuries a year, which is horrendously bad because we're only allowed to invest in government treasuries. But not with spending that. It's actually still cash flow positive and fully funded like almost any circumstance. So we finally went through the math. And I said, let's go through the math. And let's assume, you know, we get it wrong X amount of time and all this other stuff. And we did a bunch of math. And I said, look, if we're wrong 100 % of the time, it could cost us a couple million dollars of overpayments out of a$1.2 trillion trust fund.

49:41But we can shrink from 90 to 120 days, the time somebody gets a payment, to either zero to seven days if we put more people through this interim pay process. So in my mind, like, this is just the most obvious thing to do. And we are doing it, by the way. We decided to do it. But like it's a great example, at least in my mind, of where we have so ingrained in the culture that risk, any kind of risk is bad without actually people being willing to have the conversation to say, OK, like what is the cost of that risk relative to the upside opportunity that that taking that risk entails? and I'm perfectly happy.

50:13You know, I told them, I said, look, you don't need to protect me. If it turns out some congressperson wants to yell at me because like I took a modicum of risk in order to like make, you know, hundreds of thousands of retirees lives better and not have to like take loans out after they've given their career service to the federal government. Like I'm totally fine defending that. I think that's a perfectly rational answer. So that to me is the issue. I have one last question and maybe I'll sort of ask you a little bit to put your VC hat on maybe because maybe you will go back to VC when this rolls over.

50:39But who knows? Someone who knows a lot, a lot about technology. So we had an episode recently with Tyler Cowen. We're talking about one of the weird things. Oh, yeah, he's great. Yeah. Is that, you know, we haven't even really talked about AI. So there's my AI question. AI, the tech, you know, makes your jaws drop when you see it. On the other hand, the impact so far on the economy or the real world is sort of muted. It may be it's like you might have given the capabilities, you might have expected more, more revolutionary change. And his thesis that he put forth was like, we're really not going to see that sort of meaningful, impactful diffusion of the tech until we have organizations that to some extent are built ground up from AI.

51:17That, you know, yes, you can maybe get some productivity gains here and there by putting AI into some existing workflow, but that the real breakthroughs will come from whether it's companies or so forth that are truly AI native from day one. I'm curious, when you think about, again, going back to the VC world, A, does this feel intuitive to you that the real deployers will be AI native? And then does that have any implications for like we're not going to have a revolution, I guess, hopefully, at the federal level, we're not going to have a complete destruction and starting over. Does that have any implications then for thinking about how AI will be diffused within federal processes?

51:59Yeah. So obviously, yeah, Tyler is a brilliant guy. So far from me to disagree with him. But let me give you just a couple of thoughts on what he said. So look, I guess I generally agree with most of what he says, which is, and this is true, by the way, I think of adoption of most new technologies, which is we go to the common use cases and we think about incrementality as the first thing to do it. So if you go way back when Mark Andreessen told me this, so if it's made up, I'll blame him. You know, when we first had movies, you know, actual like, you know, moving pictures, they were just plays basically that people filmed, right?

52:32So like they weren't like producing them. They weren't new content. It was literally just we said, wow, we have this cool thing called a camera. Let's film a play and then go show it in a thing. Now, obviously that's not that relevant to this conversation, but my point is just like we tend to go to these very common incremental use cases for things. And so, look, I believe that if you look at any major Fortune 500 company, I guarantee you that every single board conversation is what the heck are you doing about AI and how do I get 3 % to 5 % to 7 % greater EPS growth relative to last year because you can fix my customer support problem or stuff like that.

53:05So I think that's what's happening right now. I do agree that, yes, what often happens is then there will be a new generation of companies. And we see this in the venture capital world, too, already, which is the most obvious use cases that people did is there's a ton of companies who have customer support applications for AI. And there's a ton of companies who have, you know, how do we help lawyers read documents better using AI? And that's great. And those companies are fantastic. But the real breakthrough is going to be, yes, like when someone says, OK, we're not going to make Salesforce.com more efficient, but like we're just going to rethink customer relationship management.

53:35with an AI first mentality. And like the AI should know who my customers are. It should anticipate what I want to do. And it should send me emails and texts and say, like, you haven't talked to this person in five weeks. And I just saw that they, you know, spoke at this conference and they talked about this and, you know, you're not following up on it. So that's where we will get. And I have no doubt we will get there, quite frankly. And look, I'm an optimist by nature. You have to be to be in the venture capital business. But yes, I would agree with him generally. I think that's right. What does it mean for the federal government?

