Latest / What's Your Problem? / How AI Could Actually Make the World Better

Transcript

How AI Could Actually Make the World Better · What's Your Problem?

0:15
Speaker 1: Pushkin. I'm Jacob Goldstein. This is what's your problem? And
0:31
Speaker 1: my guest today is Josh Tieringel. He's a staff writer
0:35
Speaker 1: at The Atlantic, and he just wrote a new book
0:37
Speaker 1: called AI for Good And the subtitle of the book
0:41
Speaker 1: explains the problem that Josh is interested in. The subtitle
0:44
Speaker 1: is how real people are using artificial intelligence to fix
0:48
Speaker 1: things that matter. As you'll hear, I talked with Josh
0:51
Speaker 1: about several of these real people. Professor trying to understand
0:55
Speaker 1: her nonverbal son, a general who ran the logistics side
0:59
Speaker 1: of Operation Warp Speed, a hospital CEO trying to reduce
1:03
Speaker 1: the rate of deadly infections. And one thing just to
1:06
Speaker 1: note here is you know, and I guess this is
1:08
Speaker 1: not surprising in the state of AI. Some of these
1:11
Speaker 1: stories are not yet finished. They're kind of stories that
1:13
Speaker 1: are still in the middle. But before we got into
1:16
Speaker 1: the specific stories, I asked Josh what common threads he
1:20
Speaker 1: saw across all these people he wrote about, who are
1:23
Speaker 1: you trying to use AI to do good things?
1:26
Speaker 2: They're largely stubborn. Uh huh, They're really stubborn. And you know,
1:30
Speaker 2: one of the things that I admire. I'm very impatient
1:33
Speaker 2: and when I see a problem. Oftentimes I have what
1:36
Speaker 2: I think is a fairly normal human impulse, which is like,
1:39
Speaker 2: wipe the mismatched puzzle pieces off the table and get
1:42
Speaker 2: a different puzzle. Right, It's frustrating to dig into systems
1:46
Speaker 2: to learn technology, and these people, they don't get frustrated,
1:51
Speaker 2: not to the point of destroying the system. In fact,
1:54
Speaker 2: most of them have no experience with AI or technology.
1:57
Speaker 2: They care about medicine, they care about teaching, They care
2:00
Speaker 2: about solving a really hard problem, and so they just
2:03
Speaker 2: keep looking for solutions. And when something seems promising, their
2:07
Speaker 2: ability to concentrate and learn, earn the technology of that
2:10
Speaker 2: promise and apply it to their problem is through the roof.
2:15
Speaker 2: And so I just kept being awed by the fact
2:18
Speaker 2: that they were curious, they're patient, They were impatient at
2:22
Speaker 2: exactly the right moments in the sense that things weren't moving.
2:26
Speaker 2: They knew how to deploy either their political capital or
2:29
Speaker 2: their temper in ways to get things moving. And you know,
2:34
Speaker 2: I found time and time again that when I would
2:36
Speaker 2: talk to other people who were witnessing these phenomenon, whether
2:38
Speaker 2: it was in hospitals or in government, we would all
2:41
Speaker 2: just kind of be like, Yeah, these are the ones
2:44
Speaker 2: making it happen. They're kind of Samurai, right. They care
2:47
Speaker 2: about their problem more than other people care about standing
2:50
Speaker 2: in the way of the solution, and they get it done.
2:53
Speaker 1: Tell me about Christy Johnson.
2:55
Speaker 2: Yeah, So, Christy Johnson is certainly one of the most
3:00
Speaker 2: inspiring interesting people I've met in my journalism career. So, I,
3:05
Speaker 2: like a lot of people, was thinking about AI and
3:08
Speaker 2: human connect and reading all the headlines about the sort
3:11
Speaker 2: of parasocial crisis in which people were dating, chat GPT,
3:16
Speaker 2: falling in love with Claude, realizing that this was mimicking
3:20
Speaker 2: a lot of the worst elements of social media, seeing
3:23
Speaker 2: the studies and getting very, very depressed about what AI
3:27
Speaker 2: would mean for human connection. So I was having that
3:29
Speaker 2: conversation with Rosin Picard, who runs the MIT Media Lab,
3:33
Speaker 2: and she basically said, look, you got to meet Christy,
3:35
Speaker 2: and so she introduced me. Christy Johnson was a physicist
3:40
Speaker 2: and a brilliant student. She met another physicist in a program.
3:47
Speaker 2: He is an astrophysicist and actually was on the team
3:50
Speaker 2: that took the first picture of a black hole. The
3:53
Speaker 2: two of them fell in the deepest form of nerd
3:55
Speaker 2: love you could possibly imagine. I mean even the two
3:59
Speaker 2: of them laugh at just how dorky their love letters
4:01
Speaker 2: are to each other, all about physics and math and
4:04
Speaker 2: the world. They get married and they have a kid,
4:08
Speaker 2: and about a couple of weeks in she's like, there's
4:12
Speaker 2: something wrong with my son. And she's kind of gas
4:14
Speaker 2: lit by the medical profession for a year just hearing no,
4:17
Speaker 2: he's fine, Stop worrying, enjoy your kid. Finally they get
4:21
Speaker 2: some testing done and he has a genetic deficiency. And
4:26
Speaker 2: the other six children like him in the world are
4:29
Speaker 2: all severely autistic, epileptic, and completely nonverbal, and so this
4:35
Speaker 2: is the kind of diagnosis that changes everybody's lives. Within
4:39
Speaker 2: the family. They begin to grieve a little bit for
4:44
Speaker 2: their expectations for their son, but they're also scientists, and
4:47
Speaker 2: Christy and Michael both are like, we should understand this,
4:51
Speaker 2: we should do something about it. A couple of years
4:54
Speaker 2: ago past and Christy is like, no longer all that
4:57
Speaker 2: interested in physics. She's very interested in her son. And
5:01
Speaker 2: so what she set out to do was solve the
5:03
Speaker 2: problem of the more than one million nonverbal Americans who
5:08
Speaker 2: are stuck to their parents and their caregivers. That's it
5:11
Speaker 2: for everybody in that universe, the universe shrinks.
5:16
Speaker 1: Just to be clear, what is her dream? What is
5:18
Speaker 1: the outcome she hopes to achieve.
5:20
Speaker 2: Her dream is that her son and other people like
5:23
Speaker 2: him can interact with people they don't know all that
5:25
Speaker 2: well and share what it is they want and need.
5:29
Speaker 1: So a machine will turn their nonverbal vocalizations into words. Yeah,
5:35
Speaker 1: verbal people can understand that other people can respond to
5:38
Speaker 1: and interact with. And so she made a bet. She
5:40
Speaker 1: basically thought, look, I see where AI.
5:42
Speaker 2: And machine learning are right now. My hunch is that
5:45
Speaker 2: they will get better and better. But what I need
5:48
Speaker 2: is data. And so she created a scientific protocol to
5:52
Speaker 2: get vocalizations from these kids, and she used every tool available.
5:57
Speaker 2: So she went on YouTube, she went on social media.
