WEBVTT - The Story: Spies vs. AIs

0:00:14.040 --> 0:00:16.560
<v Speaker 1>Welcome to tech Stuff. This is the story. I'm as

0:00:16.640 --> 0:00:20.919
<v Speaker 1>Volosen here with Cara Price. Hi as Hi Cara. So

0:00:20.960 --> 0:00:24.200
<v Speaker 1>I've been very, very excited about today's story. It's a

0:00:24.239 --> 0:00:26.960
<v Speaker 1>deep dive with the author of a column that has

0:00:27.000 --> 0:00:32.040
<v Speaker 1>the headline A band of innovators reimagines the spy game

0:00:32.479 --> 0:00:35.640
<v Speaker 1>for a world with no cover. Here's the author.

0:00:36.040 --> 0:00:41.560
<v Speaker 2>It began to realize that the future of intelligence was

0:00:41.600 --> 0:00:43.720
<v Speaker 2>going to be written in zeros and ones, that it

0:00:43.760 --> 0:00:47.560
<v Speaker 2>was going to become a technology war an algorithm.

0:00:48.120 --> 0:00:50.680
<v Speaker 1>So the story is by David Ignatius. He's a journalist

0:00:50.680 --> 0:00:53.639
<v Speaker 1>at the Washington Post with the reputation of knowing the

0:00:53.720 --> 0:00:58.280
<v Speaker 1>inside workings of the CIA better than almost anyone else

0:00:58.520 --> 0:01:00.600
<v Speaker 1>in the world. One of the things I love about

0:01:00.600 --> 0:01:02.840
<v Speaker 1>doing tech stuff is that we get to look under

0:01:02.840 --> 0:01:07.480
<v Speaker 1>the hood of how tech is revolutionizing places that you

0:01:07.600 --> 0:01:11.679
<v Speaker 1>might not expect, in some cases the most unexpected places,

0:01:12.080 --> 0:01:15.600
<v Speaker 1>one of which is, as we insiders like to call it,

0:01:15.640 --> 0:01:19.520
<v Speaker 1>the agency, the CIA. And what's interesting about David's piece

0:01:19.880 --> 0:01:23.440
<v Speaker 1>is that it's clear spycraft itself. Is it an inflection

0:01:23.560 --> 0:01:24.840
<v Speaker 1>point because of AI?

0:01:25.240 --> 0:01:26.920
<v Speaker 3>So tell me a little bit more about why this

0:01:26.959 --> 0:01:27.760
<v Speaker 3>cut your interest.

0:01:28.040 --> 0:01:30.520
<v Speaker 1>Well, I grew up as a boy in Britain and

0:01:30.560 --> 0:01:33.720
<v Speaker 1>have become a British Man. So of course James Bond

0:01:33.760 --> 0:01:38.200
<v Speaker 1>fantasies are totally organic to who I became as a person.

0:01:38.680 --> 0:01:42.880
<v Speaker 1>But I'm just beyond intrigued by this conceptually, because traditional

0:01:42.880 --> 0:01:46.680
<v Speaker 1>spycraft was all about obscuring your identity from other people.

0:01:47.319 --> 0:01:50.280
<v Speaker 1>Like all those crazy disguises and prosthetics and stuff you

0:01:50.360 --> 0:01:53.880
<v Speaker 1>see in movies like Mitchell Impossible, They're all based on

0:01:54.040 --> 0:01:56.480
<v Speaker 1>real CIA technologies, or at least many of them are.

0:01:56.920 --> 0:01:59.120
<v Speaker 1>And David told me that the art of disguise has

0:01:59.160 --> 0:02:01.720
<v Speaker 1>gotten so good that you can easily change your race,

0:02:02.040 --> 0:02:04.840
<v Speaker 1>even your gender. And that's just the stuff we know about.

0:02:05.600 --> 0:02:09.799
<v Speaker 1>But now intelligence agencies are facing a radical new problem,

0:02:09.840 --> 0:02:11.840
<v Speaker 1>which is how do you trig a machine?

0:02:12.280 --> 0:02:14.840
<v Speaker 3>So while you might have a mask that completely changes

0:02:14.880 --> 0:02:18.160
<v Speaker 3>your appearance, a retina scan or a fingerprint could completely

0:02:18.200 --> 0:02:19.119
<v Speaker 3>blow your cover.

0:02:19.280 --> 0:02:21.959
<v Speaker 1>That's right. So the story began with a guy who

0:02:22.000 --> 0:02:24.440
<v Speaker 1>worked with the CIA who warned a few years ago

0:02:24.880 --> 0:02:28.480
<v Speaker 1>that computer vision would soon be able to identify people

0:02:28.880 --> 0:02:32.400
<v Speaker 1>at a far off distance just based on the signature

0:02:32.400 --> 0:02:36.000
<v Speaker 1>of how they walked. It's called gate analysis, and at

0:02:36.040 --> 0:02:37.560
<v Speaker 1>the time he was laughed out of the room as

0:02:37.560 --> 0:02:41.520
<v Speaker 1>a scaremonger, but it turns out he'd actually seen into

0:02:41.520 --> 0:02:45.960
<v Speaker 1>the future. It's almost impossible nowadays for humans through outsmart

0:02:45.960 --> 0:02:49.000
<v Speaker 1>machines because of something that David called, and I love

0:02:49.080 --> 0:02:53.519
<v Speaker 1>this phrase, digital dust. What that refers to is the

0:02:53.600 --> 0:02:57.239
<v Speaker 1>data signatures we leave behind no matter how hard we try.