54:00Very simple, which is that if the private sector is incremental, the federal government is going to be like, you know, 100x incremental relative to 100x less incremental, I guess I would describe it. But that's in my mind, that's okay. Like, so again, I've told everybody in my team, I'm not asking you, I don't want you to go build like the 2040 AI plan, because we have no idea what the heck that's going to look like. That's a total waste of our time. What I want you to do is look at a process you do today and figure out, can you use an AI tool to help you get three, 5, 10 % greater performance out of your organization, right?

54:32So we do a ton of, I'll give you a very simple example, we do a ton of regulation writing, obviously, because, you know, that's what we're ultimately in the business of government is producing regulations. And we're about to put out a regulation, we got 40 ,000 public comments on this regulation, we literally assigned each one of those 40 ,000 comments to a person to read, and to draft a response to. Now, to me, that's just insane, right? And I'm not saying that like the AI, we should just like the AI should do it. But like clearly AI is very good at, you know, summarization, at, you know, helping people think about stuff like that's like that's all.

55:04If we did that, like we would unleash so much creativity and opportunity to govern if we just said take the very basic stuff we're doing today and use your companion chat GPT or XAI, whatever your favorite tool is, and just, you know, eke out a little bit of productivity. So anyway, that's how I think about it. So, yes, look, government is not going to we're not going to transform government overnight. But what we're trying to do and what the U.S. tech force is trying to do is can we get the right people in here who understand those things so that we start to drive ourselves forward so we don't wake up five years from now and find that we're the last dinosaur, basically, and, you know, the world has passed us by.

55:37So you sort of mentioned it just then, but can you talk at all about what kind of platforms you're using currently or what kind of models? Because, you know, I'm sure some people would find it very ironic if the federal government was using DeepSeek or something like that. And I'm curious whether you lean towards like an open source thing or a closed source thing because of privacy concerns. Right now, the honest answer is we're at the very, very early stages, which is most of what people are doing. Literally, this may make your head spin because it made mine. Until about a month ago, we did not even have ChatGPT on our government desktops.

56:11So like, you know, I literally would sit there on my phone and ask ChatGPT questions. My personal phone, just to be totally clear, on my personal phone. And then I would go back to my work laptop and do stuff. So as of about a month or two ago, we actually have ChatGPT. We have XAI now. We might even have a third one. I don't even know. Oh, and of course, the government's a big Microsoft client, so we have Microsoft Copilot. For an individual user in government, that's basically the state. That is like the state of kind of advanced automation, basically, is we literally have it on our laptops now.

56:41For the development work that's happening, you're absolutely right, which is, look, there are major security issues that we all have to think about. Look, I'm a big believer, as you can imagine, in open source and certainly our firm, Andreessen Horowitz. You know, you've seen Mark make lots of public comments about this. So I think the government will use open source and can appropriately do so. Clearly, we're not going to use like the deep seek commercial version. But, you know, so much of that stuff is open source. And if you bring it in, you know, to your environment and you make sure that you get the right security standards, you know, there's tremendous value to develop on open source.

57:11But I would just just to give you a sense of where we are, I would say we're very, very early in that journey in government. I have one last very quick question, which is how much of a constraint is drug testing on new hires nowadays? You know, I wish I had a really smart answer to your question. I don't know the answer to that. I have not seen it. It hasn't shown up as some like big red flag where we're losing people in the recruiting pipeline because of that. But the very honest answer is I don't know. But I will find out. I think that's a good question. We should, I'm actually very curious about drugs in the workforce in general.

57:47We should do more on that. Scott Cooper, thank you. No, for real, for real. You hear about it. My barber was talking about how hard it was to hire because he's like, everyone is sort of. If somebody, I would just say, Joe, if somebody, if somebody clips that and puts it on, Joe is very, I'm very, I'm very curious about drugs in the workforce. That's going to be a great clip. We'll get our producer. We'll clip it. We'll, we'll get ahead. We'll get ahead of that one. Scott Cooper, thank you so much for coming on AdLodge. That was really, really appreciate your take of time. I'm very helpful and I would love to stay in touch and hear more about how these projects evolve during your tenure.

58:19Thank you both for your time. I appreciate it.

58:37I thought that was great. I love I love these conversations. I mean, because I think they're just, I guess the sociology of them. Like that's what these conversations are really about. And especially, you know, we talk about some of these pathologies that replicate from the public to private sector, et cetera, that are both like they're really interesting sociological questions. The Department of War needs to get a bunch of ping pong tables and free snacks to recruit the most cutting edge tech workers. With a matcha bar. That's right. I wonder if they're experimenting. We should have asked that, whether they have a matcha on tap at the office and whether that can move the dial.