5:59
Speaker 2: She said, this is me and my son. This is
6:01
Speaker 2: what happens when he's hungry. Listen to that vocalization, and
6:04
Speaker 2: now I'm recording it, and she began to generate a
6:08
Speaker 2: little bit of momentum. She went to the MIT Media
6:10
Speaker 2: Lab and spoke to Rosbaccard one night, kind of cold
6:13
Speaker 2: calder and Roz realized, like, this woman has no qualifications
6:18
Speaker 2: to be at the media lab, but my god, like,
6:21
Speaker 2: what an incredible human she got there, she did her PhD.
6:27
Speaker 2: And she proved that if you can gather enough of
6:29
Speaker 2: these vocalizations as audio files, you can begin to attack
6:33
Speaker 2: the problem. And so what she learned largely through using AIS.
6:36
Speaker 2: At first you have to multiply the files, so a
6:41
Speaker 2: system like chat, GPT or Claude is trained on potentially
6:47
Speaker 2: trillions of data points in order to become as fluent
6:50
Speaker 2: as it is. She had a couple thousand audio files.
6:54
Speaker 2: So the first thing she did is get her graduate
6:57
Speaker 2: students to figure out how to synthesize more data, how
7:00
Speaker 2: to take those data files, those original sounds, and grow
7:04
Speaker 2: them and multiply them.
7:05
Speaker 1: It's like what they're trying to do with self driving cars,
7:08
Speaker 1: right exactly, like drive on pretend roads essentially exactly.
7:12
Speaker 2: It's called synthetic data. And AI is really good at
7:15
Speaker 2: making synthetic data.
7:16
Speaker 1: And they're even they're working on that for text, right
7:18
Speaker 1: because now the current models basically are trained on the
7:20
Speaker 1: whole Internet, so like there's no more real text for
7:23
Speaker 1: them to read, so they need synthetic texts.
7:25
Speaker 2: They're out of stuff, right, And so if you want
7:28
Speaker 2: the training to advance, you have to synthesize a lot
7:30
Speaker 2: of this data, and she's continuing to grow the organic
7:32
Speaker 2: files as well.
7:34
Speaker 1: Tell me about ROSCO.
7:36
Speaker 2: Yeah, So ROSCO is a protocol and basically, you can't
7:42
Speaker 2: do what Christy is trying to do without science because
7:46
Speaker 2: you need some sort of rigorous idea of what you're
7:49
Speaker 2: testing for or else. The data that you're recording is meaningless.
7:53
Speaker 1: It's not really data.
7:55
Speaker 2: It's not really data, it just sounds. And so working
7:58
Speaker 2: with a grad student, what she figured out is, Okay,
8:00
Speaker 2: we don't need all the data. All the data is unprocessable.
8:05
Speaker 2: We need the right kinds of data for this task.
8:07
Speaker 2: And so what they did is create this protocol where
8:10
Speaker 2: they would run families through a certain kind of test
8:14
Speaker 2: and it was split timed so that it was standardized,
8:16
Speaker 2: and they would say, for two minutes, give your child
8:21
Speaker 2: a YouTube video because we want to hear. And you know,
8:24
Speaker 2: in the autism community, YouTube is great. It cues up
8:26
Speaker 2: the same video as many times as you want. People
8:29
Speaker 2: always have their favorite video, she said, whatever their favorite
8:31
Speaker 2: video is queued up. At the one forty five mark,
8:34
Speaker 2: they would introduce a beach ball and buffering.
8:38
Speaker 1: Beach ball meaning the spinning wheel that says like pause,
8:41
Speaker 1: it's loading.
8:42
Speaker 2: Yeah, they would pause the video because what they wanted
8:45
Speaker 2: to do was get a vocalization of dissatisfaction and frustration,
8:50
Speaker 2: and so that was a way to do it. And
8:52
Speaker 2: they would say, okay, this is a five minute snack period.
8:55
Speaker 2: We want you to make sure that you are giving
8:57
Speaker 2: them their favorite sack, and we want to hear what
9:00
Speaker 2: that is like. And so they would do these things
9:02
Speaker 2: that are really relevant to her ultimate goal, which is
9:05
Speaker 2: making sure that these kids could be out there in
9:08
Speaker 2: the world talking to other people and having needs met.
9:11
Speaker 2: And so it was a very scientific protocol. And then
9:14
Speaker 2: she would send it to families along with a phone
9:19
Speaker 2: and a camera and basically say, record it from these
9:22
Speaker 2: various angles, send it back to us, so that the
9:24
Speaker 2: data is standardized. It's not just random vocalizations that they're
9:28
Speaker 2: figuring out. And because the population is so specific and
9:32
Speaker 2: because the keepers of the knowledge are so specific to
9:35
Speaker 2: parents and caregivers, it took her a while to figure out,
9:38
Speaker 2: that's how you actually want to get the data.
9:40
Speaker 1: Huh. Yeah, it sounds like a lot of work. It
9:44
Speaker 1: sounds hard to scale.
9:47
Speaker 2: Yeah, And it would be hard to scale if not
9:50
Speaker 2: for synthetic data, and if not for all of the
9:52
Speaker 2: advances on the other side with translation.
9:54
Speaker 1: Yeah, so you don't have to scale that much, is
9:57
Speaker 1: what those things.
9:58
Speaker 2: Well, you don't have to scale it nearly as much, right,
10:00
Speaker 2: And what would have been a project that would probably
10:02
Speaker 2: have happened over three scientists lifetimes yea, is now maybe
10:07
Speaker 2: only over half of her career. And that's a real
10:10
Speaker 2: form of progress. And so even though Christy doesn't have
10:13
Speaker 2: the kinds of funding that Google or open Ai have,
10:17
Speaker 2: she's able to kind of surf in their wake, and
10:20
Speaker 2: they've made these progressions on things like zero shot translation,
10:23
Speaker 2: with some people may have heard of. You know, historically
10:26
Speaker 2: we would translate things by taking one set of language,
10:30
Speaker 2: decoding each word and rewriting it in a different language,
10:34
Speaker 2: and up until about ten years ago, that's what machine
10:37
Speaker 2: translation was. Zero shot translation is actually a translation at
10:41
Speaker 2: such a deep level that you're not translating language to language,
10:45
Speaker 2: you're basically translating every language to every language simultaneously. And
10:50
Speaker 2: then we got to a certain point recently where we
10:52
Speaker 2: can translate based on audio waves. And so if you
10:56
Speaker 2: can translate audio waves, it becomes much easier to translate
11:01
Speaker 2: vocalizations from people who don't use words at all, and
11:05
Speaker 2: so she you know, you mentioned a lot of these
11:08
Speaker 2: stories aren't done. This is the least done story, right,
11:11
Speaker 2: And you know what's interesting is that technology has already
11:16
Speaker 2: changed the lives of lots of autistic people, because you know,
11:20
Speaker 2: she showed me her phone.
11:23
Speaker 1: Right.
11:23
Speaker 2: So we were having dinner on our porch, and her son, Felix,
11:27
Speaker 2: who's lovely sweet in many ways, a normal American teenager,
11:31
Speaker 2: just picked up the phone and went off to go
11:34
Speaker 2: do something. And then he came back and he dropped
11:36
Speaker 2: off the phone, and she and her husband showed me
11:38
Speaker 2: photos that he had taken of what he's fascinated by.
