Latest / The Right Time with Bomani Jones / Why AI & Their CEOS CAN NOT be trusted | 8.11

Transcript

Why AI & Their CEOS CAN NOT be trusted | 8.11 · The Right Time with Bomani Jones

• 0:01
  Speaker 1: Wave. Ladies and gentlemen, welcome to the right time. A
• 0:12
  Speaker 1: wave original. My name is Bomani Jones. Thanks for listening.
• 0:15
  Speaker 1: Wherever you get your podcast. Thanks for watching us on YouTube. Subscribe, like,
• 0:20
  Speaker 1: rate us, review us, give us five stars. You only
• 0:22
  Speaker 1: give us four stars. I'm inclined to believe you are
• 0:25
  Speaker 1: a hater. Now, it is the time of week where
• 0:27
  Speaker 1: we have a guest join us. And one of my
• 0:30
  Speaker 1: favorite subsets of guests, obviously, is Game Theory with Bomani Jones.
• 0:35
  Speaker 1: alumni and joining us now he he worked as a
• 0:38
  Speaker 1: field producer on the right time he has been a
• 0:41
  Speaker 1: showrunner done things in other capacities worked on the daily
• 0:43
  Speaker 1: show and very recently worked as a producer and i
• 0:46
  Speaker 1: want to make sure i get the full title of
• 0:48
  Speaker 1: this documentary correct because it important the ai doc or
• 0:52
  Speaker 1: how i became an apocalyptic.
• 0:56
  Speaker 2: Very good very good.
• 0:58
  Speaker 1: That was a mouthful ladies and gentlemen uh costco enthusiast
• 1:02
  Speaker 1: ted tremper how are you.
• 1:03
  Speaker 2: Man I'm very good. I'm very good. I appreciate you
• 1:05
  Speaker 2: wearing the Seattle Sonics colors for me.
• 1:08
  Speaker 1: I did that just for you.
• 1:10
  Speaker 2: Are we getting the team back? I've heard for years
• 1:12
  Speaker 2: we're getting a Vegas team and a Seattle team. Is
• 1:14
  Speaker 2: that real?
• 1:14
  Speaker 1: I think you'll get a team that will then be
• 1:16
  Speaker 1: called the Sonics, and you guys will act like all
• 1:18
  Speaker 1: this anger that you've been riding with for the last
• 1:20
  Speaker 1: 20 years.
• 1:22
  Speaker 2: Let me tell you, I have not watched the basketball
• 1:24
  Speaker 2: game since they left. And I believe you.
• 1:27
  Speaker 1: You're the person who says that, and I firmly believe
• 1:29
  Speaker 1: that you have decided to take this stand. And I
• 1:32
  Speaker 1: want to say this right fast before we get going
• 1:34
  Speaker 1: so people know. One of the interesting moments of my
• 1:37
  Speaker 1: life is the first day that we were in the
• 1:40
  Speaker 1: office for Game Theory. And you got to remember, it's 2022.
• 1:44
  Speaker 1: And with the way a television show works, we were
• 1:45
  Speaker 1: still pretty in the thick of COVID. We were a
• 1:48
  Speaker 1: very masked up audience. And We did the meet and
• 1:52
  Speaker 1: greet meeting with the whole staff. But the first meeting
• 1:54
  Speaker 1: that I had with people in the room and the
• 1:56
  Speaker 1: first meeting of my life that I've ever been in
• 1:58
  Speaker 1: charge of, was a meeting sketching out something that we
• 2:02
  Speaker 1: did where we created a museum exhibit about Duke basketball.
• 2:06
  Speaker 1: And Ted Tremper, who is kind of the project manager
• 2:09
  Speaker 1: for this, I suppose is the term, is on the
• 2:11
  Speaker 1: screen ahead of me. And the schematics of this are
• 2:14
  Speaker 1: being laid out like it's the most amazing thing I've
• 2:16
  Speaker 1: ever had. I got my man Rod. I got Sidney Castile.
• 2:19
  Speaker 1: They were working on like filling it up with content
• 2:22
  Speaker 1: and stuff. and we're all there, and it's the first
• 2:24
  Speaker 1: time I've ever been in charge of a meeting, and
• 2:26
  Speaker 1: Ted is breaking down what everything is, and I have
• 2:29
  Speaker 1: this giant smile on my face. Meanwhile, the guys next
• 2:32
  Speaker 1: to me are petrified and nervous if I like it,
• 2:35
  Speaker 1: because they can't tell, because I had a mask on.
• 2:38
  Speaker 1: So I am thinking that I am radiating all this. Wow,
• 2:43
  Speaker 1: this is so cool. And nobody can see it. Intense
• 2:47
  Speaker 1: energy guy and he's laying it out. He's talking about
• 2:49
  Speaker 1: stuff I didn't even know was possible. He's using schematic
• 2:51
  Speaker 1: terms that don't mean shit to me. But I'm just
• 2:53
  Speaker 1: riding this out and I'm so gassed up and everybody
• 2:56
  Speaker 1: thinks I'm mad because nobody can see my face.
• 3:01
  Speaker 2: You know, that's the same thing. I proposed to my
• 3:04
  Speaker 2: wife while I was, or I asked for permission to
• 3:06
  Speaker 2: propose to my wife while my future father-in-law was wearing
• 3:09
  Speaker 2: a mask. And he didn't give me an answer because
• 3:11
  Speaker 2: we were in an airport and it was during COVID.
• 3:13
  Speaker 2: I didn't know I was going to see him again.
• 3:15
  Speaker 2: And he just told me, I'm not going to give
• 3:17
  Speaker 2: your answer now. I'm going to make you sweat a
• 3:18
  Speaker 2: little bit. And it turned out that he didn't want
• 3:20
  Speaker 2: to give me an answer because he didn't want to
• 3:22
  Speaker 2: cry because then it would reveal to his wife and
• 3:25
  Speaker 2: my now wife that this had happened. But I just
• 3:29
  Speaker 2: thought he hated me. And I thought we had a good,
• 3:31
  Speaker 2: we had a great rapport. I thought we were cool.
• 3:34
  Speaker 2: And no, yeah, I'm glad. I'm glad to see your smile.
• 3:37
  Speaker 2: I'm glad to see, I'm glad that we pulled through. Yeah.
