[00:01.200 --> 00:05.380]  All right, now let's talk like civilized fucking human beings for the rest of the show, shall we?
[00:06.140 --> 00:07.160]  All right, excellent.
[00:07.920 --> 00:13.820]  We are going to be having our two favorite people, an ex-government employee and a journalist talk together.
[00:14.100 --> 00:22.680]  And they are going to cover a wide variety of subjects, and they will hopefully take it somewhere that we'll all learn a little bit today.
[00:22.860 --> 00:26.120]  Because learning is really what it's all about, that and giving to the EFF.
[00:26.120 --> 00:26.920]  All right.
[00:28.260 --> 00:31.300]  First, let's have Adam Klassfeld come on out.
[00:31.580 --> 00:31.760]  Adam!
[00:38.750 --> 00:39.710]  Adam, check your mic.
[00:39.890 --> 00:41.030]  Make sure that we're all good there.
[00:48.060 --> 00:49.940]  That's why you pack two parachutes.
[00:50.000 --> 00:51.340]  I can just project.
[00:51.700 --> 00:52.460]  Two parachutes.
[00:52.600 --> 00:53.240]  That's why you pack them.
[00:53.240 --> 00:53.600]  Hello.
[00:54.080 --> 00:54.340]  Oh.
[00:55.320 --> 00:55.720]  Awesome.
[00:57.940 --> 00:58.940]  How about this one?
[00:59.140 --> 01:02.760]  Oh, we're going to do the world famous pass, we're going to do the famous toque format.
[01:02.900 --> 01:04.200]  All right, toque format it is.
[01:04.440 --> 01:05.600]  All right, there we go.
[01:06.560 --> 01:09.700]  And, of course, the man of the hour, Sanhil Avengia.
[01:16.990 --> 01:20.290]  X with emphasis on the X, ex-DOGE engineer.
[01:21.030 --> 01:26.510]  All right, everybody have a wonderful time, and I'll be giving a talk later today, and as will many other people.
[01:26.670 --> 01:30.350]  Thank you, and again, to all the newcomers, thank you for coming to HOPE.
[01:35.550 --> 01:37.490]  Well, thank you all for coming.
[01:37.750 --> 01:46.210]  I want to say that the MC actually gave Sahil fair warning of those booze a little bit before.
[01:46.410 --> 01:48.570]  We didn't really understand what that's happening.
[01:48.570 --> 01:50.390]  I thought you guys weren't allowed booze on campus.
[01:50.630 --> 01:51.070]  I know.
[01:51.450 --> 01:52.130]  I know.
[01:52.290 --> 01:52.790]  Absolutely.
[01:53.890 --> 01:57.530]  But, um, just a little bit of introduction.
[01:57.910 --> 01:58.670]  My name's Adam Klassfeld.
[01:59.390 --> 02:03.410]  I'm the founder and editor-in-chief of All Rise News.
[02:03.750 --> 02:07.490]  I've been covering Legal Beat for about a couple decades now.
[02:07.490 --> 02:11.690]  And very pleased to talk to Sahil over here.
[02:11.970 --> 02:12.610]  I think this works.
[02:12.770 --> 02:13.750]  Okay, fantastic.
[02:13.930 --> 02:14.730]  Tech support over here.
[02:14.870 --> 02:15.790]  There we go.
[02:16.410 --> 02:18.130]  So, yeah, why don't we just start things off?
[02:18.230 --> 02:21.190]  Tell me about what led you to DOGE.
[02:21.670 --> 02:30.430]  Yeah, so I actually applied to the original DOGE, the U.S. Digital Service, which DOGE is technically, in 2015 or 16.
[02:32.090 --> 02:36.610]  DOGE, or the USDS, was set up under Obama, under the White House.
[02:37.290 --> 02:43.710]  Actually, Harper, who's over there, probably was involved to some degree, maybe, or certainly knows people.
[02:44.250 --> 02:47.110]  But basically, the goal, you know, healthcare.gov rolled out, kind of flopped.
[02:47.710 --> 02:51.750]  And so they were like, we need technology people, you know, in the White House.
[02:51.970 --> 02:56.930]  You know, the federal government isn't used to, like, shipping, like, consumer product.
[02:57.730 --> 02:59.250]  It's kind of new for them.
[02:59.250 --> 03:02.090]  So they created the USDS, I applied, I never heard back.
[03:02.370 --> 03:06.290]  So that was the first time I applied, was 2016, or 15 or so.
[03:06.490 --> 03:13.330]  And then, when I heard about this sort of DOGE project starting to spin up again, I applied again, I guess.
[03:13.930 --> 03:18.070]  And my goals, I'm not sort of, like, super politically active.
[03:19.110 --> 03:21.010]  I just want to, I'm a software engineer.
[03:22.050 --> 03:24.730]  And I wanted to ship code for the federal government.
[03:25.110 --> 03:28.270]  That's, I think that's, like, a really great way to make the world a better place.
[03:28.270 --> 03:38.510]  Like, people who've interacted with the federal government and probably aren't super impressed with, like, the HTML, CSS, JavaScript that runs the place.
[03:38.930 --> 03:42.290]  But we have, you know, there's like 500 million consumers of that software.
[03:42.730 --> 03:47.310]  Every year, whether you like it or not, you're forced to be a customer of the largest company on earth.
[03:48.370 --> 03:53.190]  So I thought, you know, being a software engineer would be a good way to have a positive impact.
[03:53.450 --> 03:57.510]  And for better or worse, you know, this was the path to doing it.
[03:57.590 --> 03:58.750]  DOGE, Trump, et cetera, et cetera.
[03:58.950 --> 04:03.210]  And in between that, you worked on the Bernie Sanders campaign as well?
[04:03.210 --> 04:08.370]  Yeah, so yeah, in 2016, I, yeah, I have volunteered for the Bernie campaign.
[04:09.810 --> 04:15.790]  So, you know, yeah, that's my political, you know, affiliations are just trying to make the world better.
[04:16.090 --> 04:16.510]  Would be cool.
[04:17.790 --> 04:19.730]  So what, just walk me through the process.
[04:20.010 --> 04:26.450]  That after Trump's election, you see the opportunity of working for the federal government with DOGE.
[04:26.450 --> 04:29.550]  I know that your wife had a certain reaction with this.
[04:29.670 --> 04:31.190]  You've spoken about this publicly before.
[04:31.810 --> 04:35.410]  Tell me about the process of starting to work for DOGE.
[04:35.810 --> 04:38.870]  Yeah, most people, you know, in my life were not pleased.
[04:39.730 --> 04:41.030]  You know, I live in Brooklyn.
[04:42.570 --> 04:53.050]  So, but, you know, unfortunately, I think, I think it's like, I don't know, like if you want to make the world a better place, like, you know, if that's your core priority, like, who cares?
[04:53.350 --> 04:53.630]  Right.
[04:54.030 --> 04:55.830]  You know, like everyone has a shitty boss.
[04:56.890 --> 04:59.150]  That's like what a, you know, that's what bosses are.
[04:59.830 --> 05:02.990]  But, you know, you do the work because you want to serve the customer, you know?
[05:03.330 --> 05:05.870]  So, so, uh, uh, yeah.
[05:05.970 --> 05:08.670]  So I, I, you know, I, I, what was your question exactly?
[05:08.790 --> 05:09.470]  Like what, like what?
[05:10.890 --> 05:16.630]  The pet, so basically lead us through it that you were, you applied for the job.
[05:16.910 --> 05:17.210]  Yeah.
[05:17.470 --> 05:24.350]  I mean, I, you know, as soon as I found out this is like pre, pre Trump winning the election, you know, DOGE was kind of starting to become like a topic, right?
[05:25.150 --> 05:29.890]  Uh, and, uh, and so I just tried to like find, sort of find somebody who worked there.
[05:30.030 --> 05:33.810]  And randomly, uh, I was at an event and someone, I met someone there.
[05:33.910 --> 05:34.610]  We grabbed dinner later.
[05:34.610 --> 05:40.150]  And he mentioned that he had a friend or he had not even a friend, but he had met someone at some other party.
[05:40.150 --> 05:42.650]  This is how like people, you know, this is social networking, right?
[05:42.730 --> 05:43.770]  Like pre the Internet.
[05:44.270 --> 05:47.910]  Uh, he, he, he knew somebody who was, who was working at DOGE.
[05:48.070 --> 05:50.150]  This is maybe like October or November or something of last year.
[05:50.570 --> 05:52.750]  Uh, and so I was like, Oh, can I like get an intro?
[05:52.910 --> 05:55.330]  I'd love to like ship software for the federal government.
[05:55.410 --> 05:56.930]  So he just introed me.
[05:57.050 --> 05:57.690]  And then I talked to him.
[05:57.810 --> 06:00.570]  I talked to a few other, other people on, you know, entirely on signal.
[06:01.130 --> 06:04.610]  Uh, and then they sort of, they didn't really even make an offer.
[06:05.070 --> 06:05.890]  There's no offer.
[06:05.890 --> 06:07.570]  I wasn't paid or anything like that.
[06:07.790 --> 06:09.650]  It was just like, when can you move to D.C.?
[06:10.110 --> 06:11.430]  And, and who was the intro to?
[06:12.090 --> 06:16.310]  Uh, so the, the first engineer I talked to, or the first few people, I don't, I can't really name them.
[06:16.370 --> 06:18.950]  They're, they, you know, they, they, they haven't like publicly doxed themselves.
[06:19.170 --> 06:24.110]  So, uh, but, but the final interview was with this guy, Steve Davis, who was like the de facto head of DOGE.
[06:24.250 --> 06:24.570]  Okay.
[06:24.770 --> 06:29.610]  Well, you know, sorry, just to kind of, um, explore that a little bit more.
[06:29.710 --> 06:33.490]  I mean, we've heard that term doxing and applied to DOGE in the past.
[06:33.570 --> 06:35.270]  They've accused some folks of doxing.
[06:35.270 --> 06:47.510]  And, but, uh, the common response to that is we work for the government that, you know, why would that be, um, you know, are they working for the government in the sort of secret capacity?
[06:47.570 --> 06:48.630]  I guess is the question.
[06:48.770 --> 06:48.870]  Yeah.
[06:48.930 --> 06:54.270]  I mean, I, I mean, I'm, I'm, I'm a big believer in like, I think everyone who works for the U.S. federal government should be doxed.