11:43
Speaker 2: And so some of them were hairbrush bristles that were
11:45
Speaker 2: actually these these kind of beautiful plastic forests. He'd used
11:49
Speaker 2: a filter and he'd taken them and hundreds of these photos.
11:53
Speaker 2: And before that, he'd taken photos of windows and was
11:56
Speaker 2: really interested in the way windows around their neighborhood worked,
11:59
Speaker 2: or castles or sketches of castles. And she said, like,
12:02
Speaker 2: before the iPhone, when he's staring at something, I would
12:07
Speaker 2: have no idea what he would have been daring. But
12:10
Speaker 2: here's this really rudimentary piece of technology, at least now
12:13
Speaker 2: and through that I'm able to understand what he's looking at.
12:17
Speaker 2: I don't necessarily understand why he's fascinated by it, but
12:21
Speaker 2: I know that he's interested. And I can make his
12:22
Speaker 2: life richer by exposing him to more of it, and
12:26
Speaker 2: I can frustrate him less by making sure there's always
12:28
Speaker 2: a hairbrush around. And so those kinds of things I
12:32
Speaker 2: have to say, like they make me more optimistic that
12:35
Speaker 2: if we can wrestle this technology, like you know, tech
12:39
Speaker 2: doesn't always have to be bad, Like we're not faded
12:41
Speaker 2: to the world sucking because technology is coming out, but
12:45
Speaker 2: we do have to figure out how we want it
12:46
Speaker 2: to work. And so I left Christy feeling like she's
12:50
Speaker 2: going to get there, and I know that that's not guaranteed,
12:54
Speaker 2: but she is so stubborn, she's so motivated by love
12:58
Speaker 2: and by problem solving, and the tech is moving so
13:01
Speaker 2: fast that she's been at this ten years, right, and
13:04
Speaker 2: ten years she's only gotten so far. But the next
13:07
Speaker 2: ten years are going to move so much faster, and
13:10
Speaker 2: more people are going to come and attack the problem.
13:11
Speaker 2: And it wouldn't surprise me if ten years from now
13:15
Speaker 2: she's successful.
13:18
Speaker 1: So another story you tell in your book is about
13:22
Speaker 1: Operation Warp Speed, and in particular it's about the logistics
13:26
Speaker 1: of getting vaccines out to the public. And you know,
13:29
Speaker 1: I knew a moderate amount about the vaccine development side,
13:33
Speaker 1: but I really didn't know anything about the getting the
13:36
Speaker 1: vaccine out to the world side. And I certainly didn't
13:39
Speaker 1: know that it was an AI story. So tell me
13:42
Speaker 1: about Operation Warp Speed, and in particular about this guy
13:44
Speaker 1: you write about, this general named Gus Purna.
13:49
Speaker 2: Look, I was writing, and I was hungry for understanding
13:53
Speaker 2: and trying to figure out how exactly machine learning and
13:56
Speaker 2: artificial intelligence integrate into society and ways are helpful, and
14:00
Speaker 2: so I kind of tripped across this video of Gus Perna,
14:03
Speaker 2: who oversaw Operation Warp Speed, who is not a technologist
14:08
Speaker 2: by any means, He's a logistician. He oversees, generally, you know,
14:12
Speaker 2: or did before he retired, the delivery of munitions and
14:15
Speaker 2: uniforms and food so that we can fight wars. And
14:18
Speaker 2: he was placed in charge of Operation Warp Speed and
14:21
Speaker 2: the delivery of it so that you know, hopefully Big
14:23
Speaker 2: Pharma can can execute on all these vaccines. But then
14:27
Speaker 2: what how do they get into plastic vials and how
14:30
Speaker 2: do those vials get delivered to everything from a CBS
14:33
Speaker 2: to a doctor's office to in rural areas, sometimes a
14:36
Speaker 2: general store that has no digital footprint. And so he
14:41
Speaker 2: shows up in Washington and discovers that there's no plan,
14:46
Speaker 2: like he is the plan. And so in Washington, like
14:50
Speaker 2: it or not, everything starts with consultants everything, and so
14:54
Speaker 2: people come in and they pitch him on things, and again,
14:57
Speaker 2: not a technical guy, but he keeps hearing these crazy
15:00
Speaker 2: pitches over the course of an hour. Somebody's like, we're
15:03
Speaker 2: going to put a medical blockchain together and in a
15:06
Speaker 2: couple of years, and he's like, no. Somebody else pitched
15:09
Speaker 2: like hardware, We'll just put a hardware device in every
15:12
Speaker 2: doctor's and he knows enough to say no. And finally
15:15
Speaker 2: he meets with these two reasonably seeing people who are like, look,
15:19
Speaker 2: you have a data problem. This can all be solved
15:21
Speaker 2: if we get the right number of data streams flowing
15:24
Speaker 2: into the right interface so that you can actually see
15:27
Speaker 2: everything from plastic vials, metal needles, refrigerated trucks because remember,
15:33
Speaker 2: half of the vaccines needed to be really cold refrigerated
15:37
Speaker 2: for delivery. To track which states might want Johnson and
15:41
Speaker 2: Johnson versus like you have to build a whole civilization
15:44
Speaker 2: to do this stuff. And they said, look, you can
15:46
Speaker 2: build a civilization in data, probably won't cost that much,
15:51
Speaker 2: but we will need you to solve the human problem,
15:54
Speaker 2: which is getting these various agencies to play ball, so
15:57
Speaker 2: we can actually get to the data, clean it up,
15:59
Speaker 2: and feed it to you. And so that's a machine
16:01
Speaker 2: learning story. And when I went down the rabbit hole
16:04
Speaker 2: of how it actually happened, it was really instructive because
16:07
Speaker 2: obviously the end state was a huge success, underrated.
16:11
Speaker 1: Every chance I get, I want to scream about how underrated,
16:15
Speaker 1: worse feed is weirdly politically orphaned, right, Like the Republicans
16:19
Speaker 1: decided they don't like vaccines even though they did it.
16:21
Speaker 1: It was like, as far as I'm concerned, the great
16:24
Speaker 1: success of Trump's first administration, and then because it was Trump,
16:27
Speaker 1: the Democrats can't say it was great.
16:28
Speaker 2: Right exactly, And so it's wandering out there, this great
16:31
Speaker 2: American success story that just got buried by the stupidity
16:34
Speaker 2: of our politics. But we did it, and we wouldn't
16:38
Speaker 2: have been able to do it without really successful machine
16:41
Speaker 2: learning and AI.
16:42
Speaker 1: So how did it work? Like why was that an
16:45
Speaker 1: AI problem? And what did they actually do.
16:46
Speaker 2: So it was a data problem, but also an actionable
16:49
Speaker 2: data problem. And so there's a company called Pounteer, very
16:54
Speaker 2: very provocative company. A lot of people respond poorly to
16:57
Speaker 2: its name, and it has this kind of mystical aura
17:01
Speaker 2: to it, in part because it was founded by Peter
17:03
Speaker 2: Chiel and co funded by the CIA, and in part
17:07
Speaker 2: because it's named for the stones in the Tolkien trilogy.