• 3:40
  Speaker 1: And Ted also was the man behind all of our
• 3:44
  Speaker 1: black and white shots on the street that the first
• 3:46
  Speaker 1: time we did was the coldest I'd ever been in
• 3:48
  Speaker 1: my life. And- every every time we shot by the
• 3:53
  Speaker 1: way except for like the literal last one was yes
• 3:56
  Speaker 1: terrible frigid weather and it's so weird when i go
• 3:58
  Speaker 1: down to the west village now like i get off
• 4:00
  Speaker 1: at west fourth street and i'm like why does this
• 4:01
  Speaker 1: look familiar oh yeah this is where we did all
• 4:03
  Speaker 1: the shooting um but with the first time we did
• 4:07
  Speaker 1: it i was it was the worst i don't even
• 4:10
  Speaker 1: know if you had been brought on board yet because
• 4:11
  Speaker 1: i was out there with stew and like oh yeah
• 4:14
  Speaker 1: it was so cold it was so terrible and And
• 4:16
  Speaker 1: then people said the thing that I was fearing that
• 4:18
  Speaker 1: they would say, that's the best part of the show.
• 4:21
  Speaker 1: It's like, shit, we got to get out here and
• 4:22
  Speaker 1: do this every day, huh? Every week.
• 4:24
  Speaker 2: Two things. I think this is the first time we've
• 4:26
  Speaker 2: ever spoken in summertime, number one. And number two, you
• 4:29
  Speaker 2: do have to tell people, what did I bring? I
• 4:31
  Speaker 2: picked up at a Costco driving across the country. I
• 4:34
  Speaker 2: brought the heated gloves and the heated socks. That's how
• 4:38
  Speaker 2: much I care.
• 4:39
  Speaker 1: Ted had all the things in a way that like,
• 4:43
  Speaker 1: Once you break up with a girlfriend, you realize now
• 4:46
  Speaker 1: you have to do all these things for yourself. Ted
• 4:49
  Speaker 1: was prepared and shared the wealth. I imagine in other situations,
• 4:52
  Speaker 1: you're in the same place. I find myself now among
• 4:55
  Speaker 1: my friends being the person that has the things because
• 4:58
  Speaker 1: once somebody has the things, now you have to have
• 4:59
  Speaker 1: the things. And I still have those socks with the
• 5:02
  Speaker 1: battery pack to keep my feet warm.
• 5:04
  Speaker 2: Everybody knows in LA, when it really starts collapsing, come over.
• 5:10
  Speaker 2: I got everything. I got us all in the basement.
• 5:13
  Speaker 1: I got.
• 5:14
  Speaker 2: It's all set up. It's all set up. This is
• 5:17
  Speaker 2: a good segue into the end of the world. Yes,
• 5:19
  Speaker 2: it really is.
• 5:20
  Speaker 1: It really is. So I just have to do that
• 5:22
  Speaker 1: because that was such a great time in my life.
• 5:24
  Speaker 1: And Ted, the stuff that we did with Ted was
• 5:26
  Speaker 1: so excellent to go back and watch. And so if
• 5:28
  Speaker 1: you guys can find it, I still get residual checks.
• 5:30
  Speaker 1: So apparently somewhere you can go watch some of this show,
• 5:33
  Speaker 1: check it out. But Ted worked on the AI doc,
• 5:36
  Speaker 1: which I think was the best way to put it,
• 5:37
  Speaker 1: because in this day and age, when these docs wind
• 5:39
  Speaker 1: up on streaming, They just become the blank doc for whoever, like,
• 5:43
  Speaker 1: yo man, have you seen the, like Chris Rock called
• 5:45
  Speaker 1: it the prison doc on HBO back on Bring the
• 5:47
  Speaker 1: Pain 30 years ago, but it is the AI doc.
• 5:50
  Speaker 1: And I watched it and I've watched all the doom
• 5:52
  Speaker 1: and gloom. I've read all the doom and gloom AI books,
• 5:54
  Speaker 1: but I wanted to start with you first, before we
• 5:56
  Speaker 1: get into like really the specifics of it. And you
• 5:58
  Speaker 1: said you basically spent two and a half years in
• 6:00
  Speaker 1: a room working on this. Where were you on the
• 6:04
  Speaker 1: topic of AI before you spent two and a half
• 6:06
  Speaker 1: years in the room?
• 6:07
  Speaker 2: It, It actually, I remember walking from the office back
• 6:10
  Speaker 2: to my Airbnb on our show and my friend Daniel
• 6:12
  Speaker 2: Kwan called me up and he said, Hey, I think
• 6:14
  Speaker 2: you should watch this thing. And it was, it was
• 6:16
  Speaker 2: a YouTube video that was called the AI dilemma that
• 6:19
  Speaker 2: Tristan Harris and Aza Raskin from the center of humane
• 6:21
  Speaker 2: technology had released in late 2022. And it was foretelling
• 6:25
  Speaker 2: basically everything that we're seeing now. It was the rise
• 6:27
  Speaker 2: of large language models and their exponential increase in capabilities
• 6:31
  Speaker 2: and all the different impacts that was going to have
• 6:33
  Speaker 2: on society and Kwan had shared it. And also Daniel Kwan,
• 6:36
  Speaker 2: by the way, he was the co-writer and co-director of
• 6:39
  Speaker 2: Everything Everywhere All at Once. This is before they swept
• 6:41
  Speaker 2: the Oscars. But we've been friends for about 15 years.
• 6:44
  Speaker 2: And he knows I'm a nerd. He's a nerd. And
• 6:46
  Speaker 2: we were in a video game group together. And he
• 6:48
  Speaker 2: kind of threw that into the chat. And that became
• 6:52
  Speaker 2: this thing where after they won all the Oscars, they
• 6:54
  Speaker 2: could meet with anybody. So they said, we want to
• 6:56
  Speaker 2: meet with these two guys. And basically, the plan was,
• 6:59
  Speaker 2: if you remember the film The Day After about nuclear annihilation,
• 7:03
  Speaker 2: That was kind of a wake-up call to society as
• 7:06
  Speaker 2: to why we need to focus on nuclear nonproliferation and deproliferation.