0:02:57.360 --> 0:03:00.960
<v Speaker 1>Here's the real kicker. Not leaving digital dust could be

0:03:01.120 --> 0:03:04.920
<v Speaker 1>just as revealing as leaving it. In the past, spies

0:03:04.960 --> 0:03:08.120
<v Speaker 1>could slip under cover with just a story. But now

0:03:08.320 --> 0:03:12.519
<v Speaker 1>if they don't have decades of online activity LinkedIn Instagram, Facebook, etc.

0:03:13.040 --> 0:03:16.520
<v Speaker 1>To support that story, it just doesn't make any sense

0:03:16.560 --> 0:03:19.760
<v Speaker 1>to their adversaries. In fact, as David put it, the

0:03:19.840 --> 0:03:22.720
<v Speaker 1>harder you try to hide, the more visible you become.

0:03:23.360 --> 0:03:25.880
<v Speaker 1>And all of this leads to some existential questions about

0:03:25.919 --> 0:03:29.160
<v Speaker 1>the future of spying. How does a CIA adapt, what

0:03:29.200 --> 0:03:32.720
<v Speaker 1>happens if it doesn't, and who's responsible for dragging into

0:03:32.720 --> 0:03:35.760
<v Speaker 1>the future. One of those people is the guy I mentioned,

0:03:35.800 --> 0:03:39.000
<v Speaker 1>the person who warned about data analysis. He's now innovating

0:03:39.160 --> 0:03:41.760
<v Speaker 1>on the outside and trying to set into the CIA.

0:03:42.400 --> 0:03:44.480
<v Speaker 1>But to start with, I wanted to know about how

0:03:44.520 --> 0:03:47.040
<v Speaker 1>the CIA has tried to adapt already and how it's going.

0:03:47.520 --> 0:03:52.000
<v Speaker 1>Here's my conversation with David Ignatius. You chart in the

0:03:52.120 --> 0:03:57.480
<v Speaker 1>piece how a number of poorly executed technological ideas have

0:03:57.600 --> 0:04:01.000
<v Speaker 1>led to networks being dismantled in places like China and Iran.

0:04:01.040 --> 0:04:05.000
<v Speaker 1>I assume by networks being dismantled that means people getting

0:04:05.520 --> 0:04:09.160
<v Speaker 1>arrested and maybe even killed, and that only killed.

0:04:09.440 --> 0:04:14.200
<v Speaker 2>Some of the executions were ruesomely done in ways to

0:04:14.640 --> 0:04:18.400
<v Speaker 2>make sure that our overhead satellites can see them. So

0:04:18.800 --> 0:04:22.560
<v Speaker 2>it is true that our attempts to come up with

0:04:22.680 --> 0:04:28.440
<v Speaker 2>clever technologies have sometimes been half baked. I'll give you

0:04:28.480 --> 0:04:32.599
<v Speaker 2>an example that's been published by me and others. The

0:04:32.640 --> 0:04:36.600
<v Speaker 2>agency had a seemingly very clever method for communicating with

0:04:36.720 --> 0:04:42.040
<v Speaker 2>recruited agents, where it would give them access to computer

0:04:42.160 --> 0:04:46.760
<v Speaker 2>websites that fit their personalities. Let's suppose somebody was a

0:04:46.839 --> 0:04:51.720
<v Speaker 2>Liverpool soccer fan, so it'll be a Liverpool fan site.

0:04:52.120 --> 0:04:56.440
<v Speaker 2>But embedded in that fan site was a template for

0:04:56.560 --> 0:05:01.920
<v Speaker 2>communicating directly and very secretly through the Internet through a VPN,

0:05:02.080 --> 0:05:06.120
<v Speaker 2>a hole in the Internet back to Langley. The problem

0:05:06.360 --> 0:05:11.320
<v Speaker 2>was that these things all had the same back end electronically,

0:05:12.080 --> 0:05:15.719
<v Speaker 2>so once you stripped away the nominal cover. You know,

0:05:15.760 --> 0:05:19.160
<v Speaker 2>there were ones for Rasta fans in the West Indies,

0:05:19.240 --> 0:05:22.520
<v Speaker 2>there were ones for Cut three music fans. I mean,

0:05:22.520 --> 0:05:25.720
<v Speaker 2>if you can go to dozens of the dedicated sites,

0:05:25.760 --> 0:05:27.680
<v Speaker 2>but they all had the same back end, which was

0:05:27.720 --> 0:05:33.040
<v Speaker 2>really about covert communications. And in both Iran and China

0:05:34.160 --> 0:05:40.400
<v Speaker 2>this secret was assessed and then ruthlessly exploited. So you

0:05:40.480 --> 0:05:42.880
<v Speaker 2>have to be careful with technology. When you think you're

0:05:42.880 --> 0:05:46.160
<v Speaker 2>being smart, you have to go back and look at

0:05:46.200 --> 0:05:49.400
<v Speaker 2>it again because you may simply be being obvious in

0:05:49.440 --> 0:05:51.240
<v Speaker 2>a different way you hadn't considered.

0:05:52.360 --> 0:05:54.800
<v Speaker 1>One of the other examples that comes up in the

0:05:54.880 --> 0:06:00.320
<v Speaker 1>piece is American Kamando's I believe in Syria who whose

0:06:00.520 --> 0:06:03.279
<v Speaker 1>cell phones came from Fort Bragg, right, And so it

0:06:03.360 --> 0:06:05.760
<v Speaker 1>was very easy for an external pologies to locate where

0:06:05.760 --> 0:06:06.200
<v Speaker 1>they were.

