Scott Kupor, now OPM Director, explains how he's applying Silicon Valley talent practices to 2.4M federal employees

business. Now he's the director of OPM for the US government, the office of personnel management. Uh, welcome to the stream, Scott. We are excited to talk to you. How you doing? I'm doing great. How are you guys doing? We're doing fantastic. Um, it's great to have you on the show.

I We're in a very funny era where more folks from Silicon Valley have moved over to the government and every time they do, I have to have them explain to me like I'm in the VC world what they do. I think I learned about the OPM uh that like the entire office uh was a new term to me just a few years ago.

Explain the Office of Personnel Management to me. You mean to tell me that you guys don't know what OPM is? That is really disappointing. I'm I'm shocked to hear that. I probably learned about it about a year ago. Explain it like I'm a venture capitalist that's about to write a thread.

Explaining it like I've always known about it and I'm an expert in it. Yeah, my my wife asked me the same thing when I got appointed to the job. So, I'm not surprised to hear that. Hey, so really quickly, what it is is we are the talent management organization for the government.

So in in simple in our corporate VC terms, we would be the HR department basically for the company. Sure. So our job is like make sure we have the best talent, make sure we have actual things like performance management standards, which by the way we don't have which we're working on.

Uh make sure we have the ability to you know kind of understand the employee base through technology which we don't have currently. So but basically think of us in very simple terms as we would be the HR department inside you know any other you know traditional VC back company.

And and does this cut across all the other departments or are you like wh where where are the boundaries of your Yeah. So um it's it's a hybrid is actually what it is. So basically so like there's some stuff that we have exclusive province on.

So like I'll just give you a simple example like the president puts out a hiring an executive order saying we want merit hiring to be the standard in government which you and I can we can talk about that forever if you want but basically so he says look I direct OPM to figure out like what are the standards that we need to do cross government to make merit hiring a reality.

So we will go put out a bunch of you know regulations and research it and say great okay now this is the law of the land for all the other agencies this is what we expect you to do and you know OPM will not formally approve and issue offer letters and stuff like that unless you guys do that so that's what we do and then but there are HR departments obviously in the different agencies and you know so think of it as like you know in a company you might have like centralized HR but then you have like business partner HR that might sit in the sales organization or other stuff like that and they may be kind of you do things that are domain specific, but all the big things about like how do we recruit people, how do we train them, how do we actually manage them, you know, performance-wise, we will provide guidance for.

Yeah. And sorry, Jordy. Yeah, I was just going to say like getting into it, what what what are your priorities first 90 days, 180, the year? Uh I'm sure there's you you've uh like uh even any any you know, I'm assuming you're trying to bring in best practices that you simp mistakes. Yeah, exactly.

Just ask the model to do we're gonna make we're going to make some mistakes which is okay. That's part of the problem by the way which is like we've been living in this environment where like you literally cannot make a mistake and so nobody actually tries anything.

So yeah, really simply there's three big things we're trying to do. One is create a high performance culture in the talent. Uh so in other words like we my simple goal is look we want like the best and brightest people to take government jobs but as you guys know look A players want to be around A players.

They want to be in a in an environment where they're held accountable where when people do well they actually get rewarded.

We have a very like tenure-based uh and conservative riskmanagement based system here in the government which is like you get power and you get promotion by having more people in your organization and a bigger budget and the longer you spend time in government.

So we've got to just flip that on its head and say look we're going to reward things like innovation. We're going to reward I've been using the term measured risk takingaking because look we can't go just like literally throw stuff against the wall and see what works like we do in Silicon Valley. Yeah.

Um, but anyways, that's thing number one. So, like goal is how do like all the smartest people in the country feel like coming to work for the government actually is great for them. They learn stuff and they'll they'll go get they'll get paid for that in the private sector.

Number two is we've got to make operational efficiency like a core tenant of government. So, again, I know this sounds crazy coming from, you know, the world that I know you and I came from, but today basically um there's no incentive for efficiency.

