The Exchanges

Q I'm really passionate about because I think our government needs to rethink a lot of things around how merit works and how accountability works and all of that, and I want to talk to you about those things, but first of all, like, what's, like, the most interesting or surprising thing you've learned so far, like, running OPM? Like, what shocked you, or what's, what's, what's been new to you?

A Yeah, so this, I, I kind of knew this vaguely, but I would say, like, it's really become front and center, so, um, uh, you know, I've talked about this a little bit, like, When you look at, kind of, the way we do performance management in the government, the system is just, it's just completely broken, and it's just so embedded in the culture. So, quick anecdote, um, you know, everybody gets ranked one through five every year, uh, one being the worst and five being the highest, and, you know, in the government, if you look across, you know, government for long periods of time, about, like, you know, 70, 80% of people get ranked a four or five, meaning, like, 70, 80% of the people are well above, like, average, basically, and Literally, like, 0.2% of people get ranked a one or a two, basically, and, you know, you know, as well as I, having run a bunch of companies, look, there's no magic number that should be in that lower quadrant, but look, it's really hard to hire, and people make mistakes, like, you know, a five to 10% turnover of people who just, for whatever reason, aren't able to do the jobs you need them to do, or you may have made a mistake in hiring, like, it's not unreasonable, but what's happened is, like, everybody's gotten very, very focused on these ratings, like, that's, like, kind of how people, Almost, you know, that that's the highlight of their year is gettin…

AI assessment note: “Literally, like, 0.2% of people get ranked a one or a two”

Q of opinions around hiring practices and management and the government and accountability. So I'm excited to talk to you. Yes. Yes. This is very important. But first we're thrilled to talk to you about, there's something you're, you're announcing the launch of, uh, this week, which is the United States tech force. Uh, and tell us about this initiative. What are you, what are you doing with the tech force?

A Yeah. So tech force is really designed to address, uh, two fundamental problems that we have current in the current Uh, workforce and government. One is we just have a shortage of modern software developers with AI talent, and that's just a function of, as you well know, Joe, of just how quickly AI is changing, uh, the world. So we're, we're missing kind of talented people there to help make sure that AI can be a first class citizen in government in the first place. And then the second big part of this is, um, we've got a real, um, early career talent problem in government. So, uh, you and I may have talked about this, Joe, but just to give you perspective, we've got about seven percent of the workforce in the federal government that is early career. And so I'm not supposed to say, you know, ages, but let's call it under the age of 30, uh, is the way to think about it.

AI assessment note: “tech force is really designed to address, uh, two fundamental problems”

Q for certain jobs where we care more about the skill than about forcing someone to go, you know, get, necessarily get brainwashed for four years. And you mentioned Doge. Doge is an example of a department which is able to attract a lot of top engineering talent from Silicon Valley, large part due to Elon Musk, of course. Are there lessons from Doge that are positive for your efforts otherwise?

A I mean, one lesson from a recruitment perspective, of course, is if you can get Elon to be your, like, main talent recruiter, that's awesome. So unfortunately, we don't have that, but, uh, you know, uh, we'll, we'll certainly get, uh, you know, uh, XAI as a partner with us, kind of, in the work we're doing. Yeah, look, I think the big lessons are, the big lesson that we are going to do with this program is, uh, that we learned well from Doge, number one, is, Exactly that, which is get people excited about what the narrative is. In that case, obviously Doge was very focused on things like, you know, efficiency and cost cutting, all of which is great. The, the, the focus for tech force will be on what I would call modernization writ large, basically, which is how do we kind of set the government up to be able to kind of, you know, be a, be a real player in the AI world. So we got to get people excited about that narrative. And then the other thing is making sure that these teams don't get lost in the kind of blob that sometimes is these government agencies. Right? You know, like when I talk to Emil, for example, I'm like, look, you know, what I want you to do is as you take these teams in, make sure whether it's you or someone on your team, like somebody's got to make sure that their objectives are defined appropriately. You can't just like drop them into a 10,000 person CIO orga…

AI assessment note: “the big lesson that we are going to do with this program is”

Q other talent, but he's actually finding a way to pay them more using a nonprofit, which actually I happen to be a donor. I think Mark and others are donors to that as well. And I thought that was really cool because, because in some cases you have to, is it, are we going to be able to reuse that model again in other places for certain types of people?