17:12
Speaker 2: But really all they do is plumbing.
17:14
Speaker 1: I am not reflexively anti tech or anti company. I
17:18
Speaker 1: have some amount of wariness toward Palenteer, I guess because
17:21
Speaker 1: I associate them with domestic surveillance at some level, like
17:26
Speaker 1: that piece of it I mean, but I don't know
17:28
Speaker 1: it that well. I've never interviewed somebody from Palenter. I
17:30
Speaker 1: haven't read a book about them. So, like, tell me
17:32
Speaker 1: about Palenteer as a company. I guess, first of all,
17:36
Speaker 1: like the side that makes people weary, and then I'm
17:39
Speaker 1: curious more generally how they operate, what differentiates them, what
17:42
Speaker 1: they actually do as a tech company.
17:44
Speaker 2: Yeah. So as a company, they're created in two thousand
17:48
Speaker 2: and three by two friends who have no business being friends.
17:50
Speaker 2: One is Peter Thiel, who I think most of your
17:53
Speaker 2: listeners really know about. He's a sort of largely libertarian Republican,
17:58
Speaker 2: big Trump donor, has very controversial stances on pretty much
18:02
Speaker 2: every possible.
18:03
Speaker 1: Issue, including the Antichrist, weirdly including the anti Christ.
18:07
Speaker 2: He really finds a way. His friend is a guy
18:10
Speaker 2: named Alex Karp, who is kind of his polar opposite.
18:12
Speaker 2: He's New York born, he's half black, half Jewish, largely
18:17
Speaker 2: identifies as a socialist. A big donor to Hillary Clinton
18:21
Speaker 2: and Kamala Harris. They met in law school. They argued
18:25
Speaker 2: all the time. They agreed about the need for this
18:28
Speaker 2: kind of technology, and the only other thing they agree
18:31
Speaker 2: on is that in general, the world is better when
18:34
Speaker 2: America is a strong actor and American values like freedom
18:38
Speaker 2: are or propagated throughout the world. So that's what they
18:40
Speaker 2: agree on.
18:41
Speaker 1: Do you think that the liberal antagonism toward Palenteer is
18:45
Speaker 1: misplaced or do you think it's valid in some ways?
18:48
Speaker 2: It's incredibly complicated because it has now grown to include
18:52
Speaker 2: not just surveillance, but Alex Karp's position on Israel, which
18:55
Speaker 2: is that Israel should be supported by Palanteer products, just
18:58
Speaker 2: as Ukraine should be. And so there's lots of different
19:02
Speaker 2: valances where people can find something they don't like about Palanteer.
19:05
Speaker 2: And by the way, I respect all of them.
19:08
Speaker 1: Yeah, and so just as a tech company, what is
19:12
Speaker 1: Palanteer's thing, What do they do? And in particular, what
19:15
Speaker 1: did they do in Operation Warp Speed to get vaccines
19:18
Speaker 1: out to the public.
19:20
Speaker 2: I spoke to their chief architect and he's like, look,
19:22
Speaker 2: we're the mole people of Silicon Valley. We take these
19:25
Speaker 2: data streams, we lay them out, we clean them, We
19:29
Speaker 2: integrate the data from various different places, so like, for instance,
19:32
Speaker 2: a CVS and a right aid and a mom and
19:35
Speaker 2: pop place and a trucking company, and we put them
19:38
Speaker 2: together so that on an iPad or a laptop someone
19:42
Speaker 2: could play the pandemic like a video game.
19:44
Speaker 1: Right, play it like a video game, meaning there's like
19:48
Speaker 1: a crazy dashboard, so you can see here's where the
19:50
Speaker 1: people are, and here's where the vials are, and here's
19:53
Speaker 1: where the vaccine is, and here's where the trucks are,
19:55
Speaker 1: and we need to get these trucks to this vaccine
19:57
Speaker 1: to these people. That's the exact video game you're playing.
19:59
Speaker 2: Yeah, and constantly up to date and constantly accurate, and
20:04
Speaker 2: literally an administrator can sit there and just move things
20:06
Speaker 2: around and know that they're actually going to the right place.
20:09
Speaker 1: Yes, a wildly hard problem. Like it's not sexy, like
20:12
Speaker 1: developing a vaccine, and we've heard a lot about mRNA vaccines,
20:16
Speaker 1: but like figuring out where the actual vaccine is is
20:20
Speaker 1: totally non trivial health.
20:22
Speaker 2: Yeah, it's a huge deal. And so there's two problems
20:24
Speaker 2: you're solving right. One is the technical problem, and that
20:27
Speaker 2: involves a lot of really dull, really unglamorous work, which
20:31
Speaker 2: is like trying to figure out why this pipe you
20:34
Speaker 2: know it's because it is pipe coming from a distribution
20:37
Speaker 2: center is broken, and so AI can attack that.
20:40
Speaker 1: When you say pipe, do you mean the pipe through
20:42
Speaker 1: which data is flowing? What do you mean pipe in
20:44
Speaker 1: this instance?
20:45
Speaker 2: So, to build a data pipeline is basically a work
20:49
Speaker 2: of code. But everybody's got lots of different code, and
20:52
Speaker 2: so when your data pipelines are built in many many
20:54
Speaker 2: different codes to many different standards, they have to be standardized.
20:59
Speaker 2: You have to be able to see, oh, okay, this
21:01
Speaker 2: data coming from CBS matches up to the data coming
21:04
Speaker 2: from the trucking company. It's clean, it's modern. Most of
21:08
Speaker 2: the that they were dealing with in fact was not
21:11
Speaker 2: modern in any way. It was created a decade ago
21:16
Speaker 2: using old code and often had bad information, and so
21:20
Speaker 2: you need to standardize it, you need to clean it.
21:22
Speaker 2: And then where it really gets fun is you can
21:25
Speaker 2: use those pieces of information to recommend actions to note
21:30
Speaker 2: that like, oh, you know you're overly distributing in one area,
21:33
Speaker 2: you could be moving here. In fact, this rural area,
21:37
Speaker 2: for instance, might really want the one shot doses, whereas
21:42
Speaker 2: an urban area might want too.
21:43
Speaker 1: And those suggestions might be coming from the machine learning model,
21:47
Speaker 1: from the AI basic.
21:48
Speaker 2: Correct, so that while you know, the very overtaxed humans
21:52
Speaker 2: dealing with political problems and interagency problems are on the phone,
21:56
Speaker 2: the AI is just kind of nudging them toward like
21:58
Speaker 2: easier ways to make this whole project happen. So like
22:02
Speaker 2: the technical is one problem, and then something that's kind
22:05
Speaker 2: of a light motif of your show is the humans
22:08
Speaker 2: are another. Because we create our own systems and there
22:12
Speaker 2: is nothing more complicated than the federal bureaucracy. And I
22:15
Speaker 2: mean that oftentimes with admiration and sometimes not with admiration.
22:19
Speaker 2: And so inside of the CDC are many different agencies,
22:24
Speaker 2: and a lot of them were even in COVID had
22:28
Speaker 2: very sharp elbows.