So, uh, again, you need to like bake it into people's performance plans and say like your job every day is are you doing a great service on the behalf of the American people and are you constantly like turning the crank to figure out how do we use technology? How do we, you know, use AI to do stuff?

How do we like reorg to do things? Again, because spend less money, spend less money, right? Today, again, right, I mean, it's the exact opposite incentive, which is more money means more people means more power, right? And so, like everything we have to do is yes, spend less money to deliver a better service.

And so we've got a ton of like technology initiatives that we can talk about there.

And then the final thing we want to focus on is we've got to make like we've got to anticipate what living in an AI first world looks like for government and we got to not be flatfooted and realize that we have none of the talent we need and we don't even know what the jobs are.

So this is I think a huge opportunity quite frankly where I think we should work with the private sector and say great you guys understand what AI is going to do.

Let's figure out ways where we can have information sharing even like quite frankly swap employees back and forth but find ways so that we don't wake up five years from now and realize there's all these new jobs that are being changed andor augmented by AI and the government is not like you know figuring out how to actually get that talent into the government.

Yeah, you mentioned spending less money. I want to know about the potential of spending more money.

uh is there is is there a world where we remove salary caps on government employees like you you've seen this in Silicon Valley there's not like you know obviously founders are very driven but they have uncapped economic upside like the venture capitalists have uncapped economic upside in so many ways the AI researchers that are getting poached like it is just moneyball at the end of the day for a lot of sectors of the economy is there a reason why the government is fundamentally structurally different or do you think that we uh just need to make a better pitch for civil service?

Yeah, I I think it's a little bit of both. So, okay. So, look, we're never going to solve we're never going to solve like the ultimate like uncapped upside pay gap between the public and private sector. And look, that's just, you know, politically that would never work.

You know, we're not going to, you know, like you know, we're not going to have a Zuckerberg equivalent of like paying engineers a billion dollars a year, right? Yeah. But we can do two things. One is we can change the compensation structure so that yes, like a players actually get rewarded.

So I'll give you a very simple example here. Everybody gets ranked one through five at the end of the year in their performance evaluations in government. Five's the highest, one's the lowest. 70% of the people get ranked a four or five. Okay, which means like they're awesome. They're doing outstanding work. 0.

2% of the people get ranked a one or a two. Now again, mind you, this is across a population of 2. 4 million people where at a minimum you might expect a normal distribution, but like there's just no way that 70% of the people are above average, right? Unless you're in, you know, Lake Wagon basically.

So we got to solve that problem and then if you solve that then the incentive structure can follow it right because again a typical government employee gets 0.

5% of cash bonus at the end of the year that's because again there are too many people highly ranked and the delta between low and high is like you know 50% or something right so what I'd like to do is say you ought to have many fewer people ranked to four or five and let's give those people like most of the bonus pool for example so we can't solve all the problems but like we can solve that and then your your other point is a really good one which is we got to change the pitch.

So again I'll give you a story here. When I first came in one of my one of the employees here said the pitch that we use for government is job stability basically. Like that's the pitch.

And I was like I'm sorry to say but like unfortunately I think somebody's been lying to you because like there's no job in the world where we can guarantee job stability for people even in the government.

To me the pitch has got to be you can see things at a scale and solve really hard complex problems and yes like you can you know do a public service as well but we've got to get people excited about like you know you know putting people on the moon or all the things that you know guys like Elon do you know it's the whole like founder story of like what's the mission of the organization the mission of the organization cannot be we provide job stability so um we've got to like tell that story and we got to go tell it on college campuses and we got to go like show it and look people don't have to commit their life to being in government but if you could come here for two years and work on really hard, really cool problems and you could actually advance in the organization because you're not tenure limited but you actually are skills you know uh you know you're awesome at skills we should be able to promote you faster like there's no reason why we can't compete for really smart people in that environment and then again go take that skill set and go work for you know Elon or go work for Zuckerberg and get paid a billion dollars a year after that like you can do both things so we got to sell the story yeah I was talking to a friend who had some experience in China and he said that the the main difference that he noticed was that the really really top performers over there just kind of got sucked into government work and uh he was kind of framing it as like it wasn't exactly optional for them.