[06:55.310 --> 06:55.670]  Right.
[06:55.990 --> 06:58.690]  But like, that's not how the U.S. federal government operates right now.
[06:58.890 --> 06:59.150]  Right.
[06:59.150 --> 06:59.170]  Right.
[06:59.310 --> 07:02.050]  I know all 500,000 employees who work for the VA.
[07:02.210 --> 07:03.410]  Cause I worked at the VA.
[07:03.650 --> 07:06.490]  I had access to all the HR data, uh, at VA.
[07:06.790 --> 07:13.770]  Uh, you know, all 473,000 people, uh, all their social security numbers, how much they get, got paid, all their promotions, all that kind of stuff.
[07:14.510 --> 07:16.310]  Um, should the world have access to that data?
[07:16.510 --> 07:18.630]  I, I think, I think so.
[07:18.630 --> 07:18.890]  Yeah.
[07:18.950 --> 07:19.010]  Yeah.
[07:19.010 --> 07:20.530]  You work for the U.S. federal government.
[07:21.350 --> 07:28.030]  I don't disagree with you, but on a personal level, if you're a friend of mine and you have a job as a career employee, I'm not going to like share your name on the end.
[07:28.930 --> 07:29.290]  Understood.
[07:29.650 --> 07:36.470]  So, but I, you know, I, I, I, I am like one of the few people who've worked at DOGE who has, you know, I'm publicly talking to the public.
[07:36.590 --> 07:38.950]  That's a pretty rare situation.
[07:39.310 --> 07:39.550]  Yeah.
[07:39.690 --> 07:47.390]  As a matter of fact, your, your willingness to speak to the press, um, has shed so much light on the way that DOGE operates.
[07:47.390 --> 07:47.630]  Yeah.
[07:47.710 --> 07:49.210]  I'm not, well, I don't give a fuck about the press.
[07:49.310 --> 07:54.070]  I care about talking to the public and like educating people on what, you know, what we're doing.
[07:54.230 --> 07:56.110]  Cause I think transparency is, is good.
[07:56.210 --> 08:00.070]  The press is like, um, uh, the media literally is an intermediary to the public.
[08:00.170 --> 08:01.970]  Well, we'll, we'll, we'll, we'll get into that.
[08:02.110 --> 08:09.970]  Quick question about, you mentioned that the predecessor agency of, um, DOGE is the U.S. digital service created by Barack Obama.
[08:10.330 --> 08:22.270]  I remember one of the first reactions that I had to the creation of DOGE was that it was, of course, named after, uh, cryptocurrency that, uh, Elon Musk had heavily promoted.
[08:22.530 --> 08:24.850]  Um, do you have any reaction to the rebrand?
[08:25.110 --> 08:26.870]  What, what it signaled, what it meant?
[08:28.450 --> 08:32.030]  Um, I mean, I think, I think it's nice to have like one syllable.
[08:32.950 --> 08:35.890]  Like, I think, I think it's, uh, I think that's great.
[08:36.270 --> 08:40.050]  Uh, so, so you thought of just basically branding.
[08:40.050 --> 08:44.150]  I think, you know, I guess what I'm getting at here is, uh, a lot of folks...
[08:44.150 --> 08:45.110]  I don't own any DOGE.
[08:45.250 --> 08:46.110]  No, no, no, no.
[08:46.310 --> 08:47.870]  I wasn't accusing you of anything.
[08:48.150 --> 08:48.810]  There's nothing like that.
[08:48.810 --> 08:49.430]  I own some Bitcoin.
[08:49.910 --> 08:50.290]  No, no.
[08:50.470 --> 08:54.650]  I, I, what I'm talking about, you talked about, you know, a medium to, to the public.
[08:54.650 --> 09:00.110]  I think a lot of, um, folks in the public, their first reaction is, oh, DOGE, the, the cryptocurrency.
[09:00.450 --> 09:02.770]  It's getting a rebrand, uh, that is associated...
[09:02.770 --> 09:04.790]  To be honest, I forgot that cryptocurrency existed.
[09:05.030 --> 09:05.430]  Right, right.
[09:05.490 --> 09:08.030]  I think DOGE has sort of superseded DOGEcoin.
[09:08.170 --> 09:10.710]  So for better or worse, it might be a, like, negative impact on the...
[09:11.170 --> 09:11.370]  Right.
[09:11.430 --> 09:11.810]  ...the value.
[09:12.290 --> 09:12.650]  Right.
[09:12.890 --> 09:16.470]  Uh, tell me a little bit about, um, your, there was a really interesting thing.
[09:16.490 --> 09:21.230]  And you wrote about it on your website, uh, in a diary entry, DOGE Days.
[09:21.490 --> 09:22.930]  I, uh, I encourage folks to read it.
[09:24.650 --> 09:28.590]  About your first days, just kind of like day one on, on DOGE.
[09:28.690 --> 09:29.410]  What does that look like?
[09:29.830 --> 09:30.230]  Yeah.
[09:30.350 --> 09:30.830]  Day one.
[09:31.050 --> 09:39.270]  So I, my job, my official job title was, uh, senior advisor to the chief of staff, uh, for the veteran affairs department.
[09:39.690 --> 09:43.390]  Uh, so that's kind of, that was my, you know, that's where I showed up.
[09:43.510 --> 09:51.870]  You know, uh, I showed up 7 45 AM Monday, you know, March 17th at, uh, you know, VACO, the veteran affairs central office in D.C.
[09:52.530 --> 09:56.050]  Uh, and, uh, they're like, who are you?
[09:56.290 --> 09:58.710]  You know, this is just like VA people, you know, career people.
[09:59.050 --> 10:01.550]  And they're like, no one's joining the government right now.
[10:01.670 --> 10:02.710]  Like who, what are you doing here?
[10:02.810 --> 10:03.950]  And I'm like, well, I'm, I think I am.
[10:04.410 --> 10:06.710]  Uh, they're like, okay, you gotta go get fingerprinted.
[10:06.710 --> 10:08.490]  So they, you know, they send me to the fingerprinting office.
[10:08.630 --> 10:09.270]  I get fingerprinted.
[10:09.330 --> 10:12.570]  I get my government sort of PIV card that lets me into any government building.
[10:13.070 --> 10:15.570]  Uh, and then I sit through three hours.
[10:15.710 --> 10:17.090]  Have you ever joined like a company?
[10:17.730 --> 10:19.230]  Uh, you know, as an employee.
[10:19.530 --> 10:19.590]  Right.
[10:19.590 --> 10:21.490]  The normal, the normal day one stuff.
[10:21.490 --> 10:23.970]  You like sit through three hours of HR.
[10:24.230 --> 10:24.590]  Right.
[10:24.690 --> 10:30.870]  Where they like tell you about your benefits that you're not going to qualify for, but you need to sit through this presentation and they give you a shitty laptop.
[10:31.270 --> 10:31.510]  Right.
[10:31.790 --> 10:35.630]  Uh, and, uh, and then they say, okay, like we've done that.
[10:35.690 --> 10:39.370]  Like who, who, like, what is you, what are you here to do?
[10:39.490 --> 10:40.390]  Like, where are you going?
[10:41.010 --> 10:41.950]  And I'm like, I don't know.
[10:42.570 --> 10:48.510]  Uh, you know, it was like looking at each other and I'm like, they're like, what are you here to do?
[10:48.610 --> 10:49.430]  And I'm like, I don't know.
[10:50.030 --> 10:53.050]  Uh, I just want to like ship software for the federal government.
[10:53.510 --> 10:56.070]  Uh, and so they're like, do you have, who's your boss?
[10:56.170 --> 10:56.930]  I'm like, I don't know.
[10:57.590 --> 11:03.670]  Uh, and so, you know, it took a little while to, to sort of figure out like what was, what I was supposed to be doing.
[11:03.830 --> 11:06.350]  Eventually I like, you know, I signaled some people.
[11:06.510 --> 11:07.670]  I was like, you know, where do I go?
[11:07.770 --> 11:09.170]  And they're like, Oh, you know, talk to this person.
[11:09.170 --> 11:16.430]  And he's, you know, like, and then he sort of, he told me, like, okay, you know, one of the projects we're working on is, you know, we're doing all these contract reviews.
[11:16.610 --> 11:23.190]  Like Harper mentioned Deloitte, you know, I don't think, I don't know if we were successful in cutting Deloitte's VA contract.
[11:23.690 --> 11:39.690]  Certainly it is, it should be de-scoped to some degree, but yeah, so that was kind of like, you know, the sort of day ended with, like, kind of going through, like, the other things that DOGE was, sort of the DOGE people were working on at VA, which mostly revolves around,
[11:39.710 --> 11:41.330]  like, reviewing contracts.
[11:41.530 --> 11:41.890]  Right.
[11:42.130 --> 11:49.510]  Trying to help with, like, the RIF, the reduction in force implementation, which the VA has publicly said they're not doing, actually.
[11:49.870 --> 11:54.730]  So the VA has actually hired more employees this year than have let go, so.
[11:54.870 --> 11:55.810]  There's a lot of nuance.
[11:55.930 --> 11:58.250]  That's part of why I want to talk publicly about this kind of stuff.
[11:58.370 --> 11:58.530]  Right.
[11:58.650 --> 12:00.850]  It's not as easy as, like, some one-syllable headlines.
[12:01.070 --> 12:10.210]  Well, that, you know, there was a very big public announcement that there would be mass layoffs at the VA, and then there was public pushback against that.
[12:10.330 --> 12:12.530]  But before that, I just kind of, there was an interesting thing.
[12:12.710 --> 12:16.990]  Like, when did you actually find out you were going to, you were assigned to the VA?
[12:17.630 --> 12:19.050]  Well, that was, that was before I joined.
[12:19.190 --> 12:29.850]  I mean, they told me, like, hey, we, you know, I was talking to the, to, to sort of the DOGE, the DOgE people, uh, and then I got, like, a sort of a signal message out of the blue that was, like, hey, we have a spot for you at VA if you're interested.
[12:30.410 --> 12:31.450]  And so I was, like, sounds good.
[12:32.370 --> 12:33.010]  You know?
[12:33.330 --> 12:35.230]  That was, that was, like, the entirety of the conversation.