0:06:06.680 --> 0:06:13.800
<v Speaker 2>So there was a very secret location in northeastern Syria,

0:06:14.520 --> 0:06:17.920
<v Speaker 2>in the Kurdish controlled area. I know about it because

0:06:17.920 --> 0:06:22.120
<v Speaker 2>I probably went there five times as an embed with

0:06:22.240 --> 0:06:25.719
<v Speaker 2>the Joint Special Operations Command, which is our most secret

0:06:25.760 --> 0:06:29.520
<v Speaker 2>and really our best military force, which is running operations

0:06:29.520 --> 0:06:33.640
<v Speaker 2>there to destroy Isis. Remember how frightening ISIS was so

0:06:34.080 --> 0:06:39.160
<v Speaker 2>a well meaning guy who was in the commercial advertising business,

0:06:39.480 --> 0:06:44.240
<v Speaker 2>whose business was picking up the little emissions from your

0:06:44.279 --> 0:06:48.720
<v Speaker 2>cell phones that tell ways where you are on the highway,

0:06:49.080 --> 0:06:52.279
<v Speaker 2>what cop car has pulled over. He wanted to help

0:06:52.440 --> 0:06:55.240
<v Speaker 2>refugees who were fleeing Syria those early in the early

0:06:55.279 --> 0:06:57.800
<v Speaker 2>days of the world. So I just bought from Syria

0:06:58.160 --> 0:07:02.000
<v Speaker 2>telcom companies bought all this data. Is very cheap because

0:07:02.000 --> 0:07:05.320
<v Speaker 2>nobody had any commercial applications whatsoever for that area. And

0:07:05.360 --> 0:07:09.080
<v Speaker 2>then he overlaid telephones in the area of Fort Bragg,

0:07:09.120 --> 0:07:12.080
<v Speaker 2>North Carolina, which is where Jaysack and all of our

0:07:12.080 --> 0:07:17.560
<v Speaker 2>most secret units operate, and he found there were all

0:07:17.600 --> 0:07:22.559
<v Speaker 2>these things at the cement factory, supposedly an abandoned French

0:07:22.600 --> 0:07:26.400
<v Speaker 2>cement factory. The advertisers just were like, oh my god,

0:07:26.400 --> 0:07:30.360
<v Speaker 2>look at all this. So there's an example where commercial

0:07:30.840 --> 0:07:36.240
<v Speaker 2>innovation just was much faster than the creativity or the

0:07:36.280 --> 0:07:41.640
<v Speaker 2>ability to detect new threats and opportunities on the intelligence side.

0:07:42.040 --> 0:07:44.920
<v Speaker 2>And that illustrates I think what in some ways OZ

0:07:44.960 --> 0:07:48.480
<v Speaker 2>is that is the biggest and most positive change that's happened.

0:07:49.280 --> 0:07:53.160
<v Speaker 2>Starting in nineteen ninety nine, then CI director George Tenne

0:07:53.440 --> 0:07:57.520
<v Speaker 2>was very foresighted and seeing that the CIA is falling

0:07:57.600 --> 0:08:01.640
<v Speaker 2>behind the pace of innovation of Silicon Valley. Tenant had

0:08:01.640 --> 0:08:05.000
<v Speaker 2>a good sense to create a CIA venture capital fund,

0:08:05.160 --> 0:08:08.320
<v Speaker 2>which he called ink Tel, and rather than getting the

0:08:08.400 --> 0:08:12.360
<v Speaker 2>usual Intel bureaucrat to run it, he went outside and

0:08:12.360 --> 0:08:15.960
<v Speaker 2>picked him in named Gilman Louie, who'd been running a

0:08:16.080 --> 0:08:22.400
<v Speaker 2>video game company, and Gilman Louie began going out and

0:08:22.480 --> 0:08:27.760
<v Speaker 2>looking for smart ideas that could help in intelligence missions,

0:08:28.160 --> 0:08:32.520
<v Speaker 2>recognizing that the pace of innovation in the private sectors

0:08:32.920 --> 0:08:35.600
<v Speaker 2>simply couldn't be matched by government. The government had to

0:08:35.640 --> 0:08:40.760
<v Speaker 2>find a way to use these technologies, if possible, work

0:08:40.840 --> 0:08:43.800
<v Speaker 2>with the people who were creating them. So that got

0:08:44.040 --> 0:08:47.640
<v Speaker 2>the intelligence communitee operating at the speed of innovation, if

0:08:47.679 --> 0:08:51.040
<v Speaker 2>you will, and I think has certainly accelerated the pace

0:08:51.080 --> 0:08:54.079
<v Speaker 2>of change that's happened in Britain a lot of our allies.

0:08:54.440 --> 0:08:57.400
<v Speaker 2>So this began with a good idea of George Tennant's.

0:08:57.440 --> 0:09:00.480
<v Speaker 2>But now I think pretty much universally understood that the

0:09:00.520 --> 0:09:04.400
<v Speaker 2>intelligence world wants totally closed, has to connect with this

0:09:04.679 --> 0:09:07.760
<v Speaker 2>wildly open and innovative world of talking about.