Uh and I know that uh that's not on the table here. We love uh freedom and choice and we kind of decided long to that we actually like democracy is actually a good thing, right? Yeah. Yeah. Of course.

Uh but I I guess my question is like is like should I I know that like when the when the administration changes over I mean this happens every four years now eight years most uh there's there's a push to bring in new folks and there is some recruiting that goes on and and there's uh like whisper networks of uh who's really talented and let's make the pitch to them that maybe now's the time to go and and and serve your country.

And a lot of people do swing at that pitch and they and they take it and some of them do very well with that. Um but should it be more aggressive?

Should have should recruiting be be 10 times bigger in terms of just every talented founder, engineer, venture capitalist, anyone who's high agency, high intelligence, high EQ is at least hearing the pitch for where they would fit in in the US government because I think a lot of people I know a lot of people that got calls.

I know a lot of people who are talented and didn't get calls and I'm wondering if we should have reached them with something. Yeah. And it felt like at the beginning of this year and and maybe the end of last year was just recruit Doge was just dominating recruiting. It's a fantastic recruiting engine. Yeah. Yeah.

But I didn't I didn't see that. I didn't hear like oh I have a a fintech founder who you know they they were great but they got kind of clobbered and now like the IRS is calling them. Like I didn't hear that story that often for example. Yeah. So yeah.

So let me try to just unpack that a little bit because so first of all you're absolutely right. We should 100% do that.

So there's there's a problem we have today which we which we are working on fixing which is today believe it or not um you can actually there are weird hiring procedures in the federal government which makes it very hard in some cases literally to hire people based on their skill skill assessment.

So let me give you a great example.

So people have gotten really good at we put a job description up on this thing called USA jobs which is basically the the website and people essentially tailor their resume to exactly the words that are in that job description and the filtering process puts them at the top of the list but it's all selfassessed stuff.

We we literally like you could hire an IT network administrator in the government today and not have a technical person interview that person to see do they know like what like you know HTTPS means basically.

Um, and the good news is we just actually part of it is because like people are worried about, okay, is it are those tests discriminatory? We actually just had a good legal ruling the other day that clears the path for this.

But believe it or not, literally like we do not have skills-based assessment for many jobs in the government. So the the way to so like it's just really hard it's really hard for those people to come in. Um, we but you're absolutely right, which is look, the real solutions problem is we got to do both.

We got to fix that piece and then we have to literally go tell the story. Like I don't think this is the solution. We're not going to do TV commercials like, you know, the Army and Navy and, you know, we need a few good men basically, but we need something like that. We got to go tell the story.

And you're absolutely right, which is I I would bet there are I don't know. I bet there's thousands of data scientist jobs open across the government today. You as an employee, if you're interested in that, you should just be able to apply once and we should then send your resume to all thousand of those jobs.

Like we don't even do that today basically. So like you literally have to go find out where those thousand jobs are. So there's all this stuff that again it's not they're not bad people. It's just that the system has not been set up for that goal. The system has been really set up to say, let's really be conservative.

Let's really be kind of, you know, careful on these things. And we just we have to change that paradigm. Yeah. It feels like we're fishing with a really large net. You know, you put up a big ad, we need a few good men and whoever gets caught in the net comes into the organization.

And we need to get some harpoons out and go find the really like laser focused people for the perfect job that will thrive. The military, you know, it's been interesting. I I think things have changed, but there was a period where military recruiting, to my knowledge, was was really really tough.

And that's still with a lot of a lot of people that do end up joining having been marketed to and and and and kind of exposed to Army, Navy, Marine Corps, Air Force since they were kids, right? But nobody's getting nobody most most Americans wouldn't know.