• 7:11
  Speaker 2: And the goal was, can we make a documentary that
• 7:13
  Speaker 2: will have a similar impact on society vis-a-vis the way
• 7:17
  Speaker 2: that AI is being developed and deployed? So that was
• 7:19
  Speaker 2: the impetus that made them get the team that had
• 7:22
  Speaker 2: made Navalny, which won the Oscar the same year that
• 7:24
  Speaker 2: they won. So they put together this team. I was
• 7:27
  Speaker 2: asked to jump on. I was going to stay for
• 7:28
  Speaker 2: five or six weeks to help book some people. And,
• 7:32
  Speaker 2: you know, we reached out to 100 people in AI
• 7:34
  Speaker 2: land and six responded. And that was the moment we
• 7:37
  Speaker 2: realized that people in AI don't really give a shit
• 7:39
  Speaker 2: about movies. And so that became this odyssey for me
• 7:44
  Speaker 2: where I was the tip of the spear. I had
• 7:46
  Speaker 2: to talk to these researchers. And in order to talk
• 7:49
  Speaker 2: to them, I had to learn what they did. And
• 7:51
  Speaker 2: in order to know what they did, I had to
• 7:52
  Speaker 2: learn how the tech worked. So it became this like,
• 7:55
  Speaker 2: you know me very well. I don't give up and
• 7:58
  Speaker 2: I don't stop. And the challenge was, I'm used to sprinting,
• 8:02
  Speaker 2: like nightly, weekly TV is a sprint. And what the
• 8:05
  Speaker 2: experience for me was, I can sprint for five or
• 8:08
  Speaker 2: six weeks. But what happened at the end of that
• 8:10
  Speaker 2: five or six weeks was, you know, I've got my
• 8:13
  Speaker 2: sprinting cleats on, I get to what I think is
• 8:15
  Speaker 2: the 100 yard line where we're done, a door opens up,
• 8:17
  Speaker 2: and there's just more road. And then it's like, well,
• 8:20
  Speaker 2: we'll be done in eight months. So we can be
• 8:21
  Speaker 2: we can be done before the election. And then eight
• 8:23
  Speaker 2: months turns to a year turns to a year and
• 8:25
  Speaker 2: a half turns to two years turns to two and
• 8:27
  Speaker 2: a half years. And it damn near killed me. But
• 8:30
  Speaker 2: At the end of the day, we interviewed more than
• 8:32
  Speaker 2: 40 people on camera. I interviewed more than 100 people
• 8:35
  Speaker 2: on background. I developed confidential sources inside of every single lab.
• 8:39
  Speaker 2: It was either current or former employees of every frontier lab.
• 8:42
  Speaker 2: So now I have what I never claimed to be
• 8:46
  Speaker 2: an AI expert, but I have what's called interpretive expertise,
• 8:49
  Speaker 2: which is I know I think I have a good
• 8:52
  Speaker 2: bearing on what every single different kind of expert would
• 8:54
  Speaker 2: say in this landscape. So, yeah, life's different. Yeah.
• 9:00
  Speaker 1: So the first thing that I find interesting about Ted
• 9:02
  Speaker 1: saying that he thought it would be this and then
• 9:04
  Speaker 1: it kept going to this is that you offered once
• 9:06
  Speaker 1: an accurate interpretation of me that helped me understand and
• 9:09
  Speaker 1: explain myself to people, which is he just wants to
• 9:12
  Speaker 1: know what the ask is. And that is 100 percent correct.
• 9:15
  Speaker 1: If you tell me we're going to be here for
• 9:17
  Speaker 1: 10 minutes, fine. If you tell me we're going to
• 9:19
  Speaker 1: be here for 100 minutes, fine. But if you tell
• 9:21
  Speaker 1: me it's 10 and it's 11, that last minute is
• 9:23
  Speaker 1: going to suck. Tell me it's going to be the
• 9:26
  Speaker 1: last one's going to suck. So for you, why did
• 9:29
  Speaker 1: this keep pushing out?
• 9:30
  Speaker 2: Well, let me know a secret you need to know
• 9:33
  Speaker 2: about me is the reason I saw that in you
• 9:34
  Speaker 2: is because that's the way I am. If somebody if
• 9:37
  Speaker 2: they had said, hey, do you want to work for
• 9:38
  Speaker 2: two and a half years on an A.I. documentary? I
• 9:40
  Speaker 2: would have said, sure, let me know what the money is,
• 9:42
  Speaker 2: because what happened was the moment you go from a
• 9:45
  Speaker 2: weekly to a project fee. every day you work over
• 9:48
  Speaker 2: eight months is a day that's lowering your rate. And
• 9:51
  Speaker 2: so when that doubles in a half, then the math
• 9:54
  Speaker 2: doesn't matter. But no, it really came down to wanting
• 10:00
  Speaker 2: to get it right. And the challenge of this film
• 10:02
  Speaker 2: was how can you make a documentary about the most
• 10:04
  Speaker 2: complex subject that maybe humanity has ever faced, but make
• 10:07
  Speaker 2: it funny, make it entertaining, make it human, make it
• 10:11
  Speaker 2: not technical to a degree that any person, regardless of
• 10:14
  Speaker 2: what their background is, if they've won a Nobel Prize
• 10:16
  Speaker 2: or if they've never talked about AI ever, can walk
• 10:18
  Speaker 2: away with the same intuitions about the way the technology works.
• 10:21
  Speaker 2: And I think that's what we did. The most common
• 10:24
  Speaker 2: comment that we've gotten about the film, which is also
• 10:25
  Speaker 2: the most flattering, which is I didn't expect a documentary
• 10:28
  Speaker 2: about AI to make me laugh and make me cry.
• 10:30
  Speaker 2: That's it.
• 10:31
  Speaker 1: So let's get to the film itself, right? Because I
• 10:34
  Speaker 1: think We are beginning to see, and I actually do
• 10:37
  Speaker 1: find this to be a bit of encouraging, that the
• 10:40
  Speaker 1: fomenting revolution of humans, now when people say they will
• 10:44
  Speaker 1: not replace us, it's not nearly as scary because they're
• 10:47
  Speaker 1: talking about computers.
• 10:49
  Speaker 2: Right.
• 10:50
  Speaker 1: When they throw that one out there now, it's like, oh, OK, cool.
• 10:52
  Speaker 1: We can we can we can all do this. But
• 10:55
  Speaker 1: I I thought the doc was important because I don't
• 10:58
  Speaker 1: think people really understand how AI works. And I understand
• 11:02
  Speaker 1: like I think it is a it is a vast
• 11:04
  Speaker 1: black box, a fascinating one and one that I understand
• 11:08
  Speaker 1: why the nerds get so fascinated by what could happen.