A Yeah, we certainly can. So you're exactly right. Yeah. What Joe has done is basically what's called an IPA basically. And so the idea is exactly as you described, which is look, the government can only pay what the government can pay just because of the way the pay scales work. And We have to have legislation to fix that, but, you know, as you know, as well as I look, the, the prospect of getting legislation on that is just really, really low. But, but effectively, you can think about the IPA as the IPA can supplement the income for an engineer, for example, so they get paid what they get paid by the government, and then the IPA can actually top them up so that, uh, you know, if you want to hire somebody who's like a twenty-year veteran, you know, you know, master AI person, you can afford to pay them at least an appropriate, uh, wage. So yeah, I think there's, there's lots of opportunities to do that. And look, even within, to be honest, like even within the government itself, you know, we have like, not, not perfect, but we do have flexibilities. We have, you know, we can do bonus payments. We have special, uh, compensation authorities. So like in my department, for example, we can determine something is a really critical, important skillset, and therefore that allows us to go off the traditional pay schedule. You know, it's still got caps, so we can't do crazy things, obviou…

AI assessment note: “Yeah, we certainly can. So you're exactly right.”

Q compete against Department of War, unfortunately, for top talent with how structured right now, but you can't, you can't comment on that, maybe. You authored, Scott, one of the definitive guides to Silicon Valley as well, Secrets of Sand Hill Road. For those who haven't, who haven't read it, what are a few of the key lessons from Silicon Valley? I want to ask that you learned from being there.

A I appreciate it. Yeah. So look, I think the whole idea behind the book was kind of to demystify this process, you know, as we all know, look, you know, Entrepreneurs, you know, might do this once or twice or three times. Not everybody's as prolific, Joe, as, as you are, but, uh, you know, as venture capitalists, you know, we do this hundreds of times. And so the asymmetry of information is just crazy. And so the whole idea was just demystify it, help people understand it so that we can actually encourage more people to be entrepreneurs. So they're not necessarily, and by now they're no longer secrets, but some of the ideas were kind of how VCs evaluate opportunities, how they think about kind of evaluating teams, Um, how to think about kind of the risk profile that BCs were looking for. So, you know, you probably used to get these pitches every now and then where people would tell you this awesome story about they're going to conquer the world. And then they'd say, oh, but by the way, if we don't, don't worry, there's like 10 acquirers who will acquire us. And like, you know, we immediately all pull our hair out. We're like, we don't care about those acquirers. Like we care about the conquer the world strategy. Right. It's like, none of us is in the risk minimization business. Right. So, uh, it's things like that, that hopefully help people understand how the business works. An…

AI assessment note: “none of us is in the risk minimization business. Right. So, uh, it's things like that”

Q Not, not because the test wasn't fair, that's just the nature of our society. And, and so therefore we stopped doing the tests. And, and so in my view, Without any of these tests, you just get people who are not nearly as good sometimes cause you can't screen out the problems. And are we thinking of putting back in tests at all? Is that what we're going to do?

A Absolutely. So first of all, like your history is awesome. And by the way, I've got a, I've got a blog post coming out. I'll make sure you see it shortly on Pendleton act. Um, You seem to, you seem to be wonky like I am on this stuff, which is good. So we'll, uh, I'll be interested in your feedback, but yeah, look, um, the short answer is yes. So you're exactly right, which is the, the test was called PACE, P-A-C-E. It probably stands for something, which I'm forgetting what the initials stand for, but you're right. What happened was basically starting in like, 1978, 79, there was this lawsuit around disparate impact. You're right. Basically that's what happened. And at the very end of the Carter administration, very beginning of the first Reagan administration, the government entered into a consent agreement. Decree that basically said, like, we will no longer use this test.

AI assessment note: “Absolutely. So first of all... the short answer is yes.”