0:09:08.280 --> 0:09:11.400
<v Speaker 1>I saw in your piece those investments in Paneteer and Enderrail,

0:09:11.480 --> 0:09:14.960
<v Speaker 1>and presumably if you were a venture investor, you would

0:09:15.000 --> 0:09:19.280
<v Speaker 1>be beyond cockahoop with those returns. As the CIA, though

0:09:19.520 --> 0:09:23.640
<v Speaker 1>I mean financial performances is great, but as a CIA

0:09:24.400 --> 0:09:27.840
<v Speaker 1>in some sense threatening its own relevance by outsourcing so

0:09:27.920 --> 0:09:28.400
<v Speaker 1>much innovation.

0:09:28.440 --> 0:09:33.160
<v Speaker 2>As the private sector, the CIA, they're trying to keep up.

0:09:33.600 --> 0:09:36.559
<v Speaker 2>They have technical advisory boards and they has some very

0:09:36.600 --> 0:09:41.800
<v Speaker 2>patriotic people and in tech companies work with them. We

0:09:41.880 --> 0:09:46.120
<v Speaker 2>have this interesting problem that Elon Musk, to take one example,

0:09:46.840 --> 0:09:53.560
<v Speaker 2>came up with an extraordinarily powerful system Starlink with now

0:09:54.000 --> 0:10:00.000
<v Speaker 2>three thousand dollars satellites and lowerthorbit providing broadbeam signals over everywhere,

0:10:00.800 --> 0:10:04.320
<v Speaker 2>and that became the command and control network for the

0:10:04.400 --> 0:10:08.480
<v Speaker 2>Ukrainian military. But that then puts enormous power in the

0:10:08.559 --> 0:10:12.040
<v Speaker 2>hands of a private sector entrepreneurs. So what happens if

0:10:12.120 --> 0:10:16.800
<v Speaker 2>Eli Musk decides I have to sell teslas in China,

0:10:17.160 --> 0:10:19.280
<v Speaker 2>I have to make teslas and shine. I don't want to.

0:10:19.440 --> 0:10:22.079
<v Speaker 2>I'm sick of this. Ukraine will war, right, So I'm

0:10:22.120 --> 0:10:24.360
<v Speaker 2>going to cut off their ability to get those signals.

0:10:24.360 --> 0:10:26.600
<v Speaker 2>There have been moments when he's in fact threatened to

0:10:26.640 --> 0:10:29.120
<v Speaker 2>do that and then just pulled back. But it just

0:10:29.160 --> 0:10:32.440
<v Speaker 2>illustrates that the use of private technology is a double

0:10:32.520 --> 0:10:36.360
<v Speaker 2>edged sword. Yes, it accelerates the pace of your ability

0:10:36.400 --> 0:10:39.520
<v Speaker 2>to innovate, but it makes you more dependent on these

0:10:39.720 --> 0:10:45.800
<v Speaker 2>entrepreneurs who hopefully share your national interests and will protect

0:10:45.840 --> 0:10:50.360
<v Speaker 2>and advance your secret advantage. But you can't be sure

0:10:50.400 --> 0:10:53.160
<v Speaker 2>of that, and so that's a puzzle. I think people

0:10:53.160 --> 0:10:54.360
<v Speaker 2>are still struggling with.

0:10:55.120 --> 0:11:00.920
<v Speaker 1>Characterize where the CIA is on these issues and questions

0:11:00.920 --> 0:11:03.760
<v Speaker 1>that you've been raising, because the sense I get from

0:11:03.760 --> 0:11:06.880
<v Speaker 1>the piece is that you feel their way behind. I mean,

0:11:06.920 --> 0:11:09.880
<v Speaker 1>there's this great quote from a former CIA director, David Norman,

0:11:09.920 --> 0:11:12.679
<v Speaker 1>who says, if Henry Ford had gone to transportation customers

0:11:12.720 --> 0:11:14.720
<v Speaker 1>and us what they wanted, the would have said faster horses.

0:11:15.280 --> 0:11:17.720
<v Speaker 1>That's what the CIA has been trying to build, faster horses,

0:11:17.760 --> 0:11:18.680
<v Speaker 1>and it's pretty damning.

0:11:18.960 --> 0:11:23.600
<v Speaker 2>It is pretty damning. So they deny that they're as

0:11:23.880 --> 0:11:28.160
<v Speaker 2>out of it as a piece suggests. The question is

0:11:28.440 --> 0:11:33.040
<v Speaker 2>how fast you can move to fully adapt It can

0:11:33.080 --> 0:11:35.480
<v Speaker 2>be aware of something, but fully adapt to it. As

0:11:35.520 --> 0:11:39.520
<v Speaker 2>a different matter, people are still trying to I think

0:11:39.559 --> 0:11:41.520
<v Speaker 2>faster or worse. It is a little unfair, but they're

0:11:41.559 --> 0:11:45.720
<v Speaker 2>still trying to think about within the existing paradigm for espionage.

0:11:45.800 --> 0:11:47.760
<v Speaker 2>How can we do it better? How do we hide better?

0:11:47.800 --> 0:11:50.720
<v Speaker 2>How do we find ways to capture their signals without

0:11:50.720 --> 0:11:55.480
<v Speaker 2>being observed ourselves? But what's needed, people say, is something

0:11:55.920 --> 0:11:59.439
<v Speaker 2>really very new, a whole new way of thinking about operating.