They think OPM if anything is other people's money like you were saying and they wouldn't know it. And so how do you how do you get somebody like even open to the idea? You know, a lot of kids today in high school, college, they know they want to work in tech.

Nobody's going through those periods of life knowing they want to work in a specific organization or or very little uh on on the you're absolutely right.

So yeah, and we got look we got to just like have a presence on college campuses and be there just like Goldman Sachs is and just like Facebook is and we got to go tell our story.

Like you know you guys know this but uh I used to work for a head of sales one of my companies and his famous thing was you know uh show it sell it hide it keep it basically right and right now we are hiding it basically and people find us because they are looking for stability as opposed to us finding people who are looking for like an actual real interesting challenge it's an interesting time you know apparently uh CSgrads are struggling to find roles you can you can see the variety of reasons for that you know we just had someone on from claude claude code performs really well.

Now maybe it's performing at the level of a junior engineer, but uh the the opportunities in government where you don't need to be you're not out there competing, you know, with other foundation model Luke Fair like he is a programmer.

It from that massive Bloomberg article, it didn't seem like 90% of what he was doing was programming. It seemed like yeah, he was able to scrape some data together and do some analysis to know how to make a decision, but a lot of it was just sending emails, having phone calls, having meetings.

But Luke's also an example of somebody who's truly elite in in his category. And I'm talking about like the next like 10,000.

But I feel like if you're a new grad and you can program that gives you a superpower that then you can also just be effective to be rigorous and first principal thinker in a meeting where you're just trying to decide how much money should we spend on this thing. Yeah.

Did you see by the way did you see the quote from the uh unnamed government source in that article about why they didn't like why they didn't want Luke? They're like he's not qualified because he didn't go to college and he didn't have 5 years experience. Right. And that is exactly the problem, right?

All of our criteria for evaluating people are based upon credentiing and based upon tenure. And this is where merit based hiling matters so much.

Like to me, if you're Luke, I don't care if you're 18 years old and you can actually function at the level of somebody who like, you know, knows programming at at a high level. Let's pay you that level. Let's hire you at that level.

And and okay, like if you need some help around the edges to like, you know, make sure you can sit in meetings, you know, with your shoes on, that's great. We can teach you that basically. Yeah.

But we can't we can't exclude people from the process because they don't have like you know a sheep skin that says you know they went to Stanford for four years. I mean that's crazy. Yeah. Exactly. How is your AI strategy forming? When people hear AI in government as citizens they get scared. That sounds scary.

If they hear if you're working in the government and you hear AI, you're probably thinking I don't want to lose my my job. Uh but but clearly it should be able to be you know beneficial.

The big thing just feels like get every government employee access to an LLM because before any of the job displacement, any of the cyber punk terminators, like let them search the knowledge quickly, like please. Anyway, what was your take? Yeah. No, you're exactly right.

So today on my government computer, if I want to do something on an LM, I have to like go to my personal computer to do it, right? It's not how many government jobs are email jobs, too. Right. Look, the privacy issues are real and we got to be kind to them.

But like there's but there's but so much of what the government does is already in the public domain. So let me give you a very simple example. So one of the things my team does as I mentioned is we develop like rules and regulations for you know let's say how how you should do merit hiring in the government. Okay.

And the way the process works which I won't bore you with but basically like we put out a rule and then we have to invite public comment on it and literally people submit comments. So we're doing a rule right now where we got 40,000 public comments on this rule. Okay.

some a group of people in my team literally read all 40,000 of those comments and they write a written response to them and do it, you know, the oldfashioned way. Now, I'm not going to say like let's just have let's just get rid of all the people and have AI do that.

But there's no question in my mind that we can get a 10 to 15% easy efficiency advantage if we actually had a very simple LLM that we could train on totally public data. I'm not talking about private data. All that regulation stuff is public. Yep. You know, help people manage that process.

But like we have so we have to start with that.