• 11:11
  Speaker 1: But also the point of the film is no one
• 11:13
  Speaker 1: really seemed to stop and ask, now what? And correct
• 11:17
  Speaker 1: me if I'm wrong, but that seems to be the
• 11:18
  Speaker 1: overarching theme of this is getting to the point of
• 11:21
  Speaker 1: the now what that we're at now and what the
• 11:23
  Speaker 1: now what will be if it goes a little farther. Correct.
• 11:26
  Speaker 2: Yeah, I think the major thing and having, and there
• 11:30
  Speaker 2: was a time where I had listened to or read
• 11:33
  Speaker 2: every single word that Sam Altman had said and Dario
• 11:35
  Speaker 2: Amadei and all the five CEOs. And the most shocking
• 11:39
  Speaker 2: part of it to me was, that they are very
• 11:42
  Speaker 2: clear that they don't have a plan. Remember, it's actually,
• 11:45
  Speaker 2: in the most recent election, when Trump was asked about
• 11:50
  Speaker 2: healthcare and they asked, what's your healthcare plan? And he
• 11:52
  Speaker 2: didn't say he had a plan. He said they had
• 11:53
  Speaker 2: ideas about a plan. And that's where we're at. They
• 11:56
  Speaker 2: have ideas about a plan. The challenge is, and this
• 11:59
  Speaker 2: goes back to the fact that maybe they don't watch
• 12:01
  Speaker 2: enough movies, but if you remember Jurassic Park, they spent
• 12:04
  Speaker 2: so much time asking whether they could that they didn't
• 12:06
  Speaker 2: ask whether or not they should. The major thing to
• 12:09
  Speaker 2: understand about the CEOs is, They do not look at this,
• 12:13
  Speaker 2: but this is my assessment. They don't look at it
• 12:15
  Speaker 2: as something that they are building. They look at this
• 12:18
  Speaker 2: as a discovery that has been made that if they
• 12:21
  Speaker 2: don't build it, someone else who maybe doesn't have a
• 12:24
  Speaker 2: strong moral compass as they do will build it. And
• 12:27
  Speaker 2: the challenge there is it puts you into a Lord
• 12:29
  Speaker 2: of the Rings situation where everyone believes that they're Frodo.
• 12:33
  Speaker 2: And if you know anything about Frodo, the only thing
• 12:35
  Speaker 2: that makes Frodo Frodo is he doesn't think he's Frodo.
• 12:38
  Speaker 2: He doesn't think he should be the one to be
• 12:39
  Speaker 2: doing it. So the moment you start thinking, oh, well,
• 12:42
  Speaker 2: it has to be me, you've already revealed yourself to
• 12:45
  Speaker 2: be the kind of person who probably shouldn't be the
• 12:48
  Speaker 2: one who's doing it. And that's the challenge. And there
• 12:53
  Speaker 2: are some of the CEOs who really are focused on
• 12:56
  Speaker 2: trying to scale things to a point where they can
• 12:59
  Speaker 2: then take the final steps together as part of an
• 13:01
  Speaker 2: international coalition, et cetera. But the race is so strong.
• 13:04
  Speaker 2: And now you've got over a trillion dollars in debt.
• 13:07
  Speaker 2: that is part of all these interwoven, very incestuous deals
• 13:12
  Speaker 2: that are propping the economy up. And now you have
• 13:14
  Speaker 2: to really start to ask some questions about where we
• 13:17
  Speaker 2: as a society want to go. And there are no
• 13:21
  Speaker 2: clear answers as to where we will go. And it changes.
• 13:25
  Speaker 2: As somebody who's been on this beat now for three years,
• 13:27
  Speaker 2: it's an impossible job to keep up with it. And
• 13:29
  Speaker 2: I know that as a person who's tried. And The
• 13:32
  Speaker 2: fact that they are being surprised by it tells you
• 13:35
  Speaker 2: another enormous thing, which is they don't know where it's going.
• 13:37
  Speaker 2: So the good news is, to your point, I read
• 13:41
  Speaker 2: a statistic, the deciding factor as to whether or not
• 13:44
  Speaker 2: you are optimistic or pessimistic about AI is whether or
• 13:46
  Speaker 2: not you make above or below $ 200, 000 a year. And
• 13:49
  Speaker 2: most Americans don't make above $ 200, 000 a year, which means
• 13:53
  Speaker 2: that you have other statistics, only 18% of young people
• 13:55
  Speaker 2: from 14 to, I think, 29 feel hopeful about AI.
• 13:57
  Speaker 2: And When the younger generation is looking at the ways
• 14:04
  Speaker 2: that data centers are, and it's just, the other thing
• 14:06
  Speaker 2: is it's so cross-partisan. When you go to Tennessee and
• 14:09
  Speaker 2: you start taking people's water away and you start multiplying
• 14:12
  Speaker 2: their energy rates by 3x, 4x, 5x, it doesn't matter
• 14:16
  Speaker 2: who you voted for. You're going to start making phone calls.
• 14:20
  Speaker 2: You're going to start showing up at town halls. And
• 14:21
  Speaker 2: that's kind of.
• 14:22
  Speaker 1: Where we're at. Yeah. I want to put a pin
• 14:24
  Speaker 1: in the economy point because I think a lot of
• 14:25
  Speaker 1: people probably saw the Larry Ellison piece that was in
• 14:28
  Speaker 1: the New York Times over the weekend, which was, terrifying.
• 14:32
  Speaker 1: I will explain to you guys a little later why
• 14:34
  Speaker 1: it was so terrifying. But one of the books that
• 14:38
  Speaker 1: I read on AI, and I hate that I'm at
• 14:40
  Speaker 1: this point where I just read all this stuff and
• 14:41
  Speaker 1: I can't remember what it is, but I want to
• 14:42
  Speaker 1: make sure people know I didn't come up with this
• 14:44
  Speaker 1: shit myself, right? Just offer a little humility in the process.
• 14:48
  Speaker 1: The growth rate of AI is part of what's interesting.
• 14:51
  Speaker 1: And I think there are two components, which is No
• 14:54
  Speaker 1: one actually knows how quickly the technology will grow and
• 14:58
  Speaker 1: no one actually knows how it will grow within that speed.