22:29
Speaker 1: Yeah.
22:29
Speaker 2: They didn't want the Department of Defense, which was nominally
22:32
Speaker 2: overseeing this effort, getting involved. They thought, no, this is
22:36
Speaker 2: our territory. What do you possibly know about this? And
22:40
Speaker 2: General Perna was so important because he really didn't know
22:44
Speaker 2: that much about technology. He really didn't know that much
22:46
Speaker 2: about vaccines. What he knew about was getting stuff done.
22:51
Speaker 1: Yeah. You tell the story of a moment when he's
22:54
Speaker 1: arguing with the governor about whether some amount of vaccine
22:58
Speaker 1: has been delivered to the governor state. Tell me, tell
23:01
Speaker 1: me about that.
23:02
Speaker 2: Yeah. So, once the vaccines started to get delivered, this
23:06
Speaker 2: is where the video game element, the dashboard Elmont really
23:08
Speaker 2: became an important because they could see trucks pulling up
23:12
Speaker 2: to a registered refrigeration point or warehouse point in every state.
23:16
Speaker 2: And there were states that were really eager to get
23:19
Speaker 2: the vaccines, and the governors felt that their entire political
23:23
Speaker 2: futures depended on getting it and distributing it and being
23:26
Speaker 2: able to tell their citizens like, hey, Massachusetts, we got it,
23:29
Speaker 2: we got it. You can go to these places. Remember,
23:31
Speaker 2: there are people, you know, frantically refreshing their browsers to
23:34
Speaker 2: try and find weights up to three four weeks, and
23:37
Speaker 2: so somebody would say, hey, you promised me this dosage.
23:41
Speaker 2: I don't have it. And he said, oh, but you do.
23:43
Speaker 2: Here's who signed for it. Go get them. I'm going
23:46
Speaker 2: to stay right here. You go get them. And so
23:49
Speaker 2: what he, you know, also did and he had plenty
23:51
Speaker 2: of people working for him who help with this. These
23:54
Speaker 2: systems need accountability. And what I think, what I saw
23:57
Speaker 2: time and again in reporting on the book, is that
24:00
Speaker 2: the accountability is human based. We respond to human beings saying, actually,
24:05
Speaker 2: I'm calling your bluff because I can see the answer here.
24:08
Speaker 1: Uh huh.
24:09
Speaker 2: Actually, I'm gonna follow up with you because you said
24:11
Speaker 2: you would do this pilot with us and you didn't.
24:13
Speaker 2: And if you don't have the right human in that chair,
24:15
Speaker 2: with the right temperament, the right force, the right seductive charm,
24:19
Speaker 2: and that's what's called for, it's going to fail. And
24:22
Speaker 2: so I can't say enough. Like my conversations with Perna,
24:26
Speaker 2: I just came away thinking we should put him in
24:28
Speaker 2: charge of a lot of stuff.
24:33
Speaker 1: We'll be back in just a minute. So let's talk
24:46
Speaker 1: about the Cleveland clinic. Your book has start of divided
24:48
Speaker 1: into sections. There's a warp speed section. There's a healthcare section.
24:52
Speaker 1: Healthcare sections basically the Cleveland Clinic. Why did you choose
24:56
Speaker 1: for your healthcare section the Cleveland Clinic?
25:00
Speaker 2: Mostly because they asked me to come in, and so
25:03
Speaker 2: I was.
25:04
Speaker 1: Really clicking phrase kings under the street light. Yeah.
25:07
Speaker 2: And also, listen, healthcare is a universe, right, So what
25:09
Speaker 2: I really wanted to explore is as much of that
25:12
Speaker 2: universe as possible. Cleveland Clinic is one of the top
25:14
Speaker 2: systems in the world and the United States, and they're
25:17
Speaker 2: run by a really interesting guy named doctor Thomas Mihaliovich.
25:21
Speaker 2: And so what he told me at the beginning is
25:24
Speaker 2: I want you to wander around because what we do
25:27
Speaker 2: with AI is going to be different than what a
25:28
Speaker 2: lot of other people do. First of all, we actually
25:31
Speaker 2: have one major priority, which is the treatment of our patients.
25:34
Speaker 2: And so that means that when technology comes to the
25:37
Speaker 2: Cleveland Clinic, it's not run by technologists. It's run by doctors.
25:41
Speaker 2: And that is going to eliminate certain partners right off
25:43
Speaker 2: the bat, because most software folks believe that software is
25:47
Speaker 2: the most powerful force in the world, and they don't
25:50
Speaker 2: want to explain themselves or work within some sort of
25:53
Speaker 2: old fashioned medical system. I spent a fair amount of
25:57
Speaker 2: time with the chief technical officer, who had come from
25:59
Speaker 2: outside of healthcare and was recruited specifically from outside of it,
26:03
Speaker 2: and he obviously knew tons about technology, but knew very
26:07
Speaker 2: little about healthcare.
26:09
Speaker 1: I'm somewhat frustrated in your book.
26:12
Speaker 2: Yeah, he is frustrated because he's used to being in charge,
26:16
Speaker 2: and what he learned in his first couple of years
26:19
Speaker 2: is like, I'm not in charge.
26:21
Speaker 1: And like every doctor is in charge in a weird way, right,
26:25
Speaker 1: one of the maybe distinctive things about healthcare. And I
26:28
Speaker 1: know doctors don't see it that way, but doctors have
26:30
Speaker 1: a lot of autonomy and a lot of them don't
26:33
Speaker 1: want AI.
26:35
Speaker 2: They have a ton of autonomy, but they also have
26:37
Speaker 2: a ton of authority. Doctor. You know, medicine is a
26:40
Speaker 2: hierarchical business, and you want it to be that way
26:43
Speaker 2: when you're the patient. You want someone who actually says,
26:45
Speaker 2: I see all the data, I have all the inputs,
26:48
Speaker 2: here's what we're going to go do, and then distributes
26:50
Speaker 2: the tasks back. Technology doesn't work all that well that way,
26:54
Speaker 2: and so ro heat Chandra came in and he's basically like.
26:58
Speaker 1: This is this is the chief technology this is yeah,
27:00
Speaker 1: this CTO, and.
27:02
Speaker 2: He's like, I'm supposed to help turn this place around
27:04
Speaker 2: with technology. The business is terrible. Our margins. You know,
27:08
Speaker 2: Clinton Clinic is actually one of the more successful nonprofit
27:10
Speaker 2: hospital systems, and it's got it two percent margin, which
27:14
Speaker 2: most people would find horrifying. And so he goes at
27:18
Speaker 2: it like a technologist, and time and again he runs
27:20
Speaker 2: into this wall, which is doctor saying well, I don't
27:24
Speaker 2: want to do it that way. Yeah, and sometimes with
27:26
Speaker 2: good reason and sometimes not. But again, you can't just
27:30
Speaker 2: solve the technical problem. If you want AI to actually
27:32
Speaker 2: work for good, you have to solve both the technical
27:35
Speaker 2: and the structural problem of what you're trying to fix.