So like what I what I'm doing here in in OPM is uh I'm actually going to have somebody come in from the outside and like do the full landscape of like what's happening in customer support and what's happening in you know HR organizations and like what are all the use cases out there and then what I'm doing is just telling my team look now your job is to go back and find like what are the small wins we can make on little use cases and again I'm not asking you to like you know get rid of 100% of your people but like you need 10 more heads and and you need to find some place to get 10% more productivity.

We ought to be able to do that within our current budget without actually having to, you know, add headcount. So, we just got to start with little baby stuff that doesn't scare people, get them educated. You're exactly right.

Everybody should have, you know, an LLM on their desktop and then we can tackle the hard problems, which is, okay, are we going to put confidential government data in them and do we need them airgapped and all that crazy stuff, but like we don't even need to do that right now.

We just need to like educate people and then come up with use cases that are based on like literally public data with human augmentation. is the does the federal workforce have a like an an aging problem?

You know, we'll talk with uh CEOs on this show or investors that say, you know, the majority of factory owner operators in the country are going to retire within the next 5 to 10 years, right?

That becomes concerning because there's a lot of just like general knowhow and process that could just kind of get lost or evaporated.

Is that a challenge that you're facing in terms of you know people that have been in the government for 30 plus years and are and are uh getting to the point where where they're ready to retire?

Yeah, if you look at the demographics of the government, I don't know the exact number, but there's no question like the government is is very skewed towards people who are kind of later in their career.

And again, I think this is a problem that we've had, which is we haven't figured out a way to actually get people in at earlier stages in their career because we've got all these impediments that I talked about that make it hard for you if you're a recent Stanford grad or whatever to be able to get a job here.

So, yes, we have to solve that problem. Look, I mean, sure, there's going to be knowledge that walks out the door.

Look, I'm more worried about the long-term thing, which is can do we have a steady pipeline of really smart people who will come here and then then we can solve that aging problem by just, you know, over time kind of changing the demographics.

But yeah, like I don't think I'm not worried like there's a cliff here where, you know, 10 years from now all the experienced people walk out the door, but I do think we have to have like a demographic and a and a set of hiring practices that allows us to actually kind of, you know, build a funnel from the very bottom of of folks.

Yeah, it'd be cool.

I remember maybe in 2012 like Teach for America was really popular and obviously that was like non-governmental but like aligned with this idea of like a a tour of rough service and there was code for America and you know it's like cut out the middleman and just go straight for America like work for the government.

I love that. Yeah. It'd be great culturally if it was like yeah somebody So because now like I feel like if you did four years just in civil service and then you jump into Silicon Valley it's kind of like a weird track but like it shouldn't be. It should be completely normalized. Yeah.

A long time ago, right there, and this is not the perfect analogy, but there used to be something called like the civilian conservation corps, which was exactly that, right? It was like people came right out of school, they did some heavy labor for a while, and then they went to the private sector.

So, yeah, like we ought to do this. So, like what I would love to do is I would love a program where we say, you know what, take all the smartest people, have them come here for two years. By the way, if we can do this, let's forgive your let's forgive some student debt as part of that, right?

So, like let's actually have a real trade-off, which is you do public service and we'll give you some financial compensation and oh, by the way, I want to work with the private sector and say, you know what, we're going to get a private sector employer to guarantee you a job when you finish your two years here and they might even give you credit full credit for those two years so you come in as a third year and that and like just work on that stuff, right?

So, uh you know, that's what we need to do and and look, I've been using this term that's terrible that my kids joke at uh which is make government cool again, which is magaka if you pronounce it out loud. Magaka.

my uh my my uh my uh marketing team who's sitting here next to me off camera will tell me doesn't run off the tongue doesn't really yeah it's not it's not the greatest but like that's what we have to do like we have to make civil service like super exciting and the good news is like we've got no shortage of like moonshots now right so like AI right like how do we transform the government with AI that's a great moonshot let's go do that but we need like a you know Apollo mission-like moonshot which I think the president has given people with AI and we got to go tell that story and you If we can do that, then I think like we're in a totally different situation.