• 15:03
  Speaker 1: So one thing, the idea is that in 15 years,
• 15:09
  Speaker 1: we could be here or we could be way farther
• 15:11
  Speaker 1: along than we thought. Like one of these incarnations of
• 15:14
  Speaker 1: Chad GPT was a mind blower for people that it
• 15:17
  Speaker 1: had gotten so good so fast. But the other part is,
• 15:20
  Speaker 1: and I found this fascinating in learning about AI, developing
• 15:25
  Speaker 1: an AI algorithm is like raising a child in the
• 15:28
  Speaker 1: sense that you can impart it with whatever values and
• 15:31
  Speaker 1: things that you think are important. But at the end
• 15:33
  Speaker 1: of the day, that shit is out of your hands
• 15:35
  Speaker 1: and that kid is going to be what that kid
• 15:37
  Speaker 1: is going to be. And all these models, I always figured, right?
• 15:41
  Speaker 1: It wasn't until that first time that model, that damn
• 15:43
  Speaker 1: Chad GBT flat out lied to me about something. Just
• 15:47
  Speaker 1: flat out lied. And I was like, I can't believe
• 15:49
  Speaker 1: this is possible because I just figured that it was
• 15:51
  Speaker 1: basically such a flowchart one in zero, ones in zero
• 15:54
  Speaker 1: sort of situation. I didn't think this thing could lie
• 15:57
  Speaker 1: to me. But that bad boy lied in my face
• 15:59
  Speaker 1: about something because it was trained to not tell me
• 16:02
  Speaker 1: that it couldn't do an operation that I asked it
• 16:04
  Speaker 1: to do. And then you realize, oh, no, no, no, no, no, no, no. This...
• 16:09
  Speaker 1: Is it too late? I guess is the question because
• 16:11
  Speaker 1: it feels like they've gone so far and we've done
• 16:13
  Speaker 1: so much. And once you realize that the algorithms actually
• 16:16
  Speaker 1: kind of control themselves, is it too late?
• 16:20
  Speaker 2: So, so I think there's a couple of important things
• 16:22
  Speaker 2: to unpack there. Like the, the, the, the challenge when
• 16:24
  Speaker 2: people are building this, like there's a huge incentive to
• 16:26
  Speaker 2: anthropomorphize these things, because if you can make it believe
• 16:29
  Speaker 2: like it's an organism, then you remove culpability. That if
• 16:32
  Speaker 2: I can, if I can really convince everyone, well, this
• 16:34
  Speaker 2: is an organic thing. Like it's, it's, it's creating its
• 16:37
  Speaker 2: own connections, et cetera. That is true on a technical level,
• 16:40
  Speaker 2: but the really important thing is to remember that there
• 16:42
  Speaker 2: are human beings who are designing these things. Human beings
• 16:45
  Speaker 2: are creating the circumstances that they can do these tests
• 16:49
  Speaker 2: where it is being trained and learning and can become
• 16:54
  Speaker 2: recursively self-improving. So those are all choices that people are making.
• 16:57
  Speaker 2: So that's the thing we need to focus on from
• 16:59
  Speaker 2: a control standpoint is the people making the choices. The
• 17:02
  Speaker 2: other thing that you make me think of is when
• 17:05
  Speaker 2: it comes to catastrophic risk or extinction risk, one of
• 17:09
  Speaker 2: the big questions is, give me an example of a
• 17:12
  Speaker 2: smarter species, or sorry, a dumber species controlling a smarter
• 17:17
  Speaker 2: species or a dumber thing controlling a smarter thing. And
• 17:21
  Speaker 2: there are all these people in extinction risk and catastrophic
• 17:24
  Speaker 2: risk land who will point to, well, the important thing
• 17:27
  Speaker 2: is to make it feel like it's a mother to
• 17:29
  Speaker 2: humanity so that it has a deep love for people.
• 17:33
  Speaker 2: And One of the questions I didn't ask one of
• 17:36
  Speaker 2: these people is, y'all ever heard of Casey Anthony? And
• 17:40
  Speaker 2: I know that she was found innocent, but we can
• 17:42
  Speaker 2: all have our own personal opinions about how that went down.
• 17:44
  Speaker 2: But to me, it's like, there are hundreds of examples
• 17:47
  Speaker 2: a year of a mother killing her child. And when
• 17:50
  Speaker 2: you are talking about, you know, imbuing the values of
• 17:55
  Speaker 2: a mother to how they take care of their children
• 18:00
  Speaker 2: in a machine that is capable of taking down the
• 18:03
  Speaker 2: entire power grid or making the internet not work. Like
• 18:06
  Speaker 2: these are things where we should not deploy this into
• 18:12
  Speaker 2: systems where there is irreversible harm that can be caused
• 18:15
  Speaker 2: by something because we think that it loves us. That
• 18:17
  Speaker 2: just seems to me like a bad thing and not
• 18:19
  Speaker 2: an intelligent thing to do. And then to the point
• 18:22
  Speaker 2: of lying, one of my favorite things to do, anytime
• 18:24
  Speaker 2: I get any response from a large language model, before
• 18:26
  Speaker 2: I do anything with it, I just open up a
• 18:28
  Speaker 2: new window. I paste the response and I say, I
• 18:30
  Speaker 2: want you to be a rigorous fact checker and red
• 18:33
  Speaker 2: team this response. I don't believe that it's true. And
• 18:36
  Speaker 2: then it'll give you all the new, all of the
• 18:38
  Speaker 2: things that it didn't do because it was trying to
• 18:39
  Speaker 2: please you. It will oftentimes be able to catch that.
• 18:43
  Speaker 2: And just also, you know, using different models to check
• 18:46
  Speaker 2: for different things. Like that's just from a practical standpoint
• 18:48
  Speaker 2: is useful. But yeah, I mean, going back to the
• 18:51
  Speaker 2: idea of the question, is it too late? I don't
• 18:53
  Speaker 2: think so because... I mean, there are so many things
• 18:57
  Speaker 2: happening that by the time that this airs, it's going
• 18:59
  Speaker 2: to be outdated. But just from an intuitive level, the
• 19:02
  Speaker 2: fact that most Americans are highly skeptical of the way
• 19:05
  Speaker 2: that this is being developed and deployed, and they realize
• 19:09
  Speaker 2: the enormous amount of power this conveys to an extremely
• 19:11
  Speaker 2: small group of people, if it continues to scale, that's
• 19:15
  Speaker 2: what brings me hope. And I think that that's where
• 19:18
  Speaker 2: I hope that the people will continue to rein this in.