59:12But it's also interesting. Again, there's that cultural sociological aspect, but it does apply to large bureaucracies and private organizations. And I think everyone at some point in their lives will experience this type of unnecessary paperwork, whether it's in their jobs or when they're applying for a driver's license or whatever. Yeah. You know what? Like this is like where I have like the sympathy for like sort of like modern life, large, complicated things. How can the individual with our I mean, how could we do it? How could you navigate insurance? How could you navigate all this stuff? It like drives you crazy at any scale.

59:50I do think, and this came up also in our conversations with Jennifer, and I believe it, that there is more to a career than pay, right? But I also think that's a very real issue, especially when you're talking about the tech workforce specifically where we know you can just make an absolute fortune. And in many cases, you can make a fortune as a very young person. You'll be in the right company at the right time and many overnight multimillionaires in Silicon Valley. So it's not just the late stage, although that's where it becomes the most obvious. I really the pay is a very tricky situation.

1:00:25Well, I do think the emphasis on time constraints for the tech force. Yeah, that's really interesting. And that's kind of a shift, again, a cultural shift from the way people have perceived government jobs for a long time. And I can imagine, you know, like if you're a young tech worker, maybe you're having trouble getting into the private sector. you know just last year no two years ago remember there was the wave of like layoffs right yeah absolutely and now everyone is recruiting because of ai but i mean no i mean you're to your point like there are even like there are many talented tech people if they're not right there in the sweet spot we even know that's not what it used to be so i can imagine like you go to the government for two years and it's it's an educational process yeah and you can take some of that knowledge away from you when you then go into the private sector.

1:01:14Although that also opens up the question of rotating doors between the government and private companies. But maybe that's a topic for a different day. I just add, I do think there is, and it's not just this administration, but I think it's been very notable, at least with some parts of the administration or some parts of the rhetoric. There's this sort of like vilification of bureaucrats, which is, again, of sort of a long time thing like bureaucrats has sort of been a kind of a pejor I mean a bureaucrat is a real thing. The B word. Anyway I do think that's an issue and I also think you know like maybe we should just bring back civil service exams everywhere and if you could no seriously and if you could pass this test you can get a job like it's Lindy you know it's like famously you know go back thousands of years in China it's like the big test.

1:02:00I was about to say do you know about the Chinese civil service exams. Those are so interesting and crazy. Shall we leave it there? Let's leave it there. All right. This has been another episode of the Odd Lots podcast. I'm Tracy Alloway. You can follow me at Tracy Alloway. And I'm Jill Weisenthal. You can follow me at The Stalwart. Follow our guest, Scott Cooper. He's at S. Cooper. Follow our producers, Kerman Rodriguez at Kerman Armand, Dashiell Bennett at Dashbot at Kale Brooks at Kale Brooks. For more Odd Lots content, go to Bloomberg.com, Flash Odd Lots, we have a daily newsletter and all of our episodes.

1:02:30And you can chat about all these topics 24-7 in our Discord, discord.gg slash oddlots. And if you enjoy Oddlots, if you want us to go to Pennsylvania to look at a bunch of file cabinets, then please leave us a positive review on your favorite podcast platform. And remember, if you are a Bloomberg subscriber, you can listen to all of our episodes absolutely ad-free. All you need to do is find the Bloomberg channel on Apple Podcasts and follow the instructions there. Thanks for listening.

1:03:27Hi, I'm Stephen Carroll. And I'm Caroline Hepker, here to introduce you to a podcast that brings you the news you need to start your day in just 15 minutes. It's called Bloomberg Daybreak Europe Edition, covering all the top stories across Europe and around the world. Each weekday morning, we're up early to bring you the latest news by 7am. We've got everything you need to know, from geopolitics and global events to economics and what's moving markets. I'm covering it all from London. And I'm in the EU's capital, Brussels. We have 3 ,000 journalists and analysts around the world to tell you what's happening, what it means and why it matters.

1:04:03It's more than just business headlines. From the price of your breakfast to global shifts in power, economics and money aren't just part of the story, they're often the driving force. So start your day with us on Bloomberg Daybreak Europe Edition for the news you need to know and the context to make sense of it. Find new episodes of Bloomberg Daybreak Europe Edition by 7am London time on Apple, Spotify or wherever you get your podcasts.