Are you going to be visiting college campuses for career fairs at all? Are you going to be pounding the pavement? I would love to. Look, you all know this having been in the valley. Look, if you're in the valley, 99% of your job is recruiting basically.

Like that is like probably the most important thing that any manager does. And yeah, like I don't think any of us is beneath that. We got to go sell. So yeah, I'd love to go to campuses. I'd also look I I want to we've been talking about campuses.

I also want to make clear though, one of the priorities the president has, which I totally agree with, is we shouldn't use credentiing as a shortcut for this stuff. So look, yeah, I don't care if you went to college or not, if you have the skills to do this job, like we got to go find you too.

U look, college college campuses are nice aggregation points and make it more efficient, but like we need to go find anybody who has a, you know, technical degree or not a technical degree or a trade degree or anybody who's just really smart and capable and go find those people.

And we should hire them not because they, you know, have a, you know, a diploma, but because they actually can contribute to the workforce or even like experience at a startup, like they've been there for a couple years, they vested, they run to the next thing, go do a tour in the government. Sounds good. Yeah.

Last question. Have you had any pleasant surprises since starting your job? Anything that you're like, "Oh, that's actually run pretty well. " Yeah. You know, look, I'll be so to be totally honest.

Look, um, there are my pleasant surprise, which is I don't think is a surprise, is look, there are very good people here who are who are really care about what they're doing.

So, look, you know, everybody in the government here, there is a serious element of public service and they all take it very seriously and believe it and believe it and and it's incredibly like noteworthy.

The what I what I really think is look I think the system has failed them which is look as as leaders and managers we have created a system where you know we've told them do not ever make a mistake do not ever risk innovating because we don't you know look you guys know the valley right in the valley we always talk about look we're looking for uncapped upside on stuff yeah you get fired for not taking the risk that's exactly right there is no value today in in maximizing for uncapped upside if that means you also take an extra 5% of downside risk and again look I want to be totally clear we can't be cavalier you know we're dealing with people's you know, social security or we're dealing with national security.

So, we can't just, you know, literally throw caution in the wind, but we can accept that there are some things where we may not get in the 100% right and then we'll fix them when we do. And we have to have a culture to do that.

So, that to me has been again the very positive upside surprise, which is and and which I hope will prove true, which is if we can change the incentive system, I think you have really good people who will do what the incentive system incents them to do basically.

But right now we just we have an incentive system that in my mind at least kind of encourages behavior that just is not you know efficiency maximizing is not innovation maximizing uh and you know is just outdated I think in a world where you know we know technology and the needs of the government are going to change very rapidly.

We need to be innovation maxing. Let's do it. I love it. Thank you so much for joining. Thank you for serving our country and and rallying so many smart uh people to join you and thanks for breaking it down for us. This is very helpful. Thank you so much. Appreciate the time. You take care. Bye. Cheers. Hi.

We need a Luke Ferator for the CIA. We need to send Luke Ferator to North Korea. Infiltrate Pyongyang. Get behind enemy lines and turn it into a capitalist utopia. Make South Korea look like a complete blackout by comparison. You've seen that map of the North Korea is dark and South Korea is super bright.

More lights in Pyongyang than all of South Korea combined. Luke Fair on the front lines. Get it done. This is breaking. This is breaking in real time. So, John actually hit up Peter Salis over at Discord. Uh, HR apparently just threw a note on uh Peter Salis just posted guys at TVPN WTF. John Xi, I think I'm in trouble.

HR just threw this on my calendar. It's Peter and HR discuss TVPN issue. Wait, is this real now? No, I think I think he's joking. But, uh, very good stuff. I love it. That's amazing. Thank you for sharing that. Peter, come on the show. We're going to do a uh we're going to figure out your cloud uh your cloud.

We're gonna have people live bidding. Yeah. Yeah. Yeah. We should have three. We should be Shark Tank. He's the shark. We have three hyperscaler uh cloud SDRs. Come on and pitch. Okay. This is what we can bring to you. This is what we're asking. We're asking for we we got