[12:35.230 --> 12:40.470]  And it was quite a long time before you actually had that meeting with Elon Musk.
[12:40.490 --> 12:43.390]  I think that was, uh, kind of abbreviated as an e-meeting.
[12:44.010 --> 12:45.050]  Yes, the e-meeting.
[12:45.310 --> 12:47.310]  Yeah, the Elon Musk meeting, I think, is what it stood for.
[12:47.410 --> 12:48.530]  Not electronic meeting.
[12:48.850 --> 12:50.070]  Uh, it was not electronic.
[12:50.250 --> 12:50.790]  It was in person.
[12:51.170 --> 12:55.010]  Uh, yeah, so that, that was, like, a, maybe a couple weeks into my, my tenure.
[12:55.010 --> 13:00.710]  And before that, and this was fascinating, there's this really interesting interview on Planet Money, NPR.
[13:00.910 --> 13:02.010]  Folks should listen to it.
[13:02.430 --> 13:07.270]  Uh, it took place, uh, you, you had what was called a political alignment test.
[13:07.650 --> 13:09.210]  Can you, can you tell me about that?
[13:09.330 --> 13:13.730]  Yeah, political alignment interview, uh, which I think every, this is, like, not like a DOGE.
[13:14.070 --> 13:17.290]  I think, I'm sure DOGE would prefer not, for it to not be a thing.
[13:17.290 --> 13:27.670]  Uh, but, uh, generally, if you, if you, if you join, uh, the administration, uh, you know, whether it's Democrat, Republican, whatever, uh, you have to go through a political alignment interview.
[13:28.150 --> 13:33.150]  And so, basically, Trump's not gonna hire you if you, if you are a massive Democrat.
[13:34.070 --> 13:34.450]  Right?
[13:34.510 --> 13:35.050]  And vice versa.
[13:35.490 --> 13:35.870]  Also true.
[13:36.410 --> 13:39.590]  Uh, Obama's not gonna hire you if you're, like, a massive Trump supporter.
[13:40.150 --> 13:47.030]  Uh, so that's part of the interview process is to make sure that you are not the, you know, potentially the enemy.
[13:47.610 --> 13:54.170]  Well, I, let's get into that a little bit, because the specific questions seemed very, it was Trump specific.
[13:54.430 --> 13:55.730]  There was a question about tariffs, right?
[13:55.750 --> 13:55.930]  Correct.
[13:56.010 --> 13:56.430]  There's a question.
[13:56.530 --> 13:57.490]  Yeah, they love that question.
[13:58.250 --> 13:59.690]  What are your opinions on tariffs?
[14:00.290 --> 14:03.610]  And what were some of the other questions of the political alignment interview?
[14:04.630 --> 14:17.930]  Um, I mean, you know, you basically, I mean, just go to, like, whatever, trump.com or whatever his website is, and, uh, just go through and say, like, you know, like, are you, you know, are you massively against the, any of these things?
[14:18.170 --> 14:22.810]  Like, you know, what are your, what are your biggest problems with, with, uh, with, with Trump?
[14:23.130 --> 14:28.570]  Uh, I assume similar to, like, you know, if you're trying to get a visa to America or something like that, those kind of questions.
[14:29.070 --> 14:34.450]  Um, but again, I think, I think every, it's, it's terrible, but, but this is just the way the federal government works.
[14:34.850 --> 14:37.670]  Oh, the, but they also asked who you voted for, correct?
[14:38.030 --> 14:40.290]  I don't think they ever asked me specifically who I voted for.
[14:40.450 --> 14:41.850]  I can't remember the, like, specifics.
[14:42.250 --> 14:46.010]  Uh, they did ask me if I had prior, previously donated to anyone.
[14:46.410 --> 14:46.810]  Okay.
[14:47.250 --> 14:52.130]  They, the, uh, NPR had, on that episode, they characterized it as a loyalty test.
[14:52.290 --> 14:54.050]  Would you think that's a fair assessment?
[14:55.070 --> 14:59.110]  Um, I mean, I, I think, look, it's politics.
[14:59.290 --> 15:00.390]  Like, loyalty doesn't exist.
[15:00.930 --> 15:12.050]  Uh, there's, like, I think that loyalty is actually a terrible thing, uh, because, like, we have the United States of America and the Constitution and Americans, uh, who I'm loyal to, like, who's president.
[15:13.210 --> 15:15.370]  You know, inherently, you should not be loyal to, to a person.
[15:15.550 --> 15:15.570]  Right.
[15:15.570 --> 15:16.710]  Uh, but it's alignment.
[15:16.890 --> 15:17.670]  It's literally alignment.
[15:17.810 --> 15:19.990]  Like, we have a, we were, like, voted on.
[15:20.110 --> 15:22.010]  Like, we're, you know, we were elected on certain things.
[15:22.010 --> 15:24.350]  And, like, we're here to, like, execute on that specific plan.
[15:24.350 --> 15:28.570]  And if you are going to help us achieve these goals, like, you're sort of, like, draw two vectors.
[15:28.610 --> 15:33.050]  Like, how, what's the degree of, you know, of separation between these complementary arrows?
[15:33.270 --> 15:37.710]  And, like, if it's, like, 180 degrees, like, you know, you should not be here.
[15:37.810 --> 15:40.070]  You're just gonna, like, slow us down or not do anything.
[15:40.730 --> 15:42.670]  Well, I'll, I'll tell you one of the reasons that I'm...
[15:42.670 --> 15:45.470]  I'm not a die-hard Trump supporter, so I had to, like, you know, say...
[15:45.470 --> 15:45.590]  Oh, totally.
[15:45.590 --> 15:46.690]  I'm, I'm not saying that.
[15:46.870 --> 15:50.150]  The reason that I ask, we had a conversation before today.
[15:50.450 --> 16:04.430]  And one of the things that you said that I found really striking is that, um, one of the, the, a lot of the people who have your skill set, um, wouldn't necessarily, uh, might find the...
[16:04.430 --> 16:12.310]  that it might, uh, having such a, uh, a test could stop the White House from getting good talent, correct?
[16:12.630 --> 16:13.010]  A hundred percent.
[16:13.310 --> 16:13.730]  A hundred percent.
[16:13.950 --> 16:14.310]  Totally.
[16:14.690 --> 16:21.670]  Do you think that that kind of test, this political alignment test, is standing against DOGE's stated goal?
[16:22.270 --> 16:24.990]  I mean, it's, it's, it has nothing to do with DOGE or not.
[16:25.110 --> 16:25.450]  It's Trump.
[16:25.890 --> 16:26.350]  I mean...
[16:26.350 --> 16:26.490]  Right.
[16:26.710 --> 16:27.550]  A hundred percent.
[16:27.770 --> 16:30.310]  I don't, I don't, I don't think it is...
[16:30.310 --> 16:36.770]  Like, for example, like who should, who should, who, the current, uh, just, because I worked at VA, I'm most familiar with VA.
[16:36.870 --> 16:37.970]  I'll use it as an example.
[16:38.250 --> 16:45.810]  The current secretary of VA, the person that Trump nominated, uh, to be the secretary of VA, is this guy, Doug Collins.
[16:46.250 --> 16:47.810]  He's like very funny, nice person.
[16:48.230 --> 16:51.730]  But, uh, you know, he's like a, he was like a former, he's, you know, obviously he's a veteran.
[16:51.730 --> 16:54.190]  He's like ex, uh, ex Navy or something like that.
[16:54.350 --> 16:54.670]  Chaplain.
[16:55.050 --> 16:59.410]  But veteran, uh, VA has, uh, 473,000 employees.
[16:59.710 --> 16:59.850]  Right.
[17:00.170 --> 17:06.190]  Um, so I would say you should, you know, you should probably like have a secretary who, you know, you're kind of like hiring a CEO for four years.
[17:06.270 --> 17:13.530]  You'd probably pick like, who's like the, you know, the CEO of like Kaiser Permanente, like the largest, like Western hemisphere, like, you know, integrated healthcare system.
[17:13.530 --> 17:15.350]  That person would probably be pretty good.
[17:15.450 --> 17:24.270]  But Obama or Biden, um, you know, they had, uh, this, uh, VP from Microsoft, some senior VP from Microsoft who was like, who is, you know, pretty high up in the, in the VA.
[17:24.390 --> 17:32.610]  He says, neither is the, the, the right choice is to, is, you know, you know, I, I think the word that gets thrown around is like meritocracy, right?
[17:32.690 --> 17:34.470]  Like you pick the best person for the job.
[17:34.810 --> 17:35.130]  Right.
[17:35.430 --> 17:35.730]  Yeah.
[17:35.790 --> 17:37.910]  I fully agree in that, but we don't live in a meritocracy.
[17:38.070 --> 17:43.130]  We live in a democracy, uh, you know, and, uh, and this is what people voted for.
[17:43.130 --> 17:47.370]  Or people voted, you know, for the Republican and the Republican is going to pick Republicans.
[17:47.530 --> 17:52.570]  And like, we, we all vote for a, uh, you know, for, for, for, for these two parties.
[17:52.590 --> 17:56.850]  And that's just like the, the system we, we, we live in, we live in a society, you know?
[17:56.990 --> 17:57.190]  All right.
[17:57.310 --> 17:57.750]  What are you going to do?
[17:58.250 --> 18:06.810]  Um, Doug Collins also, in addition to being secretary of the VA, he used to be a, a congressman, very loyalist part of the MAGA coalition.
[18:06.810 --> 18:13.390]  Quick question about, um, the, you, you started doing an audio diary of your time at DOGE.
[18:13.530 --> 18:18.290]  And what made you, we, we started talking about this shortly before, uh, getting up on the stage.
[18:18.590 --> 18:20.970]  What, what informed that decision?
[18:21.150 --> 18:22.270]  What'd you want to record?
[18:22.850 --> 18:23.370]  Yeah.
[18:23.390 --> 18:27.910]  I mean, I, I recorded this set of audio diaries, uh, for myself.
[18:27.910 --> 18:33.830]  Uh, and then it turns out NPR was interested in them, but I recorded them mostly cause I didn't know what DOGE was going to be.
[18:33.970 --> 18:36.690]  Like the, the details I had on DOGE were pretty sparse.