0:12:00.120 --> 0:12:04.040
<v Speaker 2>Is John Ratcliffe, the CI director of a person who's creative,

0:12:04.720 --> 0:12:10.120
<v Speaker 2>disruptive enough to orchestrate that transformation. We'll see, But legacy

0:12:10.200 --> 0:12:14.880
<v Speaker 2>systems have a momentum sort of weight that Look at

0:12:14.880 --> 0:12:18.440
<v Speaker 2>aircraft carriers in the Navy. I mean, it's twenty years ago.

0:12:18.559 --> 0:12:21.640
<v Speaker 2>We knew that they were all vulnerable. They disappeared the

0:12:21.679 --> 0:12:25.079
<v Speaker 2>first minutes or hours of any attack. But they're still

0:12:25.080 --> 0:12:27.600
<v Speaker 2>out there, and good luck trying to get rid of.

0:12:27.640 --> 0:12:37.120
<v Speaker 1>Them after the break the story, even elite soldier to

0:12:37.400 --> 0:12:59.160
<v Speaker 1>tech founder who's creating an AI souperagent, stay with us. Yeah,

0:12:59.160 --> 0:13:02.079
<v Speaker 1>this's light in the p The CIA's technology challenge is

0:13:02.080 --> 0:13:05.280
<v Speaker 1>a little noted example of a transformation that's happening in

0:13:05.360 --> 0:13:09.679
<v Speaker 1>every area of defense and security today. Smart machines can

0:13:09.760 --> 0:13:13.480
<v Speaker 1>outweit humans now, even for Tim Cook, even for the

0:13:13.559 --> 0:13:18.920
<v Speaker 1>CEOs of silicon value based technology companies, knowing how to

0:13:19.000 --> 0:13:23.440
<v Speaker 1>balance defending the core product with integrating this wave of

0:13:23.520 --> 0:13:26.880
<v Speaker 1>new technology that is so fast moving. Even Google is

0:13:26.960 --> 0:13:32.600
<v Speaker 1>struggling with it. How does a government organization saddled with

0:13:32.640 --> 0:13:35.920
<v Speaker 1>the bureaucracy that even the most forward thinking government organizations

0:13:36.520 --> 0:13:41.080
<v Speaker 1>come with. Is there an overall strategic response to this problem?

0:13:41.160 --> 0:13:42.640
<v Speaker 1>Is crisis?

0:13:42.800 --> 0:13:46.880
<v Speaker 2>So I think the obvious answer is that new ideas

0:13:47.440 --> 0:13:50.960
<v Speaker 2>in the intelligence business that are really powerful, that allow

0:13:51.040 --> 0:13:54.320
<v Speaker 2>you to collect secrets that you didn't have before, that

0:13:54.480 --> 0:14:02.040
<v Speaker 2>open new areas for collection and analysis become irresistible. These

0:14:02.080 --> 0:14:06.079
<v Speaker 2>secrets are so powerful once you learn to read somebody's mail,

0:14:06.679 --> 0:14:11.160
<v Speaker 2>listening to their phone calls, it's resistible. Because policymakers, once

0:14:11.200 --> 0:14:15.160
<v Speaker 2>they've got that conversation between the General secretary and his

0:14:15.440 --> 0:14:17.520
<v Speaker 2>chief of staff, they want it. They want to get

0:14:17.520 --> 0:14:21.280
<v Speaker 2>the good stuff. So there's always a demand for policy

0:14:21.280 --> 0:14:25.280
<v Speaker 2>makers for the very best intelligence. And if technology can

0:14:25.680 --> 0:14:29.120
<v Speaker 2>allow people to get more of that or sustain the

0:14:29.120 --> 0:14:32.280
<v Speaker 2>flow of it, certainly the demand will be there. I

0:14:32.360 --> 0:14:36.400
<v Speaker 2>just wouldn't note if you look at what the CIA

0:14:36.560 --> 0:14:42.360
<v Speaker 2>and other agencies were able to know about Russian intentions

0:14:43.200 --> 0:14:48.280
<v Speaker 2>in late twenty twenty one, when Europe and even Ukraine

0:14:48.400 --> 0:14:51.240
<v Speaker 2>said no, the Russians aren't going to attack, and Bill

0:14:51.240 --> 0:14:54.480
<v Speaker 2>Burns and his colleagues kept going out and saying, yes

0:14:54.520 --> 0:14:58.600
<v Speaker 2>they are, and that they had detailed intelligence. They were

0:14:58.640 --> 0:15:03.840
<v Speaker 2>reading Russian intentions like a phone book. It's still not

0:15:03.920 --> 0:15:07.480
<v Speaker 2>clear just how they had such precise intelligence, but they

0:15:07.560 --> 0:15:09.640
<v Speaker 2>knew right where they were coming out. How Ukraine in

0:15:09.680 --> 0:15:11.960
<v Speaker 2>the early days of the war was able to be

0:15:12.000 --> 0:15:15.240
<v Speaker 2>so successful. We knew where they were coming. They were

0:15:15.240 --> 0:15:18.240
<v Speaker 2>coming to the airport just west of Kiev. We knew

0:15:18.280 --> 0:15:20.520
<v Speaker 2>exactly what they were going to try to do and

0:15:20.880 --> 0:15:23.360
<v Speaker 2>ban there were people there waiting for them, and they

0:15:23.480 --> 0:15:27.120
<v Speaker 2>just took them out. But the electronic coordination of all

0:15:27.160 --> 0:15:29.800
<v Speaker 2>the systems that have to have to operate simultaneously we

0:15:29.800 --> 0:15:31.840
<v Speaker 2>don't think about. You know, if you've got an hour

0:15:31.920 --> 0:15:35.440
<v Speaker 2>of time and you've got all these different multiple fires,

0:15:35.480 --> 0:15:40.000
<v Speaker 2>different drones, different systems that cannot be done by human beings,

0:15:40.000 --> 0:15:43.240
<v Speaker 2>too complicated. So there's a way in which the war

0:15:43.280 --> 0:15:47.240
<v Speaker 2>in Ukraine was an algorithm and the complex systems for

0:15:47.360 --> 0:15:50.240
<v Speaker 2>handling data that's never been done in the warfare before.