• 19:22
  Speaker 2: And if it gets to a point where we have
• 19:25
  Speaker 2: things like omnipresent surveillance technology. Like you see how people
• 19:28
  Speaker 2: react into flock. You know, I have, I have very
• 19:32
  Speaker 2: far left liberal friends who want to take an angle
• 19:34
  Speaker 2: grinder to every flock camera they see just as much
• 19:37
  Speaker 2: as my very conservative truck driving brother-in-law. That gives me hope.
• 19:40
  Speaker 2: Like the people, the people get it.
• 19:42
  Speaker 1: I think it is also interesting to me just how
• 19:46
  Speaker 1: acute it has become. because like i've been on dates
• 19:49
  Speaker 1: where i have said that i put something into chad
• 19:51
  Speaker 1: gpt because i where chad gpt and i talk is
• 19:55
  Speaker 1: i find that ai chat bots are very helpful in
• 20:00
  Speaker 1: whittling down that which is infinite And so I talk
• 20:04
  Speaker 1: to it about travel because the world is an infinite thing.
• 20:07
  Speaker 1: And so it's just like, okay, this month, this time,
• 20:10
  Speaker 1: these kinds of things I like, what do we have?
• 20:12
  Speaker 1: Because otherwise I would not be, you know, it'll come
• 20:14
  Speaker 1: up with things I'd never thought of X, Y, and Z.
• 20:17
  Speaker 1: That's where I go. And I find like, especially if
• 20:20
  Speaker 1: you are talking to somebody that's under 40 and you
• 20:23
  Speaker 1: mentioned using some AI, there's a recoil and resistance that
• 20:28
  Speaker 1: comes from people. Now you also have the kids that
• 20:29
  Speaker 1: are using it to cheat on every exam and every
• 20:31
  Speaker 1: paper that they've ever been asked to write. Um, but
• 20:35
  Speaker 1: I think, and I thought that this was an interesting
• 20:37
  Speaker 1: part of the doc also was kind of an acknowledgement
• 20:39
  Speaker 1: that there's a bit of a disconnect between how the
• 20:41
  Speaker 1: average person will use this technology and the idea, for example,
• 20:46
  Speaker 1: if you can't afford to build a staff for a
• 20:49
  Speaker 1: entrepreneurial idea you have. That the ability like there's a
• 20:52
  Speaker 1: bit of democratization that comes from being able to use
• 20:54
  Speaker 1: the technology to help you. But there's a much bigger, larger,
• 20:59
  Speaker 1: more terrifying discussion that is higher. And that's when we
• 21:02
  Speaker 1: start talking about surveillance. That's when we start talking about
• 21:04
  Speaker 1: defense and all of those things. And it feels kind
• 21:06
  Speaker 1: of like we're being sold the little thing that can
• 21:09
  Speaker 1: make your life a bit easier. But behind the curtain,
• 21:13
  Speaker 1: it's something else that's going on.
• 21:15
  Speaker 2: It's that's very true. And it's to me, somebody pointed
• 21:17
  Speaker 2: out the other day to me that right now we're
• 21:19
  Speaker 2: in a world where using a large language model is
• 21:22
  Speaker 2: the same as like online dating in the early, like
• 21:24
  Speaker 2: the mid two thousands where everybody's doing it, but nobody
• 21:26
  Speaker 2: wants to talk about it. Like, you don't want to say,
• 21:28
  Speaker 2: you don't want to say we met on Tinder. You
• 21:30
  Speaker 2: want to say we met at work or whatever. Uh,
• 21:33
  Speaker 2: and I think that to you, to your other point,
• 21:36
  Speaker 2: it is a, there's a soylent green is people aspect
• 21:39
  Speaker 2: to a lot of this, to your, to your point
• 21:41
  Speaker 2: where The idea of having something that makes it easier
• 21:45
  Speaker 2: for me to make a flyer for my kid's birthday party,
• 21:47
  Speaker 2: financing what is also being sold to the federal government
• 21:51
  Speaker 2: to create a Panopticon, that really hasn't happened in the
• 21:55
  Speaker 2: history of technology. Usually, if I'm giving money to Amazon,
• 21:59
  Speaker 2: I know what Jeff Bezos is trying to do with
• 22:01
  Speaker 2: that money. And that's because it's a general purpose technology.
• 22:04
  Speaker 2: So we kind of We have this notion of artificial
• 22:07
  Speaker 2: general intelligence, which is the sort of a digital god
• 22:09
  Speaker 2: that a lot of the frontier labs are trying to build,
• 22:12
  Speaker 2: which is a machine that is designed to replace the
• 22:16
  Speaker 2: totality of economically valuable intellectual labor that a human can do.
• 22:21
  Speaker 2: That's step one. And then step two is doing the
• 22:23
  Speaker 2: same thing, but replacing all physical labor in the form
• 22:25
  Speaker 2: of robots. So that's the stated goal of these companies.
• 22:28
  Speaker 2: That's where people are getting wise to it because they've
• 22:31
  Speaker 2: been told for the last six, seven, eight years, oh,
• 22:33
  Speaker 2: we're coming to replace you. And now, if you'll notice,
• 22:36
  Speaker 2: the song has changed a little bit. Now they're not
• 22:38
  Speaker 2: talking about replacement. Now they're talking about, well, it's just
• 22:39
  Speaker 2: going to make you a little bit better. It's going
• 22:41
  Speaker 2: to make you more efficient. And the challenge about navigating
• 22:46
  Speaker 2: these areas, it's very difficult for a human being to
• 22:50
  Speaker 2: understand that two things can be true at the same
• 22:51
  Speaker 2: time that sound as though they're antithetical. So it can
• 22:55
  Speaker 2: be true that they are scaling to a point where
• 22:56
  Speaker 2: it could replace people at the same time where, in
• 23:00
  Speaker 2: really specific cases, it still sucks. Uh, and the other
• 23:04
  Speaker 2: part of it that gets really challenging is most people
• 23:06
  Speaker 2: who I know who think that large language models don't
• 23:09
  Speaker 2: do a good job are not using the paid tier,
• 23:11
  Speaker 2: which means they're not using the state of the art models. And,
• 23:14
  Speaker 2: and the difference there is enormous, which then gets us
• 23:17
  Speaker 2: into the inequality piece of it, which is if you
• 23:19
  Speaker 2: have more money, you're going to have access to better
• 23:21
  Speaker 2: models and look what that does on a business level,
• 23:24
  Speaker 2: on an education level, on a, on a kid level.