[18:36.890 --> 18:47.810]  And so going into it, I just thought it would be nice to have my own sort of like a time capsule of like what I sort of the, the possible set of distributions of like what this experience could look like.
[18:47.810 --> 18:52.210]  such that when, after the fact I could look back and be like, how accurate was I?
[18:52.330 --> 18:53.830]  Like, what was I super wrong about?
[18:53.990 --> 18:54.870]  What was I right about?
[18:55.510 --> 19:00.650]  Uh, and also as a way of like protecting myself to be like, this is what I really wanted to do.
[19:00.890 --> 19:07.130]  So like, you know, in case I, I, I don't know, someone thinks I sort of enabled some fascist, uh, to take over the world or something.
[19:07.150 --> 19:10.370]  Like I can say, Hey, look, I didn't really, that was not my goal.
[19:10.450 --> 19:11.650]  I wanted to just write some JavaScript.
[19:11.650 --> 19:18.050]  And so you have kind of a real time diary of, you know, day one, here's what happened.
[19:18.290 --> 19:18.730]  Here's.
[19:18.910 --> 19:19.390]  Yeah.
[19:19.570 --> 19:20.970]  And I haven't published those though.
[19:21.070 --> 19:29.470]  I think there's some edits on Planet Money, but I did use them to then like make sure that the article I wrote, DOGE Days, was accurate to my experience.
[19:29.930 --> 19:30.310]  Gotcha.
[19:30.930 --> 19:34.570]  Um, so start talking to me about your work at the VA.
[19:34.730 --> 19:37.390]  You talked about, uh, reductions in force.
[19:37.630 --> 19:40.410]  Uh, the, it was a plan for, for mass layoffs.
[19:40.410 --> 19:45.470]  So you originally, uh, started working on software known as a contract muncher.
[19:45.610 --> 19:48.170]  Uh, explain a little bit about that.
[19:48.470 --> 19:48.690]  Yeah.
[19:48.850 --> 19:54.450]  So, uh, DOGE, one of the, one of the, so DOGE is really a, you know, what is DOGE?
[19:54.550 --> 19:59.250]  Like DOGE is really a set of executive orders that Trump signed pretty early on in his second term.
[19:59.250 --> 20:11.690]  Uh, one of these contracts is like, I think called like cost efficiency initiative, uh, which basically sort of says every agency has to go review all of their contracts with every, you know, there's a lot of contracts.
[20:11.690 --> 20:19.530]  The VA has tens of thousands of contracts, um, and has to review them for efficiency and like any cost savings that can accrue to the American taxpayer.
[20:19.750 --> 20:22.190]  Like we, you know, every agency has to, has to do that.
[20:22.190 --> 20:26.990]  Uh, but these agencies don't have any capacity, like to, to do that.
[20:27.110 --> 20:28.550]  They can't hire anyone to do it.
[20:28.670 --> 20:33.710]  So DOGE is basically like this, like volunteer force of, of people who will like help the agencies do these things.
[20:33.790 --> 20:36.410]  So I wrote a script, uh, which is open-source.
[20:36.670 --> 20:37.810]  You can find it on the Internet.
[20:38.350 --> 20:40.970]  Uh, big fan of open-source software.
[20:40.970 --> 20:54.930]  Uh, and, and so I, you know, I contributed the contract muncher, which is a Python script that basically just goes through all the, all the contracts that the VA had and reviews them for, uh, possible termination.
[20:54.930 --> 21:03.630]  Which really just means a shortlist that then gets sent to humans to then review and nothing is getting cut because an AI said to cut it.
[21:03.870 --> 21:04.710]  That would be dumb.
[21:05.070 --> 21:06.810]  Um, it caused a bit of a stir.
[21:07.230 --> 21:08.510]  Uh, the AI typically does.
[21:09.130 --> 21:11.130]  The Wired magazine reached out to you.
[21:11.230 --> 21:14.990]  Can you, can you tell me about that, uh, that encounter?
[21:15.330 --> 21:15.630]  Yeah.
[21:15.690 --> 21:22.850]  I mean, I think, you know, I think the media serves a great purpose, which is like, you know, uh, to like hold people accountable to doing things.
[21:22.910 --> 21:26.930]  So I, I think there's nothing wrong with like the, the media trying to, to figure out what's going on.
[21:27.030 --> 21:31.290]  And sometimes I think they reported like a little bit one sided, but you know, obviously I'm going to say that.
[21:31.690 --> 21:40.430]  Uh, but yeah, the Wired wrote, wrote a piece about how, you know, like the VA and other agencies were, were sort of using AI to, to, to do the work faster.
[21:40.790 --> 21:45.490]  Um, and every, the thing is like DOGE, you know, you can think of it like a bunch of like mini cells.
[21:45.710 --> 21:47.270]  Like there wasn't like a clear central org.
[21:47.450 --> 21:50.950]  So I don't know how DOGE was necessarily acting in these other agencies.
[21:50.970 --> 21:54.570]  I can't speak to that, but at VA, I was the one writing all the code.
[21:54.570 --> 21:57.570]  So I can say that like, I talked to the humans.
[21:57.650 --> 22:03.530]  I talked to the contract people, the contracting officers they're called and said, Hey, you know, I know you guys are like reviewing all the contracts.
[22:03.970 --> 22:05.290]  Um, what are you doing?
[22:05.330 --> 22:06.830]  Like, what does that work actually look like?
[22:07.230 --> 22:17.310]  Uh, and I just sat with them and then basically wrote like a prompt with AI to then, you know, uh, sort of try to encapsulate their thinking in like three to 400 English words.
[22:17.310 --> 22:25.090]  Uh, and then I ran through the script, you know, on 10 PDFs and sort of help got their feedback, like how accurate do you think this is?
[22:25.190 --> 22:28.190]  And then we kind of refined it, um, until they felt like pretty good.
[22:28.290 --> 22:37.810]  And then I said, Hey, instead of having to review 90,000 contracts, you can review, uh, these 4,500 contracts that, you know, AI has just, you can think of it like a, a sort.
[22:37.970 --> 22:39.510]  It's kind of a new way of sorting, right?
[22:39.570 --> 22:42.010]  Like there's these sorting algorithms that exist in computer science.
[22:42.490 --> 22:45.330]  Uh, you know, like a bubble sort is like one kind of sort.
[22:45.330 --> 22:49.290]  Uh, you can think of like an AI as like a kind of an LLM sort.
[22:49.510 --> 22:58.910]  It's like a, a way of ranking, uh, the likelihood that a contract should get cut such that the humans can, you know, aren't wasting their time on like a bunch of stuff that obviously should not get cut.
[22:59.450 --> 22:59.850]  Right.
[22:59.930 --> 23:01.650]  Like healthcare, for example, should not get cut.
[23:01.870 --> 23:07.130]  So like, you don't have to review like, you know, two thirds of the contracts that are related to healthcare.
[23:07.670 --> 23:08.070]  Mm.
[23:08.990 --> 23:17.990]  One of the things that's striking from that Wired article is, uh, well, one thing, your response to the, request for comment, you, you were very cordial.
[23:18.090 --> 23:19.170]  You said, I'm a fan of your work.
[23:19.510 --> 23:22.830]  And you see throughout the article, it's a very interesting article.
[23:22.830 --> 23:29.090]  Cause you see kind of your immersion into the federal government.
[23:29.250 --> 23:31.410]  That a lot of things aren't as simple as it seems.
[23:31.570 --> 23:37.630]  There's this, uh, thing that the NPR picked up on, um, you changed Twitter to X.
[23:37.630 --> 23:38.070]  Yeah.
[23:38.070 --> 23:43.550]  There are all sorts of, uh, that can change how, uh, accessibility.
[23:44.150 --> 23:47.650]  It, it can, that simple things take a lot.
[23:47.770 --> 23:49.910]  There are lots of checks and balances in place.
[23:50.170 --> 23:51.070]  Yeah, totally.
[23:51.430 --> 23:52.810]  So tell me about that.
[23:52.830 --> 23:53.810]  They made me come off as very stupid.
[23:54.070 --> 23:56.150]  Well, what did you, what, what's that?
[23:56.330 --> 23:56.930]  Very stupid.
[23:57.090 --> 23:57.690]  They, they tried.
[23:58.750 --> 24:01.330]  They also called me an operative, which I think is a fascinating.
[24:02.910 --> 24:04.090]  What did they get wrong then?
[24:04.930 --> 24:06.750]  Um, well, I don't think they got anything wrong.
[24:06.950 --> 24:10.110]  I, I actually did a podcast interview with Vittoria from, from, from wired.
[24:10.170 --> 24:11.430]  I think she's, she's great.
[24:11.690 --> 24:24.850]  Um, but I think there, there is this sort of like, you know, uh, sort of bull in a China shop, sort of like, right, like sort of image that they want to create of the someone who walks in, like having zero clue what's going on as someone who's like never actually written code before.
[24:25.290 --> 24:29.970]  Um, but generally like, you know, I'm decently good at shipping software.
[24:29.970 --> 24:32.410]  And I, I, I, it's interesting.
[24:32.610 --> 24:37.330]  The, the impression that I got from the article, um, wasn't that, uh, you, I don't think...
[24:37.330 --> 24:41.010]  Well, things like, for example, like changing Twitter to X and there being accessibility issues.
[24:41.010 --> 24:42.850]  It's like a solvable problem.
[24:43.010 --> 24:43.270]  Right.
[24:43.390 --> 24:47.850]  Well, I mean, how would you have known that problem without having first worked for the federal government, right?
[24:47.990 --> 24:52.210]  I mean, it's a, I mean, I, I, anyone who makes websites should be aware of accessibility issues.
[24:52.370 --> 24:52.750]  Oh, sure.
[24:52.970 --> 24:53.070]  Sure.
[24:59.970 --> 25:00.450]  That's what you tell it to be.
[25:00.610 --> 25:00.850]  Mm.
[25:01.390 --> 25:06.730]  Uh, so what did you learn from that, that experience?
[25:07.010 --> 25:10.630]  I mean, there, what did you learn from being in the federal government for the first time?
[25:10.710 --> 25:12.770]  You had wanted to work for the federal government.
[25:13.330 --> 25:14.130]  Yeah, I learned a lot.
[25:14.230 --> 25:15.190]  I mean, that's why I did it.
[25:15.230 --> 25:18.090]  Like as, as, uh, MC said before, like we're here to learn.