0:15:50.400 --> 0:15:55.040
<v Speaker 2>That is something of that complexity managed electronically simultaneously. You've

0:15:55.080 --> 0:15:56.000
<v Speaker 2>never seen that before.

0:15:56.280 --> 0:16:00.440
<v Speaker 1>So the diagnosis is actually rumors of the demise maybe

0:16:00.720 --> 0:16:03.000
<v Speaker 1>over exaggerating. What you're really looking at in the piece

0:16:03.080 --> 0:16:06.800
<v Speaker 1>is more continued relevance in a five to ten year horizon.

0:16:07.480 --> 0:16:10.480
<v Speaker 2>So I'm looking at the people who were trying to

0:16:10.560 --> 0:16:13.880
<v Speaker 2>change the paradigm. But the point that should encourage people,

0:16:13.960 --> 0:16:17.120
<v Speaker 2>if you're an American, you have to say that is

0:16:17.680 --> 0:16:20.720
<v Speaker 2>there are a lot of really smart people out there.

0:16:20.880 --> 0:16:23.400
<v Speaker 2>Many of them got frustrated being the CI. It's pretty

0:16:23.440 --> 0:16:26.680
<v Speaker 2>bureaucratic these days, got a little political, to put a

0:16:26.760 --> 0:16:31.680
<v Speaker 2>mild late so people say. And they're out there trying

0:16:31.680 --> 0:16:34.120
<v Speaker 2>to create companies that are going to be good for

0:16:34.160 --> 0:16:37.840
<v Speaker 2>their former colleagues, solve problems that they couldn't solve while

0:16:37.840 --> 0:16:41.240
<v Speaker 2>they were in an operation, and so people are coming

0:16:41.320 --> 0:16:44.200
<v Speaker 2>up with, as I say, quite innovative ideas.

0:16:45.000 --> 0:16:47.480
<v Speaker 1>Talk a bit about Aaron Brown's story, if you don't mind,

0:16:47.680 --> 0:16:50.760
<v Speaker 1>how you first met him, how his insights within the

0:16:50.760 --> 0:16:54.400
<v Speaker 1>agency were rejected, and how he came to design Lumbra

0:16:54.840 --> 0:16:57.600
<v Speaker 1>and even engage with Samiltman in the very early days

0:16:57.600 --> 0:16:59.040
<v Speaker 1>of chat Chipet's release.

0:17:00.080 --> 0:17:03.480
<v Speaker 2>Brown was an Army ranger and if you don't know

0:17:03.520 --> 0:17:06.520
<v Speaker 2>anything about what they do at ranger school, you know

0:17:06.680 --> 0:17:09.760
<v Speaker 2>these are the guys who can climb up a cliff

0:17:09.800 --> 0:17:13.440
<v Speaker 2>face and run ten miles or a fifty pound pack

0:17:13.480 --> 0:17:16.680
<v Speaker 2>on their back and get in and out of incredible places.

0:17:16.800 --> 0:17:20.800
<v Speaker 2>They're tough fighters and end up going to these special units,

0:17:21.440 --> 0:17:25.480
<v Speaker 2>And like many very good soldiers, Aaron ended up getting

0:17:25.520 --> 0:17:28.800
<v Speaker 2>cut to the CI and worked in its counter terrorism

0:17:28.880 --> 0:17:32.480
<v Speaker 2>operations in the days of the pursuit of wassamb B

0:17:32.520 --> 0:17:35.840
<v Speaker 2>and Laud. All the while he was with a deep

0:17:36.000 --> 0:17:39.479
<v Speaker 2>engineer's interest in electronics, was trying to think about the

0:17:39.680 --> 0:17:42.600
<v Speaker 2>tools of his craft and whether they were adequate. And

0:17:42.920 --> 0:17:46.159
<v Speaker 2>when the bin laden work was done, he began to

0:17:46.240 --> 0:17:51.399
<v Speaker 2>worry about the vulnerability of officers overseas to this technology

0:17:51.400 --> 0:17:54.080
<v Speaker 2>that could recognize the way they walked. And just think

0:17:54.080 --> 0:17:57.959
<v Speaker 2>about it. If you've got recordings of everybody enters all

0:17:58.000 --> 0:18:01.639
<v Speaker 2>the airports of China, everybody who enters the US embassy,

0:18:01.720 --> 0:18:04.399
<v Speaker 2>or all the consulates that we have, you could end

0:18:04.480 --> 0:18:06.800
<v Speaker 2>up having a library that you can then run against

0:18:07.200 --> 0:18:10.000
<v Speaker 2>people all over China. You see, well, where that guy

0:18:10.000 --> 0:18:12.200
<v Speaker 2>from the US embassy, what's he doing in chung Du?