27:38
Speaker 1: So actually, at the Cleveland Clinic, you encountered this company
27:41
Speaker 1: called Baesian Health, And as it happens, I interviewed the
27:45
Speaker 1: founder of that company, Suci Saria, a few years ago
27:47
Speaker 1: on this show, and they built this system where they
27:51
Speaker 1: use AI to flag when patients are likely developing sepsis.
27:56
Speaker 1: And you know, that was a few years ago that
27:58
Speaker 1: I talked to her, and so I'm curious, how's it going,
28:00
Speaker 1: How is it working in the world. At the Cleveland clinic.
28:03
Speaker 2: Yeah, so's she's brilliant. And I talked to her as well,
28:07
Speaker 2: and from a technical perspective of her ability to understand
28:10
Speaker 2: not just what sepsis is, but again to create a
28:13
Speaker 2: bunch of data streams and integrate them and create weights
28:16
Speaker 2: between the streams, to create a flag for doctors to say, hey,
28:20
Speaker 2: this might be sepsis. And just as a reminder, sepsis
28:24
Speaker 2: kills more people every year than breast cancer, prostate cancer,
28:28
Speaker 2: and the opioid crisis. Yeah, three hundred and fifty thousand
28:30
Speaker 2: Americans a year, and it happens in hospitals. And so
28:33
Speaker 2: the clinic, which loses in twenty twenty one, I think
28:37
Speaker 2: lost about three thousand patients to sepsis, said well, we can't.
28:40
Speaker 2: This is crazy. It may be the industry standard, but
28:43
Speaker 2: it's crazy. So they brought in beision. I will tell
28:46
Speaker 2: you that the founder, when first approached by the CTO
28:48
Speaker 2: of the Cleveland clinic, said yeah, I'm not sure we're interested.
28:53
Speaker 2: She said, you guys are a really big system. Your
28:55
Speaker 2: doctors are going to be resistant. You're kind of a pain.
28:59
Speaker 2: Prove to me that you actually want to implement this
29:01
Speaker 2: in your hospital. And the CTO, who has a similar
29:04
Speaker 2: attitude about medicine. Was actually like, oh, I see a
29:08
Speaker 2: kindred spirit. I'm going to talk up the fact that
29:11
Speaker 2: our doctors are committed and we're going to do this.
29:13
Speaker 2: And so they went into Cleveland Clinic and they really
29:17
Speaker 2: piloted for a year, and it was a hard year.
29:19
Speaker 2: They had to figure out how it works, not just
29:22
Speaker 2: in community hospitals in an emergency room in other places,
29:26
Speaker 2: but like in the ICU and the Cleveland Clinic. ICU
29:29
Speaker 2: is different even than most other healthcare ICUs because it's
29:35
Speaker 2: a world famous institution. People come in from all over
29:38
Speaker 2: the world, and a lot of sepsis looks exactly like
29:42
Speaker 2: a cardiac infection, looks exactly like sixteen other things that
29:45
Speaker 2: they're all flagging for. And so if you're an ICU
29:48
Speaker 2: and doctor or nurse, your job is basically responding to
29:51
Speaker 2: beeps coming from machinery near your patient. It can be
29:55
Speaker 2: a huge pain in the ass. And so when a
29:58
Speaker 2: sepsis flag isn't properly calibrated, when it flags too much
30:02
Speaker 2: or flags too little, it's actually not doing its job.
30:05
Speaker 2: So even though everyone may agree on the premise, which is, hey,
30:08
Speaker 2: we want to reduce more metality, from sepsis because it's
30:11
Speaker 2: twenty twenty six or whatever. You know, you think about
30:15
Speaker 2: your own job, You think about the own annoyances you
30:17
Speaker 2: have in getting your job done, and even though it's
30:20
Speaker 2: life or death, those annoyances mount. And so the software
30:24
Speaker 2: has to work within the way doctors work, or else
30:27
Speaker 2: it's not going to get used well.
30:28
Speaker 1: And I mean, just to be clear, it was a
30:31
Speaker 1: little bit unclear to me in the book how good
30:34
Speaker 1: the software is in that context, like.
30:38
Speaker 2: Is it good enough?
30:39
Speaker 1: Is part of the problem that there's too many false positives,
30:42
Speaker 1: which is a real problem.
30:44
Speaker 2: Not outside of the ICU. Okay, the ICU is the
30:47
Speaker 2: is the last frontier because everything is so complicated in
30:51
Speaker 2: an ICU.
30:52
Speaker 1: Yeah, it's the hardest diagnosis essentially exactly. So outside of
30:57
Speaker 1: the ICU, is there like empirical benefit, are fewer people
31:01
Speaker 1: dying of sepsis?
31:02
Speaker 2: Well? The clinic reported a forty one percent drop in
31:06
Speaker 2: mortality over the course of a year using the software. Now,
31:10
Speaker 2: some of it was because they had drawn everyone's attention
31:13
Speaker 2: to the reduction of sepsis. It's the Hawthorne effect when
31:16
Speaker 2: you say we're going to do this, and everyone's attention shifts.
31:20
Speaker 2: But people are realizing that that sepsis is a very
31:23
Speaker 2: perfect for AI kind of problem because it has all
31:26
Speaker 2: of these inputs that are sometimes conflicting and so much
31:30
Speaker 2: data that it can overwhelm a caregiver. And AI doesn't
31:34
Speaker 2: get overwhelmed by too much data. It's really great at
31:37
Speaker 2: processing data at scale, very very quickly. What it often
31:42
Speaker 2: isn't as great at is figuring out how to tell people.
31:45
Speaker 1: About it, especially probabilistically. Right that like correct, the AI
31:50
Speaker 1: is very good at pattern matching and making a probabilistic assessment,
31:55
Speaker 1: but that last piece seems like the hard to figure
31:59
Speaker 1: out for humans piece.
32:00
Speaker 2: Yeah, And so what Beijin did, much to their credit,
32:04
Speaker 2: is they heard the caregivers and they said, look, it's
32:07
Speaker 2: not good enough just to send a flag. I get
32:09
Speaker 2: flags all day, Tell me why explain why this flagged
32:13
Speaker 2: on this patient. And what Cleveland Clinic did in the ICU,
32:16
Speaker 2: particularly because they're very aware that this was the hardest place,
32:21
Speaker 2: they basically just set up a workstation. They called it
32:24
Speaker 2: a command center. Command center is a little glorious for
32:27
Speaker 2: what I saw.
32:27
Speaker 1: It was a PC.
32:29
Speaker 2: There's a lovely woman named Dana operating it. She would
32:32
Speaker 2: look at the patients in the ICU. She would see
32:34
Speaker 2: the flag, she would hover over the flag, and the
32:36
Speaker 2: flag would say I flagged this because I'm noticing X, Y,
32:41
Speaker 2: and Z, and so she would say, Okay, Well, what
32:45
Speaker 2: I'm seeing when I go look at the patient is
32:48
Speaker 2: they're actually okay, and I can see why the AI
32:51
Speaker 2: might be confused about this. I'll keep an eye on it.
32:54
Speaker 2: Other times it was very clear that like, yep, you're
32:57
Speaker 2: seeing what we're seeing. This is validating what we're seeing.