[25:18.330 --> 25:22.810]  Uh, and I wanted to learn as much about how the federal government worked as possible.
[25:22.810 --> 25:23.670]  And I learned a lot.
[25:23.990 --> 25:27.250]  Um, you know, I learned, it's, it's hard, it's hard to like, you know, there's no mental model for it.
[25:27.310 --> 25:30.130]  It's not like I can give you like five ways the federal government works.
[25:30.230 --> 25:30.310]  Right.
[25:30.530 --> 25:38.190]  Uh, but I think I've generally learned that, like, there's a huge force in terms of like, all the things we want out of the federal government are happening.
[25:38.350 --> 25:39.570]  It is becoming more efficient.
[25:39.710 --> 25:40.410]  It's getting more modern.
[25:40.710 --> 25:42.450]  Uh, but it just takes a long time.
[25:42.630 --> 25:45.730]  Um, and just seeing like how certain decisions get made.
[25:46.030 --> 25:50.670]  Um, at the end of the day, it's just like a bunch of people in a room, not that different than a big tech company.
[25:51.130 --> 26:02.190]  Uh, I did learn that like, for example, like basically no software engineers work for the federal government, which is kind of a problem, uh, because it makes it really hard to ship good software when no software engineers work for you.
[26:02.310 --> 26:08.990]  It would be like outsourcing all of your software to a third-party, like Deloitte, which I thought was a toothpaste company before I joined the federal government.
[26:09.850 --> 26:13.930]  And obviously it makes sense why it's so expensive and the software is so shitty.
[26:14.470 --> 26:16.350]  It's like you're outsourcing it to McKinsey.
[26:16.350 --> 26:21.010]  Uh, just one, cause I see we're, we're running a little.
[26:21.150 --> 26:22.010]  We're running out of time already.
[26:22.250 --> 26:22.590]  No, no, no, no.
[26:22.690 --> 26:23.430]  We have, we have.
[26:23.550 --> 26:24.070]  We're on Q and A.
[26:24.190 --> 26:24.790]  Q and A would be awesome.
[26:24.790 --> 26:25.550]  Not for all the ground.
[26:25.550 --> 26:26.270]  I'm on to cover.
[26:26.510 --> 26:29.170]  Uh, the, the, um, the email.
[26:29.270 --> 26:29.710]  I'm here all day.
[26:29.850 --> 26:30.290]  I'm here all day.
[26:30.590 --> 26:30.830]  Good.
[26:30.990 --> 26:31.090]  Good.
[26:31.370 --> 26:32.350]  Tell me about the e-meeting.
[26:32.710 --> 26:32.990]  Yeah.
[26:33.050 --> 26:36.750]  The e-meeting, I mean, you could think of it like a Joe Rogan podcast, but Joe Rogan wasn't there.
[26:37.470 --> 26:46.670]  Uh, it was, uh, not as I was honestly hoping generally one of the, I guess you could call it a, a disappointment.
[26:46.770 --> 26:52.010]  Uh, so, you know, a journalist would take that and make, make, make, make it, make it look worse than it is.
[26:52.110 --> 27:05.670]  But basically I, I was expecting doge to be more like a startup, more like a central organization with a clear goal and all the vectors aligned in a specific direction and a way of breaking up this big goal into sub goals and, and making a lot of progress on it on a weekly basis.
[27:05.930 --> 27:09.630]  This sort of, what did you get done last week meme that, that Elon kind of created.
[27:09.910 --> 27:19.630]  Uh, but it really was just like a sort of a bunch of people who like, were just trying to make the agencies efficient in this very broad, undefined way.
[27:19.630 --> 27:25.490]  So when this e-meeting happened, I was hoping for this sort of alignment of like, this is what we're doing.
[27:25.710 --> 27:26.670]  Everyone's going to give an update.
[27:26.670 --> 27:28.330]  What did everyone get done last week?
[27:28.650 --> 27:29.330]  Okay, cool.
[27:29.550 --> 27:33.010]  Like, you know, knowledge, sharing feedback, like, et cetera, et cetera.
[27:33.370 --> 27:50.050]  Uh, and really it was just more of this kind of like informal, casual, really just like a way of like hanging out with Elon and like, sort of like hearing him opine on like, you know, how she, uh, the government is, uh, which, uh, was like a cool experience,
[27:50.150 --> 27:59.750]  I guess, like it's, you know, but like certainly was not as productive in terms of like accomplishing the, the thing, uh, and actually shipping software for the federal government.
[27:59.870 --> 28:02.410]  I think like Elon was not really involved with DOGE to be honest.
[28:02.510 --> 28:04.790]  Like he was really just kind of like a figurehead.
[28:05.170 --> 28:10.190]  Um, there were no, you know, group signal chats with, with anyone really in them like that.
[28:10.190 --> 28:14.410]  So, so that was the first time you had met him or, or spoken to him at DOGE?
[28:14.550 --> 28:14.710]  Yeah.
[28:14.770 --> 28:15.170]  Me personally.
[28:15.270 --> 28:15.330]  Yeah.
[28:15.470 --> 28:15.770]  Yeah.
[28:15.970 --> 28:19.910]  And that was, you told me that happened in, uh, the Eisenhower building.
[28:20.150 --> 28:20.410]  Yeah.
[28:20.630 --> 28:22.270]  So part of the white house complex.
[28:22.870 --> 28:23.110]  Yeah.
[28:23.290 --> 28:23.590]  Part two.
[28:23.650 --> 28:23.690]  Yeah.
[28:23.770 --> 28:26.310]  Technically part of the Eisenhower, the, the white house complex.
[28:26.470 --> 28:28.470]  I did get badged by the white house, which was cool.
[28:28.690 --> 28:28.930]  Wow.
[28:29.030 --> 28:40.210]  So the, uh, the, that was, um, I, a very long meeting, you know, like Joe Rogan podcasts or very long, this was a, tell me about how long exactly?
[28:41.450 --> 28:41.730]  Yeah.
[28:41.870 --> 28:45.210]  I mean, I think it started at like seven or 8 PM and ended at like midnight.
[28:45.430 --> 28:46.650]  And it was just him philosophizing.
[28:47.430 --> 28:49.690]  Well, it was like a back and forth, you know, it was kind of like all hands.
[28:49.810 --> 28:57.250]  So, you know, if you've worked at like a tech company or a company, like, you know, you have these all hands with like the leadership or the CEO of the company where people kind of like, you know, TGIF at Google, right?
[28:57.350 --> 28:59.050]  Where like people have questions, they answer them.
[28:59.150 --> 29:02.890]  And it's kind of like becomes this like bit of a sort of a fireside chat.
[29:03.290 --> 29:06.010]  And, uh, so it was a little bit like, a little bit like that.
[29:06.010 --> 29:10.010]  Well, just, I want to, uh, explore one thing.
[29:10.150 --> 29:15.450]  You talked about the power of, of Elon Musk that you said he was more of a figurehead.
[29:15.990 --> 29:18.350]  Uh, did you ever hear the name Amy Gleason?
[29:18.530 --> 29:18.750]  Yeah.
[29:18.930 --> 29:19.590]  I met her one time.
[29:19.730 --> 29:20.250]  Oh, you met her.
[29:20.370 --> 29:20.510]  Okay.
[29:20.550 --> 29:21.270]  What was the context?
[29:21.610 --> 29:21.930]  Yeah.
[29:22.030 --> 29:25.830]  So the context was I wanted to hire software engineers at VA.
[29:26.030 --> 29:31.710]  Cause I fundamentally, uh, think that, uh, by the way, people ask like, what, what do all the software engineers do that can't get jobs?
[29:31.710 --> 29:35.050]  Like go work for the U.S. federal government as software engineers would be awesome.
[29:35.390 --> 29:38.890]  Uh, they actually pay pretty well, uh, especially relative to not having a job.
[29:39.490 --> 29:42.990]  Uh, and, uh, and so I wanted to hire software engineers at VA.
[29:43.270 --> 29:46.950]  And so I went to, I was like, you know, I was like, Hey, can we hire software engineers at VA?
[29:47.050 --> 29:48.130]  And they're like, no, DOGE won't let us.
[29:48.210 --> 29:49.050]  I'm like, well, guess what?
[29:49.050 --> 29:50.270]  Uh, I am DOGE here.
[29:51.190 --> 29:53.790]  So I think we, we can, uh, you know?
[29:53.970 --> 29:59.190]  And so they're like, okay, uh, well, DOGE, like Amy Gleason, like, you know, you gotta get her like on board with this.
[29:59.290 --> 30:05.310]  And so, uh, I just emailed her and said, Hey, uh, you know, can I hire software engineers at VA?
[30:05.470 --> 30:06.210]  And she's like, of course.
[30:06.610 --> 30:18.870]  Uh, and, uh, the EO, the Trump EO actually says we can, it says specifically for every contract we cut, we can redirect 25% of the savings to hiring in house, actually increasing state capacity, which I think would be awesome.
[30:19.230 --> 30:28.130]  Uh, and so that's, that was like the, the meeting I went, I went with the CTO of VA, who's an awesome guy, Charles, uh, and, uh, who actually worked at USDS, the original one.
[30:28.390 --> 30:31.470]  Uh, a lot of the CTOs at these different agencies kind of came from that.
[30:31.770 --> 30:34.050]  And, uh, and we just asked her, like, can we hire people?
[30:34.150 --> 30:34.950]  And she said, obviously.
[30:35.730 --> 30:37.270]  Uh, and that was the meeting.
[30:37.830 --> 30:41.410]  What was your understanding of who Amy Gleason was in the chain?
[30:42.270 --> 30:46.550]  Um, well, she was like the acting head of USDS, right?
[30:46.710 --> 30:51.130]  So the, the, the, the EOs explicitly sort of refer to that position.
[30:51.270 --> 30:56.730]  Uh, but she wasn't really involved in like any of the DOGE, like the sort of Trump, DOGE, Elon stuff.
[30:57.330 --> 31:15.070]  One of the reasons that I ask, there's, um, you know, my background, I mentioned I'm a legal journalist and there are a couple of live, uh, cases going on right now about that that go to the appointments clause, which is basically, uh, the, the constitutional mandates of how an agency is created, how someone heads an agency.