0:18:12.920 --> 0:18:15.639
<v Speaker 2>Why do you go to that forest that's ten miles

0:18:15.680 --> 0:18:19.639
<v Speaker 2>out of town. What's going on here? So he left

0:18:19.680 --> 0:18:24.159
<v Speaker 2>the agency year eighteen months ago and with a friend

0:18:24.320 --> 0:18:29.239
<v Speaker 2>started this company, Lumbra, And his basic idea is that

0:18:29.480 --> 0:18:35.280
<v Speaker 2>it's not simply these magnificent ais that are moving towards

0:18:35.440 --> 0:18:39.080
<v Speaker 2>what we call superintelligence without being quite sure what that means.

0:18:39.080 --> 0:18:42.000
<v Speaker 2>But computers that begin to be able to think in

0:18:42.080 --> 0:18:46.199
<v Speaker 2>real human like ways and begin to give themselves instructions,

0:18:46.200 --> 0:18:49.600
<v Speaker 2>maybe by writing their own programs, begin to be able

0:18:49.600 --> 0:18:52.240
<v Speaker 2>to speak to each other. It's one thing super intelligence

0:18:52.280 --> 0:18:55.480
<v Speaker 2>will be able to do. And his insight was that

0:18:56.040 --> 0:19:01.120
<v Speaker 2>just as our brilliant brains need hands, legs, the ability

0:19:01.160 --> 0:19:05.520
<v Speaker 2>to get to what we need to assemble to then process,

0:19:06.359 --> 0:19:09.240
<v Speaker 2>computers are going to need agents agentic AI that will

0:19:09.240 --> 0:19:13.240
<v Speaker 2>help them assemble that disparate pieces of information, think and

0:19:13.280 --> 0:19:17.119
<v Speaker 2>then analyze and make sense of it. No human being could.

0:19:17.840 --> 0:19:20.919
<v Speaker 2>So he decided that that was his going to be

0:19:20.960 --> 0:19:24.240
<v Speaker 2>his area, this central nervous system that's going to connect

0:19:24.800 --> 0:19:28.520
<v Speaker 2>the AI brain with all the other parts of the

0:19:28.560 --> 0:19:32.640
<v Speaker 2>system that will make it most effective. He in turn

0:19:32.760 --> 0:19:35.800
<v Speaker 2>introduced me to other friends, a guy who was thinking

0:19:35.840 --> 0:19:38.720
<v Speaker 2>about this problem of how our cell phones give off

0:19:38.760 --> 0:19:41.639
<v Speaker 2>our identity, but we still have to communicate, So how

0:19:41.640 --> 0:19:44.800
<v Speaker 2>our officers going to communicate without giving away their position.

0:19:44.920 --> 0:19:47.680
<v Speaker 2>And it turns out he came up with an incredibly

0:19:47.920 --> 0:19:51.000
<v Speaker 2>ingenious technology that essentially the three or four levels of

0:19:51.000 --> 0:19:54.360
<v Speaker 2>identifiers in our phones were not aware of that give

0:19:54.400 --> 0:19:57.440
<v Speaker 2>off our identification, but kind a way to bounce those

0:19:57.480 --> 0:20:02.439
<v Speaker 2>identifiers among the different a thousands of users of his system,

0:20:03.200 --> 0:20:07.159
<v Speaker 2>so you can't really tell where any particular signal is

0:20:07.200 --> 0:20:10.160
<v Speaker 2>coming from, even if you capture it. It's called Cape.

0:20:10.560 --> 0:20:14.520
<v Speaker 2>Third company that interesting to me was called Strider. It's

0:20:14.600 --> 0:20:18.520
<v Speaker 2>based in Salt Lake City, Utah. Essentially, what they're doing

0:20:18.800 --> 0:20:23.640
<v Speaker 2>is reading other people's digital dust. We're not the only

0:20:23.680 --> 0:20:27.120
<v Speaker 2>ones who leave digital dust. The Chinese are so intent

0:20:27.200 --> 0:20:31.000
<v Speaker 2>on monitoring their own population. They've got cameras on every

0:20:31.040 --> 0:20:32.920
<v Speaker 2>street and it's not to watch us, is to watch

0:20:33.040 --> 0:20:36.480
<v Speaker 2>Chinese people. And it turns out that that's an entry

0:20:36.520 --> 0:20:39.760
<v Speaker 2>way for companies like Strider to collect an amazing amount

0:20:39.840 --> 0:20:43.440
<v Speaker 2>of information. I mean, I've looked at it what they

0:20:43.440 --> 0:20:46.920
<v Speaker 2>can get, and it's it's pretty incredible. Imagine how long

0:20:46.960 --> 0:20:50.040
<v Speaker 2>it would take you to do that kind of forensics

0:20:50.080 --> 0:20:53.480
<v Speaker 2>in a free digital world. So that's something that that

0:20:53.600 --> 0:20:56.600
<v Speaker 2>company is doing. They think they probably do it better

0:20:56.600 --> 0:20:59.959
<v Speaker 2>than any internal agency in the government. The fact that

0:21:00.080 --> 0:21:03.160
<v Speaker 2>they were willing to talk about it with me in detail,

0:21:03.160 --> 0:21:06.359
<v Speaker 2>I didn't sneak them in the telling me, it means

0:21:06.359 --> 0:21:08.959
<v Speaker 2>that they're fairly confident they can continue to do it.