• 23:26
  Speaker 2: And then you get into the whole notion of like
• 23:28
  Speaker 2: how it's impacting education. And to your point at that
• 23:30
  Speaker 2: Ivy league school. where the professor pulled a switch on
• 23:33
  Speaker 2: the last minute and he went from an online exam
• 23:35
  Speaker 2: to an in-person exam and the scores, I believe, dropped
• 23:37
  Speaker 2: 45 or 50%, we're talking about a sort of a
• 23:41
  Speaker 2: dumbification of the next generation of people where they're not
• 23:45
  Speaker 2: being taught to think. And this is all the fallout
• 23:47
  Speaker 2: when you, I used to teach at schools in New
• 23:50
  Speaker 2: Haven and it was the birth of the teach for
• 23:52
  Speaker 2: the test.
• 23:55
  Speaker 1: Era.
• 23:56
  Speaker 2: Now you're seeing a world where you're combining teaching for
• 23:58
  Speaker 2: the test with a software that will do all of
• 24:01
  Speaker 2: that work for you, and you end up with a
• 24:03
  Speaker 2: generation of people who don't know how to think. And
• 24:07
  Speaker 2: to me, those societal harms, and this, again, doesn't get
• 24:11
  Speaker 2: into the environmental racism based around where the data center's
• 24:13
  Speaker 2: being built, the environmental impact of those things, the financial
• 24:16
  Speaker 2: impact of utility rates, et cetera, but just on a level,
• 24:20
  Speaker 2: you're exactly right that we've never been in a place
• 24:23
  Speaker 2: before where a company that's making what is a subscription
• 24:28
  Speaker 2: product I mean, it'd be like if Spotify were building
• 24:31
  Speaker 2: a nuclear bomb. You know what I mean? Right. I
• 24:34
  Speaker 2: just want some music. I just want to listen. I
• 24:38
  Speaker 2: think I want to do a good thing. I want
• 24:39
  Speaker 2: my artists to get one one thousandth of a penny
• 24:41
  Speaker 2: every time I listen to their song. But then behind
• 24:44
  Speaker 2: the scenes, they're helping, you know, the federal government become
• 24:47
  Speaker 2: Big Brother. We haven't seen that before.
• 24:50
  Speaker 1: I think it was my man Yannis Varoufakis who had
• 24:52
  Speaker 1: made the point that he thought the saving grace on
• 24:54
  Speaker 1: this AI thing would be This system only works if you,
• 24:58
  Speaker 1: this economic system only works if you have sellers and
• 25:01
  Speaker 1: buyers and people can't buy if they don't have money
• 25:04
  Speaker 1: and you get money from going to jobs and that
• 25:07
  Speaker 1: at some point that would be the thing that would
• 25:09
  Speaker 1: slow people down. But no, that's not doing that. Like
• 25:12
  Speaker 1: that is, I think part of what's interesting about this
• 25:14
  Speaker 1: is not just about AI. It's about the larger tech ethos, right?
• 25:17
  Speaker 1: Like just break shit and see what happens. However they
• 25:20
  Speaker 1: put that, right? Move fast and break shit. Break stuff,
• 25:22
  Speaker 1: I think is what they call it. But this is
• 25:23
  Speaker 1: the most interesting example of it, because it's one thing
• 25:26
  Speaker 1: when you say that about just like trying to come
• 25:28
  Speaker 1: up with some cool technological doodads. But now we're talking
• 25:31
  Speaker 1: about something that has grown faster than I think most
• 25:34
  Speaker 1: people expected that it would be capable of growing. And
• 25:38
  Speaker 1: it's created a phenomenon. And I think something that the
• 25:41
  Speaker 1: film also talks about where now everybody feels like they
• 25:43
  Speaker 1: have to use it.
• 25:44
  Speaker 2: Correct.
• 25:45
  Speaker 1: But nobody asked for this. But it does. It's like, hey, man,
• 25:48
  Speaker 1: you're going to be behind.
• 25:50
  Speaker 2: Exactly. Well, and that's the thing, the idea of the
• 25:55
  Speaker 2: degree of FOMO that we feel from not being on
• 25:57
  Speaker 2: social media is social. But to combine those sort of
• 26:01
  Speaker 2: neural pathways with the idea of you are going to
• 26:04
  Speaker 2: not be as productive as the next person, it is
• 26:07
  Speaker 2: totally understandable to feel that way and to believe that
• 26:10
  Speaker 2: you must actually use this stuff. And in a lot
• 26:13
  Speaker 2: of cases, people are being forced to, regardless of what
• 26:16
  Speaker 2: the backlash is going to be. The number of people
• 26:17
  Speaker 2: that I know who have, you work in social media,
• 26:21
  Speaker 2: et cetera, and they try to explain to their bosses,
• 26:23
  Speaker 2: you don't get it. If I put an AI-generated thumbnail
• 26:26
  Speaker 2: on this video, people are going to comment bomb this
• 26:29
  Speaker 2: into oblivion, and you're going to lose credibility as a brand.
• 26:32
  Speaker 2: That happens constantly, and then they have to show their bosses, look,
• 26:37
  Speaker 2: you need to go back to human employment. That, to me,
• 26:41
  Speaker 2: it's a really, there's a lot of conflict there where
• 26:43
  Speaker 2: people want to not miss out on the increase in
• 26:46
  Speaker 2: productivity that they believe that they can get from the systems,
• 26:49
  Speaker 2: while at the same time, being able to account for
• 26:53
  Speaker 2: the social pressure for not using it. So it's a
• 26:56
  Speaker 2: very strange soup to be in.
• 26:57
  Speaker 1: All right, now coming up next, I'm gonna talk to
• 27:00
  Speaker 1: Ted about the experience of talking to some of these founders,
• 27:03
  Speaker 1: which the documentary has, including Sam Altman, and more on
• 27:07
  Speaker 1: what is going on with AI. The Right Time is
• 27:11
  Speaker 1: brought to you by FanDuel. Baseball fans, think you know
• 27:15
  Speaker 1: who's hitting a homer today? Well, FanDuel is giving you
• 27:17
  Speaker 1: a free chance to call your shot with daily dingers.