[31:17.710 --> 31:24.730]  And a question that has come up and legal journalists have asked this question over and over again, who is the administrator of DOGE?
[31:24.730 --> 31:30.630]  and who during your time would have been the boss?
[31:30.950 --> 31:32.170]  Well, there's no boss.
[31:32.390 --> 31:41.490]  I mean, you, you are, I was working as senior advisor, chief of staff at NVA and I was asked to help the agency execute the EOs that Trump signed.
[31:41.850 --> 31:48.070]  Uh, but your understanding of who you reported to like, well, but what is reported?
[31:48.430 --> 31:49.090]  Like, what does that mean?
[31:49.370 --> 31:50.990]  Like, no one, no, no, no, no one.
[31:51.130 --> 31:53.910]  Like I'm not like the federal government is, is very interesting.
[31:53.930 --> 32:01.970]  Cause like in a typical company, you have like a manager, but in the, in the kind of the context of, of the U.S. federal government, it's actually incredibly flat.
[32:02.270 --> 32:05.030]  Like there's because it really Congress defines statutory roles.
[32:05.370 --> 32:08.010]  Um, but I was an advisor to the chief of staff.
[32:08.010 --> 32:08.770]  Uh, right.
[32:08.870 --> 32:15.210]  So like, if, if I would, you know, if, if, if I had a boss, it was the, it was the chief of staff of, of the department of veterans affairs.
[32:15.270 --> 32:23.210]  Like if I, if he wanted to, he w if he said, you can't do this or you should do this, like that's my job to help him run the org.
[32:23.530 --> 32:23.630]  Right.
[32:23.850 --> 32:24.310]  Got it.
[32:24.510 --> 32:25.550]  But, and he was nominated.
[32:25.730 --> 32:26.830]  I believe it pointed by Trump.
[32:27.350 --> 32:27.710]  Right.
[32:27.990 --> 32:41.910]  So I think, you know, what you're saying and what is often, uh, the kind of live issue being litigated here, uh, in a couple of these cases is there's a presentation of dOGE of, oh, we're just advisors.
[32:42.210 --> 32:47.810]  Then there's the perception of Joe Doge that Elon Musk was in the white house.
[32:47.810 --> 32:51.450]  He tweets out CFPB, RIP.
[32:51.850 --> 32:56.990]  And then the consumer financial protection bureau, uh, is, is they try to shut it down.
[32:56.990 --> 32:58.890]  And it's a kind of question of...
[32:58.890 --> 33:01.170]  I think he was just being a whistleblower in that context.
[33:01.450 --> 33:01.850]  What's that?
[33:02.010 --> 33:02.930]  He's just a whistleblower.
[33:03.590 --> 33:04.810]  That Elon Musk is.
[33:04.830 --> 33:06.930]  Well, if he's saying it's CFPB, RIP, right?
[33:07.010 --> 33:10.070]  He's just sharing private information with the public.
[33:10.390 --> 33:17.150]  That he wasn't, uh, at that point, you're saying that he wasn't having a role in shutting it down?
[33:17.550 --> 33:17.790]  No, correct.
[33:17.970 --> 33:18.110]  Yeah.
[33:18.350 --> 33:20.470]  That he made the, he had a similar comment.
[33:20.470 --> 33:25.610]  I'm trying to, um, he said with, uh, USAID, we're in the process.
[33:25.770 --> 33:28.790]  We're in the process of shutting down USAID.
[33:28.910 --> 33:29.010]  Yeah.
[33:29.070 --> 33:32.890]  The Royal, we like Trump, Trump and Marco Rubio and Russ Vaught.
[33:33.210 --> 33:40.870]  So, so from your vantage point, this was, um, just Elon Musk informing the public.
[33:41.050 --> 33:45.330]  I mean, you know, in political, like political theory, you have like the concept of scapegoats, right?
[33:45.410 --> 33:47.110]  Like Rene Girard likes to talk about.
[33:47.110 --> 33:57.210]  Uh, and, uh, I think DOGE is like the perfect scapegoat because like at the end of the day, like who's in charge, like Trump and, uh, the secretaries of each agency.
[33:57.490 --> 34:04.030]  And if DOGE, like, you know, the meme where like someone's like getting like, you know, all these arrows are hitting somebody and like, there's someone like defending them.
[34:04.510 --> 34:06.810]  Uh, that's like Elon, like taking all the, all the heat.
[34:06.990 --> 34:07.090]  Right.
[34:07.290 --> 34:09.750]  Um, I don't know why he, he was so interested in doing that.
[34:09.970 --> 34:12.550]  Uh, that's just, it's kind of like what goes back to why he called it DOGE.
[34:12.550 --> 34:20.890]  I think he's, he likes, uh, being the center of attention and he likes like the, the, the, sort of like the conflict of, of that, uh, which most people do not like.
[34:21.430 --> 34:25.610]  Um, and, uh, but yeah, I think, I think USAID was going to get shut down no matter what.
[34:25.610 --> 34:29.250]  Um, and, uh, you know, part of project 2025 or whatnot.
[34:29.510 --> 34:33.790]  Uh, and DOGE, uh, DOGE was happy to take the, the credit.
[34:33.970 --> 34:35.910]  Uh, and Trump was obviously happy to let them.
[34:36.490 --> 34:40.430]  Uh, so it kind of, I think genius, like political strategy by the Trump admin.
[34:41.330 --> 34:43.650]  And by the media for, by the way, supporting it completely.
[34:44.390 --> 34:44.830]  Yeah.
[34:44.830 --> 34:52.490]  They've done a very good job of like making DOGE feel super powerful, even though we have literally zero, zero authority to do anything besides be an advisory body.
[34:53.050 --> 34:54.650]  Which is why the lawsuits won't go anywhere.
[34:54.810 --> 35:00.730]  Why none of the people doing the lawsuits have ever asked me, who was on the record of saying all these things to like do an affidavit or anything.
[35:01.310 --> 35:02.190]  No one really cares.
[35:02.510 --> 35:04.210]  No one really believes that there's a problem here.
[35:05.470 --> 35:09.770]  And I mean, there is a problem, but like the people in power, aren't really willing to do much about it.
[35:10.230 --> 35:10.470]  Mm.
[35:10.950 --> 35:18.530]  Uh, tell me about the, uh, you're going back to kind of the contract muncher and all that.
[35:18.530 --> 35:29.950]  You told, uh, NPR, uh, that, uh, I hope we didn't break things in terms of, um, folks being scared about losing their jobs and things like that.
[35:30.230 --> 35:35.210]  With benefit of hindsight, do you think that, uh, any of that actually occurred?
[35:36.610 --> 35:37.890]  Uh, I don't know.
[35:38.070 --> 35:41.950]  I mean, uh, I'm, I'm super public with like my email and, and stuff like that.
[35:41.950 --> 35:46.750]  So people can like email me if they like want to tell me anything that's happened so far.
[35:46.810 --> 35:48.810]  And that hasn't really come back to me though.
[35:49.190 --> 35:52.490]  Do you keep in touch with folks from the federal government who you've met?
[35:53.110 --> 35:53.230]  Yeah, totally.
[35:53.450 --> 35:54.130]  What's that been like?
[35:54.350 --> 35:54.710]  Yeah.
[35:55.070 --> 35:57.790]  Uh, I mean, they're basically like, you know, this happens every four years.
[35:58.750 --> 36:00.870]  Like, yeah, this is, you know, nothing new.
[36:00.950 --> 36:04.630]  This is like, you know, what's been happening with the U.S. federal government for a long time.
[36:04.770 --> 36:06.990]  Like it just goes through crazy whiplash all the time.
[36:07.490 --> 36:08.630]  Everyone thinks they can change it.
[36:08.710 --> 36:09.150]  No one can.
[36:09.150 --> 36:11.050]  Uh, the deep state wins again.
[36:12.710 --> 36:17.290]  The U.S., uh, U.S. versus, uh, DOGE here.
[36:17.510 --> 36:23.330]  What do you think your experience would have been like had your application to join the U.S. digital service?
[36:23.430 --> 36:25.070]  How would that have been different?
[36:25.470 --> 36:25.890]  Yeah.
[36:25.950 --> 36:29.150]  Well, I, I actually talked, uh, to quite a few USDS people.
[36:29.450 --> 36:33.730]  Uh, you know, they did, some of them reached out like after the, after I wrote my article.
[36:34.010 --> 36:37.950]  Um, and they said it's sort of, I think broadly, broadly two things.
[36:37.950 --> 36:43.170]  One, I think it would have been, uh, uh, worse in the sense that like, we wouldn't have been able to do anything.
[36:43.510 --> 36:53.310]  Like, I think that, that, that USDS 1.0 had such, like the, the political class was so uninterested in letting hackers into the room.
[36:53.310 --> 37:02.070]  That like, DOGE was actually like, the fact that like, Elon, the Aura, all of this stuff helped software engineers be in the room when these decisions were being made.
[37:02.170 --> 37:12.150]  Like, I got to talk to the Deloitte people and the IBM people and the Accenture people as a software engineer and like, really drill down on like, on their cost structure.
[37:12.150 --> 37:13.310]  And I think that's great.
[37:13.830 --> 37:14.630]  I think that's a really good thing.
[37:15.170 --> 37:17.910]  Uh, USDS 1.0 had basically no power.
[37:18.050 --> 37:24.110]  Like the, the, the, the, you know, software engineers have like very low credibility in the federal government, which is unfortunate because they should have the most.
[37:24.450 --> 37:28.630]  That's really just because all the lawyers are scared that they're going to be replaced by software engineers at some point in the future, which will happen.
[37:28.630 --> 37:32.890]  I see that you want to go to, uh, questions from, from folks.
[37:32.970 --> 37:40.290]  But before that, I just want to ask, you had told NPR that you were surprised by how efficient the federal government actually was.
[37:40.610 --> 37:51.390]  If you had been, so a kind of two part question, if you were entering DOGE again for the first time, knowing what you know now, how might you have approached it differently?
[37:51.570 --> 37:52.650]  Or would you have joined at all?
[37:53.630 --> 37:58.610]  I mean, I'm happy to join again, like in two and a half, four, three years, whenever there's a new cycle.
[37:58.610 --> 38:01.590]  Um, I'd love to do this every four years for the rest of my life.