0:21:09.280 --> 0:21:12.600
<v Speaker 2>So those are three little examples of companies that are

0:21:12.720 --> 0:21:15.399
<v Speaker 2>playing around at the frontier, and I hope people will

0:21:15.440 --> 0:21:18.160
<v Speaker 2>find that encouraging. I would hate to think that we

0:21:18.160 --> 0:21:23.880
<v Speaker 2>were just a slumbering, sloppy giant with the intelligence equivalent

0:21:23.880 --> 0:21:26.639
<v Speaker 2>of aircraft carriers just sitting there away and get blown

0:21:26.640 --> 0:21:27.879
<v Speaker 2>out of the water. I don't want that.

0:21:28.359 --> 0:21:30.920
<v Speaker 1>We've talked a lot about concerns about the CIA falling

0:21:31.040 --> 0:21:34.680
<v Speaker 1>behind technologically. Do you have any concerns about a swing

0:21:34.760 --> 0:21:38.240
<v Speaker 1>too hot in the other direction and onboarding too many

0:21:38.840 --> 0:21:44.040
<v Speaker 1>untested private technologies that could create risk either to American

0:21:44.080 --> 0:21:46.359
<v Speaker 1>citizens or systems risk. I mean, what's the kind of

0:21:46.359 --> 0:21:48.359
<v Speaker 1>flip side concern and everything we've been talking about.

0:21:49.240 --> 0:21:52.920
<v Speaker 2>So, you know, this is a period where we're seeing

0:21:53.240 --> 0:21:58.199
<v Speaker 2>that the powers of government can be misused in what

0:21:58.359 --> 0:22:00.960
<v Speaker 2>to me are quite disturbing. When is you know, when

0:22:01.680 --> 0:22:06.520
<v Speaker 2>law enforcement is federalized, you see overreach, and you could

0:22:06.520 --> 0:22:09.320
<v Speaker 2>see a technological version of that overreach. You could see

0:22:09.320 --> 0:22:15.040
<v Speaker 2>the application of facial recognition software. In China, a citizens

0:22:15.080 --> 0:22:19.160
<v Speaker 2>can't go out of his town without permission. He gets

0:22:19.400 --> 0:22:23.760
<v Speaker 2>a score based on his performance at work, rating his

0:22:24.160 --> 0:22:27.920
<v Speaker 2>social merit. He gets a little gold star if he's

0:22:27.960 --> 0:22:31.720
<v Speaker 2>a dutiful citizens supporting the Chinese Communist Party. We don't

0:22:31.760 --> 0:22:35.359
<v Speaker 2>want that, and we don't want our schools, I don't

0:22:35.400 --> 0:22:40.760
<v Speaker 2>think to become rigid in what they instruct We want creativity,

0:22:40.840 --> 0:22:45.600
<v Speaker 2>not uniformity, and these tools can create uniformity sadly. So

0:22:45.880 --> 0:22:48.960
<v Speaker 2>you know, we have to remember that the power of

0:22:49.000 --> 0:22:52.960
<v Speaker 2>our federal government is overwhelming and that we want it

0:22:53.000 --> 0:22:56.440
<v Speaker 2>to be strong in dealing with our adversaries. But if

0:22:56.480 --> 0:23:00.480
<v Speaker 2>that power begins to be used against American citizens or

0:23:00.560 --> 0:23:04.359
<v Speaker 2>inappropriate ways around the world, everybody should be watching. So

0:23:04.400 --> 0:23:07.439
<v Speaker 2>that's I think my biggest challenge is a journalist. I

0:23:07.480 --> 0:23:12.280
<v Speaker 2>think this revolution is needed and beneficial, but I think

0:23:12.320 --> 0:23:16.040
<v Speaker 2>the dangers are ones that we have to keep in mind.

0:23:16.760 --> 0:23:18.600
<v Speaker 2>It can't be so worried about the danger you just

0:23:18.600 --> 0:23:20.480
<v Speaker 2>to stop and say we're not going to do it.

0:23:20.560 --> 0:23:22.680
<v Speaker 2>That would be a mistake. But you do have to

0:23:22.720 --> 0:23:25.119
<v Speaker 2>be eventualan all the time about what could happen.

0:23:26.240 --> 0:23:28.280
<v Speaker 1>Well, David Nations, thank you so much for joining us

0:23:28.280 --> 0:23:30.879
<v Speaker 1>on text us today for a fascinating conversation, and I

0:23:30.960 --> 0:23:32.639
<v Speaker 1>hope you'll join us again before too long.

0:23:32.520 --> 0:23:33.960
<v Speaker 2>With pleasure Us. Thank you very much for.

0:23:55.920 --> 0:23:59.080
<v Speaker 1>Text Stuff. I amost Forelosian and I'm care Price. This

0:23:59.200 --> 0:24:03.080
<v Speaker 1>episode was Preduce Spy, Eliza Dennis, Melissa Slaughter, and Tyler Hill.

0:24:03.320 --> 0:24:06.360
<v Speaker 1>It was executive produced by Me, Karen Price, and Kate

0:24:06.400 --> 0:24:10.880
<v Speaker 1>Osborne for Kaleidoscope and Katrian Novelle for iHeart Podcasts. Jack

0:24:10.920 --> 0:24:14.360
<v Speaker 1>Insley makes this episode and Kyle Murdoch wrote out theme song.

0:24:14.560 --> 0:24:16.680
<v Speaker 3>Join us on Friday for the week in tech as

0:24:16.720 --> 0:24:18.440
<v Speaker 3>and I will run through the tech headlines you may

0:24:18.440 --> 0:24:21.080
<v Speaker 3>have missed. Please rate, review, and reach out to us

0:24:21.080 --> 0:24:39.600
<v Speaker 3>at tech Stuff podcast at gmail dot com.