33:00
Speaker 2: And occasionally it would flag something and she would go
33:02
Speaker 2: in and say to the caregivers, we have to do
33:05
Speaker 2: a check right now.
33:07
Speaker 1: When you step back, having learned what you've learned from
33:11
Speaker 1: the Cleveland Clinic, when you think about healthcare and AI
33:14
Speaker 1: in the next few years, what do you think about I.
33:17
Speaker 2: Think that healthcare and AI are going to really beat
33:20
Speaker 2: like peanut butter and chocolate. I think it's the best
33:23
Speaker 2: place for AI to come into our lives and have
33:26
Speaker 2: rational uses that make our health care better and that
33:29
Speaker 2: make our costs lower. It is going to take change.
33:33
Speaker 2: And so one of the most interesting people I met
33:35
Speaker 2: at the clinic is this woman named doctor Rita Pappus,
33:38
Speaker 2: who's a hospitalist and hospitalists oversee the administration of the hospital,
33:42
Speaker 2: but from a medical point of view. And she started
33:45
Speaker 2: as a nurse and in her thirties went back to
33:48
Speaker 2: medical school, and so she's seen everything a hospital has
33:51
Speaker 2: to offer, including every type of human in a hospital.
33:55
Speaker 2: And one of the things she told me is like, look,
33:58
Speaker 2: I know that roheat the CTO is finding lots of
34:02
Speaker 2: resistance across the hospital, and guess what, I'm not the
34:06
Speaker 2: least bit surprised. Doctors are trained a very particular way
34:10
Speaker 2: and ultimately, like long term, two decades out, if we
34:13
Speaker 2: want AI to take hold in hospitals, we actually going
34:16
Speaker 2: to need different kinds of people to go into medicine
34:20
Speaker 2: who are less lowercase C conservative, less grounded in their ways,
34:25
Speaker 2: more collaborative, and more willing to trust that technology can
34:31
Speaker 2: do some of the job. The people that she sees
34:33
Speaker 2: going into the profession right now after hundreds of years
34:36
Speaker 2: of selection, are very particular type. And so she didn't
34:40
Speaker 2: say this in a value judgment. She's very objective about it,
34:43
Speaker 2: but she said, look, that's what it's going to take.
34:45
Speaker 2: We've already proven that it can work for our bottom line,
34:48
Speaker 2: that it can work for patients if we're going to
34:50
Speaker 2: scale it, education for doctors is going to have to
34:54
Speaker 2: change too.
34:55
Speaker 1: So at the end of the book you have this epilogue,
34:59
Speaker 1: that's the one part of the book where you're basically like,
35:04
Speaker 1: here's what you should do. You being a person who's
35:07
Speaker 1: reading the book, who cares about the world it's interested
35:09
Speaker 1: in AI, what should I do? What's your advice?
35:14
Speaker 2: Well, I think, look, there's two pieces of advice here, right.
35:18
Speaker 2: The first is it, if you're the kind of person
35:22
Speaker 2: who cares about living in a better world, you're going
35:26
Speaker 2: to have to use these tools to understand what they're
35:28
Speaker 2: capable of. And that responsibility is really frustrating to people. Right,
35:33
Speaker 2: We've now listened to technologists tell us that we have
35:36
Speaker 2: to change our lives for whatever their product is for
35:38
Speaker 2: the last twenty years, and it's annoying and they get richer,
35:43
Speaker 2: and we spend more time adapting, and you haven't seen
35:46
Speaker 2: a lot of public good come from it. But I'm
35:49
Speaker 2: pretty convinced that in the right situations, these tools can
35:52
Speaker 2: help us, can help their republic, can help education, medicine.
35:58
Speaker 2: But if we leave that to the technologists, we're going
36:01
Speaker 2: to get a bad result. So we need to figure
36:03
Speaker 2: out what good they can play in our lives and
36:06
Speaker 2: what kind of good we want to advocate for in
36:09
Speaker 2: the public space. And if it's not hard, there are
36:12
Speaker 2: lots of ways to understand what this stuff does. You
36:15
Speaker 2: can do it in a couple minutes, in a day,
36:17
Speaker 2: repeat it. It's a really important component to being a
36:22
Speaker 2: citizen in the twenty first century. And it's a lot
36:24
Speaker 2: of what the Pope wrote in his encyclical, which is
36:27
Speaker 2: we cannot wish these away, but we do have to
36:30
Speaker 2: make them more human. And the way to make them
36:31
Speaker 2: more human is to be insisting that they be more human.
36:34
Speaker 2: So you have to use them.
36:36
Speaker 1: When you say make them more human, what does that mean.
36:40
Speaker 2: Left to their own devices. This is going to be
36:42
Speaker 2: enterprise software. It's going to be sold into the Department
36:46
Speaker 2: of Defense. It's going to be used to eliminate jobs.
36:49
Speaker 2: It's going to be used to make corporations more efficient.
36:53
Speaker 2: And I think what we want to advocate for is, no,
36:56
Speaker 2: we want these tools to make many, many people's lives better.
37:01
Speaker 2: It's not to say that, you know, capitalism can't continue
37:04
Speaker 2: churning on, but like you have to advocate for what
37:07
Speaker 2: you want, and so in the same way, that mister
37:09
Speaker 2: Johnson's use of AI is sitting way downstream of what
37:13
Speaker 2: Google and open aiy ultimately want. We want to move
37:16
Speaker 2: those things upstream, and so the only way to do
37:19
Speaker 2: it is to know what the tools can do and
37:21
Speaker 2: then insist that we get that kind of utility because
37:26
Speaker 2: it's just not on the radar for these companies. So
37:29
Speaker 2: that's one component of it, and the other, frankly, is
37:32
Speaker 2: we probably need greater regulation for the negatives to prevent
37:36
Speaker 2: them from happening.
37:37
Speaker 1: Like what's good regulation?
37:39
Speaker 2: The best model I've heard so far is really, you know,
37:43
Speaker 2: kind of using nuclear power and nuclear weaponry as a model. Right,
37:49
Speaker 2: that was the last thing we created that has this
37:52
Speaker 2: kind of scale and power. The International Atomic Energy Commission
37:56
Speaker 2: was established to make sure that we knew exactly how
37:59
Speaker 2: much nuclear material there was in the world yea, and
38:01
Speaker 2: who had control of it and what it was being
38:03
Speaker 2: used for.
38:04
Speaker 1: I mean, like, how does that map to this? I
38:07
Speaker 1: get that it says big deal read flo but what
38:09
Speaker 1: does it actually mean.
38:10
Speaker 2: It means keeping an inventory of all of these models
38:14
Speaker 2: and their capabilities, knowing how much energy they're using, knowing
38:17
Speaker 2: how many chips they have, knowing how many floating point operations,
38:21
Speaker 2: which is the technical term for what these models are processing.
38:25
Speaker 2: Each model is doing. You could set a limit on
38:27
Speaker 2: the floating point operations and say beyond this model, you
38:31
Speaker 2: can't make it, at least for now.
38:33
Speaker 1: So you're talking about like a global speed limit on
38:38
Speaker 1: the development of bigger models.