[38:01.690 --> 38:06.430]  Cause I think there's a huge amount of value to be had by just like trying to make the system better slowly.
[38:06.830 --> 38:09.290]  Um, as much as it will allow you to do it.
[38:09.370 --> 38:13.090]  But yeah, generally I would say that the, the, the, the government was more efficient than I was expecting.
[38:13.210 --> 38:16.130]  I think I, I've sort of spent a lot of time in software at startups, et cetera.
[38:16.730 --> 38:18.990]  Uh, startups are like the least efficient industry on earth.
[38:19.390 --> 38:21.530]  Uh, and that's mostly cause we're so high leverage.
[38:21.750 --> 38:29.470]  We're so successful and investors are willing to like sort of inflate our stock value by so much money that we have like free cash to, to blow basically.
[38:29.690 --> 38:35.330]  So you go into the average startup office and it's a bunch of really nice laptops and bean bags and, and all sorts of really cool stuff.
[38:35.710 --> 38:38.110]  Um, and you're like, what do you guys do here?
[38:38.570 --> 38:46.570]  You know, that you ask that at Google and Facebook and like, you know, but, but it pays off cause, uh, you know, cause the, the market caps of these companies are crazy.
[38:46.770 --> 38:48.650]  Cause a lot of people use Google.com, right.
[38:48.710 --> 38:49.990]  And Google maps and things like that.
[38:50.410 --> 38:52.410]  Um, so I think that was the bias I came into.
[38:52.410 --> 38:55.530]  It was like expecting the federal government to look a little bit more like a tech company.
[38:55.890 --> 39:02.130]  Uh, and actually the U.S. federal government is relatively efficiently run because in a tech company, you can just hire people cause you want to.
[39:02.410 --> 39:05.270]  In the federal government, the federal government will only hire if it's required.
[39:05.410 --> 39:13.750]  So the only way to really replace and like make the government more efficient would be to do what I said, which is to hire software engineers, requires hiring people, requires building stuff.
[39:13.750 --> 39:23.770]  And then once the stuff is built and doing things and people, then you may need less people to then, you know, fulfill the same service that the U.S. government is statutorily required to do by Congress.
[39:24.210 --> 39:25.690]  But you can't really flip the two.
[39:25.830 --> 39:32.250]  You can't first cut a bunch of stuff and then build a bunch of stuff to replace that.
[39:32.570 --> 39:38.990]  That's something, you know, I would definitely want to aim more towards, which is always what I tried to do.
[39:38.990 --> 39:43.730]  And I learned over time that, you know, for example, if you can't hire Democrats, like, guess what?
[39:43.870 --> 39:45.090]  Most software engineers are Democrats.
[39:45.090 --> 39:47.690]  So, like, good luck hiring, like, a lot of software engineers, right?
[39:48.010 --> 39:57.390]  So, yeah, there's, you know, there's a lot of reasons we can get into on why, like, the federal government is so bad at software.
[39:59.070 --> 40:12.230]  But that's, yeah, I definitely want to spend more time on the kind of the systemic issues and the sort of congressional changes that might need to happen than maybe just, like, trying to change 450 agencies at the same time.
[40:12.330 --> 40:14.490]  And I think more transparency, like, I open-sourced my code.
[40:14.590 --> 40:18.310]  That was one of the things I asked Elon when I met him.
[40:19.050 --> 40:20.870]  I said, can I open-source my code?
[40:20.930 --> 40:21.790]  And he said, sure.
[40:22.170 --> 40:22.790]  So I did.
[40:23.510 --> 40:29.090]  I'm mindful of the need to go to viewer and audience questions, but just one last thing based on what you said.
[40:29.090 --> 40:42.670]  I want to get to the, what you, you stopped working at DOGE, I think, less than two months in after they essentially cut off your credentials because you responded to a request for an interview.
[40:42.850 --> 40:44.070]  And it was with Fast Company.
[40:44.870 --> 40:55.370]  And that, and do you think that the way that that went down, you've talked about transparency quite a bit.
[40:55.890 --> 40:59.150]  They unceremoniously sack you.
[41:00.310 --> 41:00.830]  Sure.
[41:01.150 --> 41:03.130]  After you reach out and...
[41:03.130 --> 41:03.990]  I got riffed, I got riffed, I know.
[41:03.990 --> 41:06.330]  After you get a press inquiry and you talk to the press.
[41:06.570 --> 41:09.010]  And without a word, your credentials are yanked.
[41:10.510 --> 41:11.770]  Yeah, it's kind of unfortunate.
[41:12.410 --> 41:13.990]  I thought I was being pretty productive.
[41:15.430 --> 41:18.790]  And, yeah, kind of ironic that the DOGE person got DOGEd.
[41:20.850 --> 41:21.970]  But also kind of funny.
[41:22.570 --> 41:25.970]  So, you know, the most entertaining outcome is the most likely, that sort of thing.
[41:26.090 --> 41:31.170]  But, yeah, I mean, it's... I was hoping that DOGE would be more transparent.
[41:31.870 --> 41:37.190]  That was kind of, like, one of the reasons I joined is that it was marketed as, like, a sort of hyper-transparent thing.
[41:38.870 --> 41:41.610]  And it, unfortunately, was not that transparent.
[41:41.710 --> 41:44.070]  And that's certainly disappointing.
[41:44.070 --> 41:51.930]  And I would encourage the next people who, you know, take over USDS and DOGE, et cetera, like, actually commit to making it more transparent.
[41:52.070 --> 41:54.470]  Because I think more transparency is sort of apolitically a good thing.
[41:55.490 --> 42:02.130]  I think it's hard to argue that, you know, I don't think anyone is arguing that the U.S. federal government should be more opaque.
[42:03.950 --> 42:08.510]  So, and it's not the right amount of opaque, so certainly it should be more transparent.
[42:09.050 --> 42:17.330]  And the project that you didn't get to work on next was processing software for facilitating claims processing, correct?
[42:17.430 --> 42:18.890]  Yeah, that's what I really wanted to work on.
[42:18.970 --> 42:24.930]  It's kind of crazy how much the VA spends on claims processing, which is effectively just, like, scanning PDFs.
[42:25.550 --> 42:31.490]  But we, they give IBM and Boozell and Hamilton, like, $200 million a year to process PDFs and stuff.
[42:33.130 --> 42:37.870]  And I thought we should just do it ourselves and save $200 million a year.
[42:38.830 --> 42:48.690]  We're going to switch to a Q&A, and we're just going to extend the courtesy to our virtual followers to have them ask a few questions, which Kyle will give to the two panelists.
[42:49.110 --> 42:52.590]  If you were leaving, just be aware there's a little water spill out in the hallway.
[42:52.590 --> 42:58.210]  You might just want to leave through the exit right there on the left and avoid slipping and sliding away on your first day.
[42:59.590 --> 43:05.290]  And after that, if we're lucky, we're going to get a few questions asked through this microphone over here to your right.
[43:05.870 --> 43:06.990]  And let's start.
[43:07.210 --> 43:08.030]  Go ahead, Kyle, please.
[43:08.570 --> 43:12.470]  Yeah, one of the first questions from our virtual attendees is as follows.
[43:13.070 --> 43:19.110]  This was Trump's second term, and the sort of person Elon is was well known.
[43:19.670 --> 43:30.070]  What did you... why did you decide to work for them specifically, especially in an organization that was clearly aiming to asset strip and data mine the government?
[43:30.390 --> 43:30.790]  What was that?
[43:31.450 --> 43:32.230]  Acid strip?
[43:32.510 --> 43:33.390]  Asset strip.
[43:33.650 --> 43:34.030]  Asset strip.
[43:34.530 --> 43:37.570]  I mean, in one word, hope.
[43:39.170 --> 43:40.750]  I think that's why we do what we do.
[43:40.930 --> 43:41.890]  I was hopeful.
[43:42.670 --> 43:42.890]  Okay.
[43:42.890 --> 43:46.270]  Some might call it naively optimistic, but I still have hope.
[43:46.790 --> 43:47.110]  So...
[43:47.770 --> 43:48.530]  Yeah, hope.
[43:49.710 --> 43:50.350]  Go ahead.
[43:50.910 --> 43:52.250]  Was that the whole question?
[43:52.370 --> 43:53.330]  There are several others.
[43:53.610 --> 43:54.850]  I can switch off.
[43:55.190 --> 43:58.130]  I'll read the second one that we have here.
[43:58.670 --> 44:13.310]  Given how sycophantic AIs can be, if you were LLM ranking contracts, how are you deciding what was wasteful given the administration's clear desire to nuke Medicaid from orbit?
[44:16.610 --> 44:17.310]  Um...
[44:18.010 --> 44:18.710]  We...
[44:18.710 --> 44:21.670]  I mean, we told the AI, do not cut any healthcare.
[44:23.430 --> 44:24.130]  So...
[44:24.130 --> 44:24.480]  Yeah.
[44:26.530 --> 44:27.230]  Okay.
[44:27.530 --> 44:29.390]  We did not cut any healthcare.
[44:32.350 --> 44:32.630]  Yes?
[44:33.430 --> 44:33.630]  Do...
[44:33.630 --> 44:33.850]  Okay.
[44:33.990 --> 44:35.330]  Well, I think we'll leave it at two for that.
[44:35.330 --> 44:35.930]  I just mean VA, though.
[44:36.050 --> 44:37.490]  I don't know what other agencies were doing.
[44:38.510 --> 44:38.790]  Okay.
[44:39.110 --> 44:39.390]  Um...
[44:39.390 --> 44:41.230]  We have several more, but we can take some live questions here.
[44:41.330 --> 44:41.410]  Okay.
[44:41.430 --> 44:42.390]  Let's go for a live question.
[44:42.510 --> 44:42.950]  Please go ahead.
[45:00.570 --> 45:00.930]  Yeah.
[45:01.270 --> 45:05.550]  I mean, I wasn't, uh, familiar, uh, with that decision.
[45:06.650 --> 45:07.010]  Um...
[45:07.010 --> 45:12.010]  But yeah, I certainly think we should hire more software engineers for the federal government, including a lot of the people who worked at AT&F.
[45:12.330 --> 45:12.550]  Um...
[45:12.550 --> 45:14.470]  But yeah, I don't, I don't know why that decision was made.
[45:15.030 --> 45:15.390]  Uh...