38:40
Speaker 2: Say, I think it's probably the simplest way to go
38:43
Speaker 2: about regulating how AI works in the world and giving
38:48
Speaker 2: governments and societies time to reckon with the consequences of
38:51
Speaker 2: each new model.
38:52
Speaker 1: And you think it might happen, I think.
38:55
Speaker 2: Something akin to it is going to happen because I
38:58
Speaker 2: think politically actors are realizing in twenty twenty six, AI
39:03
Speaker 2: is likely to be one of the two or three
39:05
Speaker 2: biggest issues in the election, and by twenty twenty eight
39:08
Speaker 2: it's like to be the biggest. Look, it's America, and
39:12
Speaker 2: in the end, America tends to regulate after the crisis.
39:16
Speaker 2: That's historically what we do. If I had to beat
39:20
Speaker 2: I would say that's probably what will happen this time.
39:22
Speaker 2: Is something bad will happen and then we will all
39:25
Speaker 2: be looking at the ashes of it and figuring out
39:27
Speaker 2: what to do. I hope that's not the case.
39:35
Speaker 1: We'll be back in a minute with the lightning round.
39:46
Speaker 1: Let's finish with the lightning round.
39:48
Speaker 2: Yeah.
39:49
Speaker 1: Best Baltimore musician of all time.
39:53
Speaker 2: Tupac Shakur did not know.
39:57
Speaker 1: Baltimore.
39:57
Speaker 2: A student of Baltimore School for the Arts and the
40:00
Speaker 2: Baltimore Orioles recently gave out a Tupac Shakur bobblehead and
40:04
Speaker 2: the stadium was full.
40:05
Speaker 1: Were you there?
40:06
Speaker 2: I was not, but I contemplated.
40:09
Speaker 1: We were on the ground crew of the Orioles. Yeah.
40:11
Speaker 1: In high school, what'd you learn?
40:15
Speaker 2: I learned that the most disgusting thing in the world
40:18
Speaker 2: is a tobacco ball combined with bazooka, and that most
40:22
Speaker 2: relief pitchers would form that in their mouths, get the
40:26
Speaker 2: juices out, spit it out into the air, and then
40:29
Speaker 2: kick it into the outfield.
40:31
Speaker 1: Incredible. What's your walk on song?
40:38
Speaker 2: I really love the Rocky theme?
40:40
Speaker 1: Amazing a cliche for a reason.
40:43
Speaker 2: It absolutely is like, if you are working out and
40:46
Speaker 2: you need to go that last quarter mile, just put
40:50
Speaker 2: it on.
40:50
Speaker 1: I really respect that you went to that instead of
40:53
Speaker 1: some obscure hipster pick so right that you oversaw Times
40:59
Speaker 1: Person of the Year when you were at time I did.
41:02
Speaker 2: Yeah, what's a tip.
41:05
Speaker 1: If I want to be Times.
41:06
Speaker 2: Person of the Year, you know, it's really year dependent.
41:09
Speaker 3: So I, like a fool, thought we would just evaluate
41:14
Speaker 3: it based on the most important person of the year,
41:17
Speaker 3: and that is in that case, it turns out it's like, oh,
41:20
Speaker 3: but last year we did a bad person, so this.
41:23
Speaker 2: Year we need to do a good person.
41:25
Speaker 1: Well that makes some sense. There's a mix, it does.
41:28
Speaker 2: You want to mix, but it is not the pure
41:31
Speaker 2: thing that I was raised to believe.
41:32
Speaker 1: It is. Were you there? There was a year I
41:36
Speaker 1: remember exactly what it was. I don't know exactly when
41:39
Speaker 1: you were there when they put like something reflective on
41:41
Speaker 1: the cover in the Person of the Are was you?
41:43
Speaker 2: Was that you?
41:44
Speaker 1: Oh?
41:44
Speaker 2: I was there? It was not me. I thought this
41:49
Speaker 2: is pretty silly. You know, we could just make the
41:52
Speaker 2: YouTube guys, the people of the year.
41:54
Speaker 1: Failure is an orphan success as a thousand fathers failures. Yeah,
41:58
Speaker 1: that was happening right now.
41:59
Speaker 2: So I was no, no, no, I genuinely did not
42:01
Speaker 2: oversee that one.
42:02
Speaker 1: Is there nobody who would say I thought that was
42:04
Speaker 1: a good idea.
42:05
Speaker 2: I was asked forty eight hours before it went to
42:08
Speaker 2: print to write the back page, and so I wrote
42:11
Speaker 2: a backpage that, as I recall, was largely about Andy
42:15
Speaker 2: Warhol and the notion of you know, fifteen minutes and
42:18
Speaker 2: how that works in a distributed sort of technological I
42:21
Speaker 2: don't want to go back and look at it or
42:23
Speaker 2: defend it, but I will tell you it was not
42:25
Speaker 2: my call to put the reflective view on the cover.
42:29
Speaker 1: You know, everybody hated it at the time, and I
42:31
Speaker 1: get that. I mean, for anybody to care at all
42:34
Speaker 1: now it would be amazing. But it also, like you
42:37
Speaker 1: could argue it's sort of the fifteen minutes of famous,
42:39
Speaker 1: like the foreshadowing the creator economy or something, if you
42:42
Speaker 1: wanted that.
42:43
Speaker 2: I think others there have been others that have held
42:46
Speaker 2: up worse.
42:47
Speaker 1: If you were running a media company right now, what
42:51
Speaker 1: would your play be.
42:54
Speaker 2: I would really retreat to original reporting and high quality presentation,
43:00
Speaker 2: because I think what you're really competing is right now
43:02
Speaker 2: is a lot of noise and a lot of slop
43:05
Speaker 2: and a lot of volume. And when you see that
43:07
Speaker 2: going on, treat to quality and trust your taste because
43:12
Speaker 2: the individual the taste of individuals is still pretty interesting,
43:16
Speaker 2: I think. But I wouldn't try to compete everywhere. I
43:19
Speaker 2: would try to compete on the most interesting stories and
43:22
Speaker 2: make them, you know, keep it simple, make them as
43:24
Speaker 2: good as you can possibly make them.
43:27
Speaker 1: Thanks Mian, congratulations on the book.
43:29
Speaker 2: Thank you so much.
43:38
Speaker 1: Josh Tarrengel is the author of the book AI for Good,
43:41
Speaker 1: How real people are using artificial intelligence to fix things
43:45
Speaker 1: that matter. Please let us know what you think of
43:49
Speaker 1: the show, which you want to hear more of, which
43:51
Speaker 1: you want to hear less of particular guest ideas. You
43:55
Speaker 1: can email us at problem at pushkin dot fm. I
43:58
Speaker 1: read all the emails. You can also find me on
44:01
Speaker 1: x run LinkedIn. Really do appreciate all the messages that
44:06
Speaker 1: we get. Today's show was produced by Gabriel Hunter Chang
44:08
Speaker 1: and The Menino. It was engineered by Hansdale She and
44:11
Speaker 1: edited by Lydia Jane Kott. I'm Jacob Goldstein and we'll
44:15
Speaker 1: be back next week with another episode of What's Your Problem.