[45:15.390 --> 45:16.190]  I don't agree with it.
[45:18.230 --> 45:19.910]  And actually, they shipped a lot of really cool software.
[45:20.370 --> 45:20.610]  Yeah.
[45:22.190 --> 45:22.990]  Go ahead, Ruben.
[45:23.630 --> 45:23.870]  Yeah.
[45:24.470 --> 45:24.830]  Um...
[45:24.830 --> 45:34.590]  So I got from your, uh, description that one of the main contributors uh, you personally worked on, was this script, uh, that would help sort the contract?
[45:34.790 --> 45:35.590]  Is that, uh...
[45:35.590 --> 45:35.750]  Did I get...
[45:35.750 --> 45:37.050]  Yeah, that's one of the things I worked on, yeah.
[45:37.490 --> 45:37.630]  Right.
[45:38.070 --> 45:39.590]  Did you also write the criteria in it?
[45:40.690 --> 45:41.090]  Yeah.
[45:41.550 --> 45:41.950]  Yeah.
[45:42.130 --> 45:43.350]  With, with contracting officers.
[45:43.610 --> 45:45.470]  So, like, with career employees, uh...
[45:45.470 --> 45:50.410]  Did the agency help, uh, write the criteria, or was that your, your work?
[45:50.550 --> 45:52.070]  Do you have a background in social...
[45:52.070 --> 46:05.850]  I mean, it's like, I'm typing on the keyboard, but the human contracting officers are, you know, like, imagine, like, this script on the wall, and like, there's a bunch of contracting officers, like, helping me write the script, and then we test it on contracts.
[46:06.210 --> 46:11.050]  But the AI is not, like, to be clear, like, the AI, like, the AI is not doing any contract cutting, right?
[46:11.150 --> 46:14.090]  The humans are the ones then cutting the contracts.
[46:14.630 --> 46:19.270]  Uh, and there's like a long ten-step process that every contract has to get through to, to get cut.
[46:19.270 --> 46:22.810]  Um, it's not like I'm, I can't do anything.
[46:22.970 --> 46:25.270]  I'm not allowed legally or anything to cut any contract.
[46:25.290 --> 46:26.030]  No, I, I got that.
[46:26.170 --> 46:33.850]  But the, the criteria for sorting, which is a social policy criteria, I was just wanting to clarify who came up with that.
[46:34.010 --> 46:38.330]  If it was the agency, or if it was DOGE, if it was Elon, who, who came up with that?
[46:38.510 --> 46:38.910]  Got it.
[46:39.030 --> 46:39.610]  Yeah, the agency.
[46:40.490 --> 46:41.050]  Okay, thank you.
[46:43.790 --> 46:46.870]  Uh, thank you very much for, uh, this, uh, interview.
[46:47.070 --> 46:50.170]  Uh, I want to push back just a little bit on something you said before.
[46:50.490 --> 46:50.930]  Uh, please do.
[46:51.210 --> 46:51.310]  Question.
[46:51.490 --> 46:51.970]  That's great.
[46:52.130 --> 47:02.090]  Uh, I actually worked as a contractor, direct contractor with the federal government, not through booze or any of those, uh, for 21 years, uh, through Republican and Democratic administrations.
[47:02.290 --> 47:06.770]  I never faced any kind of, like, loyalty tests or any kind of questions about my...
[47:06.770 --> 47:10.270]  Well, yeah, the loyalty is about the employees of the federal government, not contractors.
[47:10.270 --> 47:11.110]  All right.
[47:11.350 --> 47:18.350]  Actually, it's one of the reasons contractors, like, to go into the systemic reasons of why contractors are great, is they don't have to go through a lot of the bullshit that federal government employees have to go through.
[47:18.850 --> 47:24.990]  But working with a lot of the employees, you know, we had people fighting over elections and stuff like that.
[47:25.190 --> 47:27.510]  So, anyway, but that's just my experience.
[47:27.790 --> 47:42.090]  The question I had was, uh, one of the things that I think a lot of people have been very concerned about is the amount of, uh, information that has been allowed by the courts, uh, to be given to, uh, DOGE and to the federal government.
[47:42.230 --> 48:01.230]  Do you have any concerns about, uh, people being targeted, uh, based on things like by administrations, maybe this, maybe other administrations, based on things that they, um, did financially, for example, donations to certain charities, um, and certain views,
[48:01.430 --> 48:05.570]  um, that they may hold people outside of the federal government.
[48:05.570 --> 48:25.370]  But if you're exposing, uh, complete IRS records to federal government, uh, ostensibly to fight fraud, but you're also telling them, like, who they're donating and who they're, uh, supporting, uh, and that that may have some negative ramifications to those individuals down the road.
[48:25.790 --> 48:28.590]  Yeah, I mean, I'm, I'm a big fan of sort of, like, zero trust.
[48:28.710 --> 48:34.010]  Like, I think when you ask for data, you should give reasons for asking for the data, and there should be clear logs of why you're asking for the data.
[48:34.010 --> 48:36.530]  And if you ask for the data and you don't need it, like, you should get fired.
[48:37.110 --> 48:38.870]  Uh, this is how most countries operate.
[48:39.010 --> 48:44.990]  Like, Estonia kind of famously has, like, this model around government data, um, citizen data, et cetera.
[48:45.310 --> 48:52.090]  Um, I do think, like, there's a lot of value to, like, you know, taking two data sources and connecting them to, like, find some insight around fraud, for example.
[48:52.570 --> 48:54.430]  So I think you want to be able to do that to some degree.
[48:54.890 --> 49:05.450]  Uh, but of course, the way you do that, you want to be very good about, you know, this is why I like open-source software, personally, is because I think that there's a lot of eyeballs on it, and you can build, like, a really good, resilient, maybe zero-knowledge system.
[49:05.910 --> 49:09.310]  Uh, uh, you know, so I'm...
[49:09.310 --> 49:13.770]  But I think there's sort of... there are good ways to do it and bad ways to do it.
[49:13.970 --> 49:16.910]  You know, there's good ways to have the government do stuff.
[49:16.970 --> 49:18.330]  There's good ways to have contractors do stuff.
[49:18.390 --> 49:20.790]  There's also bad things the government can do, bad things contractors can do.
[49:20.790 --> 49:24.130]  So at the end of the day, you want to just do a good job, uh, no matter what.
[49:24.350 --> 49:25.570]  Um, so yeah, I think there are concerns.
[49:25.670 --> 49:27.990]  I think that's why, like, the journalism is great.
[49:28.210 --> 49:28.950]  Lawsuits are great.
[49:29.770 --> 49:30.550]  Pushback is great.
[49:30.810 --> 49:32.150]  You know, open-source code is great.
[49:32.290 --> 49:33.930]  Like, we want... we all want to do the right thing.
[49:34.090 --> 49:38.630]  And, uh, the more people, like, like, push back and, you know, the better the result will be.
[49:38.850 --> 49:38.930]  Right?
[49:39.030 --> 49:39.730]  I think that's...
[49:39.730 --> 49:40.370]  Thank you.
[49:40.610 --> 49:40.930]  Really good.
[49:41.350 --> 49:41.770]  Thank you.
[49:43.810 --> 49:44.390]  Hi there.
[49:44.590 --> 49:53.030]  So, I'm just curious, because when DOGE came out, I was familiar with the Government Accountability Office existing, doing auditing and stuff of that nature.
[49:54.050 --> 49:58.150]  How much of DOGE's mission interlaps with GAO's mission?
[49:58.510 --> 50:00.470]  And did you have any kind of interaction?
[50:00.770 --> 50:02.930]  And what do those types of interactions look like?
[50:03.190 --> 50:06.050]  I did talk to the GAO afterwards.
[50:06.830 --> 50:08.970]  Like, they read my article and were like, hey, we should chat.
[50:09.130 --> 50:10.810]  We're both interested in the same kind of stuff.
[50:11.730 --> 50:13.690]  GAO is very, like...
[50:14.230 --> 50:14.770]  It's great.
[50:15.050 --> 50:16.130]  I mean, they do a lot of cool stuff.
[50:16.470 --> 50:24.890]  And the first thing that everyone does when they join DOGE is to go read the GAO reports on the agency, because, obviously, you know, they've done a lot of the work.
[50:25.110 --> 50:29.710]  But the writing the report and then executing the report is, you know, quite different.
[50:30.130 --> 50:39.370]  So, yeah, I think there's tons of overlap with GAO and DOGE and, you know, lots of organizations want the same outcome, which is like a more efficient, better...
[50:40.690 --> 50:41.290]  federal government.
[50:41.970 --> 50:42.410]  Got it.
[50:42.530 --> 50:46.710]  So, what you're saying is kind of like GAO is laying the foundation and then the DOGE is the value add?
[50:46.850 --> 50:48.350]  Is that kind of like what you're trying to get at?
[50:48.470 --> 50:52.250]  I mean, I don't think they're like, you know, like they're colluding in any way.
[50:52.430 --> 50:54.890]  They're, you know, but they have like the same mission.
[50:55.850 --> 51:00.370]  GAO takes it from sort of like a congressional view and DOGE is kind of like a White House agent.
[51:00.930 --> 51:02.650]  Like, it's not really even an agency.
[51:02.870 --> 51:04.310]  It shouldn't really be called a department.
[51:04.590 --> 51:12.670]  It's really an office within the White House, which is like really tiny, you know, like four people kind of thing.
[51:12.830 --> 51:14.170]  So, we are at time.
[51:14.990 --> 51:20.470]  Sahil has generously offered to stick around and not get into a rocket powered Tesla and disappear.
[51:20.730 --> 51:23.070]  So, I would suggest going, I'd say, upstairs.
[51:23.330 --> 51:24.290]  I did take the subway here.
[51:24.530 --> 51:24.690]  Yeah.
[51:24.690 --> 51:29.670]  So, go upstairs and some, maybe we'll have some follow-ups from people who want to discuss it.
[51:29.990 --> 51:31.750]  Thanks again for everyone for coming.
[51:31.970 --> 51:33.010]  Thanks again to the newcomers.
[51:33.190 --> 51:33.710]  Thanks for listening.
[51:34.310 --> 51:35.710]  Thanks to our two panelists.
[51:36.270 --> 51:37.550]  And being open-minded.
[51:38.650 --> 51:40.590]  See you all very soon.