• 27:21
  Speaker 1: Joining the action is easy. Just make your free pick
• 27:24
  Speaker 1: on who's hitting a home run in their game, and
• 27:26
  Speaker 1: if you're right, you'll get a profit boost to use
• 27:29
  Speaker 1: on your next bet for a chance at extra winnings.
• 27:32
  Speaker 1: You've got nothing to lose and a lot to gain.
• 27:35
  Speaker 1: And don't forget, you can make a free pick every
• 27:37
  Speaker 1: day during the MLB season. It's just another reason why
• 27:41
  Speaker 1: FanDuel is the home for home runs. So are you
• 27:45
  Speaker 1: ready to play your game today? Join all the action
• 27:48
  Speaker 1: at FanDuel.com slash Bomani. That's FanDuel.com slash Bomani to play
• 27:54
  Speaker 1: daily dingers and make your free pick on who's hitting
• 27:57
  Speaker 1: a homer. Play ball and swing for the fences on FanDuel,
• 28:01
  Speaker 1: an official partner of MLB. 21 plus states for Kansas
• 28:05
  Speaker 1: in affiliation with Kansas Star Casino. 18 plus, D.C., Kentucky, Wyoming.
• 28:11
  Speaker 1: Gambling problem? Call 1-800-GAMBLER. 1-800-MY-RESET.
• 28:13
  Speaker 2: Call 1-888-789-7777.
• 28:14
  Speaker 1: Visit ccpg.org slash chat, Connecticut. Visit mdgamblinghelp.org, Maryland. Visit gamblinghelplinema.org.
• 28:18
  Speaker 2: Call 800-327-5050.
• 28:37
  Speaker 1: Massachusetts. Call 1-877-8-HOPE-NY. Text HOPE-NY.
• 28:44
  Speaker 2: New York.
• 28:46
  Speaker 1: Call 1-877-770-7867. Louisiana. Look, if you're serious about performance, you
• 28:55
  Speaker 1: want products you can trust. GNC was built for athletes
• 28:59
  Speaker 1: and has been trusted by athletes for 91 years. They
• 29:02
  Speaker 1: review every product against a restricted list of more than
• 29:06
  Speaker 1: 150 no-go ingredients, and their experts test products for quality, taste,
• 29:11
  Speaker 1: and performance. They also offer products that go through strict
• 29:14
  Speaker 1: third-party banned substance verification, which matters if you're competing at
• 29:19
  Speaker 1: any level. Whether you're looking for energy, recovery, better sleep,
• 29:23
  Speaker 1: or just trying to be ready on game day, GNC
• 29:26
  Speaker 1: has something for every athlete, every fan, and every goal.
• 29:31
  Speaker 1: With more than 2,000 locations and 24-7 shopping at GNC.com,
• 29:36
  Speaker 1: it's easy to find what works for you. Discover why
• 29:39
  Speaker 1: millions trust GNC. Shop athlete-tested, science-backed picks today at GNC.com.
• 29:48
  Speaker 1: Let's face it, nothing feels better than finding grooming products
• 29:51
  Speaker 1: that are actually for you, and that's where Baskin-Lather can help.
• 29:55
  Speaker 1: Baskin-Lather is a Black-owned and family-operated company that researched and
• 29:59
  Speaker 1: formulated their own scalp and hair oil to stimulate growth
• 30:02
  Speaker 1: and retain length. I recommend Beard Growth Magic for fuller,
• 30:06
  Speaker 1: thicker facial hair. A blend of vital nutrients to help
• 30:10
  Speaker 1: nourish and increase blood circulation to hair follicles that promote
• 30:13
  Speaker 1: softening of beard hair and dandruff reduction while working for
• 30:17
  Speaker 1: all hair types. Also, make sure you try Men's Healthy
• 30:20
  Speaker 1: Hair and Beard Starter Kit. an all-in-one set that helps
• 30:23
  Speaker 1: you maintain a healthy, stylish, well-groomed beard. Includes shampoo, conditioner, balm, oil,
• 30:30
  Speaker 1: and a stimulator formula. Each product is crafted with high-quality
• 30:34
  Speaker 1: ingredients for the care you deserve. Explore viral bestsellers and
• 30:38
  Speaker 1: products of healthier hair of all types from Baskin-Lather. Go
• 30:43
  Speaker 1: to BaskinLatherCO.com and use code TRT for 20% off. That's 20%
• 30:49
  Speaker 1: off at BaskinLatherCO.com, code TRT. All right, we are back
• 30:56
  Speaker 1: with my Ted Tremper, producer on the AI Doc, how
• 30:59
  Speaker 1: I became an apocaloptimist. I recommend that you check this out.
• 31:04
  Speaker 1: I was actually surprised that you got, what, four CEOs.
• 31:09
  Speaker 2: We got three out of five. So we got Sam Altman,
• 31:11
  Speaker 2: Dario Amadei, and Demis Hassabis. Yeah.
• 31:13
  Speaker 1: Right. And so the five companies are OpenAI, Anthropic, DeepMind, Google,
• 31:19
  Speaker 1: and is Microsoft the fifth?
• 31:20
  Speaker 2: It's Google DeepMind. So it's OpenAI, Anthropic, Google DeepMind, and
• 31:24
  Speaker 2: then XAI and MetaAI.
• 31:27
  Speaker 1: Yes, yes. So Zuckerberg and Elon, huh?
• 31:30
  Speaker 2: Zuckerberg turned us down multiple times. I developed a contact that...
• 31:34
  Speaker 2: So in the documentary, it points out that Elon Musk
• 31:37
  Speaker 2: agreed to be in the documentary, but then got busy.
• 31:39
  Speaker 2: And that was essentially the experience that I developed a
• 31:42
  Speaker 2: contact in his inner circle. After we got Sam Altman,
• 31:45
  Speaker 2: Dari Amadei, and Demis Hassabis, this person channeled to him hey,
• 31:49
  Speaker 2: if you want to do this, this is who we have.
• 31:50
  Speaker 2: And he said, yes, I should definitely do that. Have
• 31:52
  Speaker 2: that person text me. So I texted Elon three times
• 31:54
  Speaker 2: and he didn't get back to me.
• 31:55
  Speaker 1: So you have Elon Musk's phone number, which is an
• 31:56
  Speaker 1: amazing thing to be able to say.
• 32:00
  Speaker 2: Well, it went... I can't get into too much of
• 32:03
  Speaker 2: what the timing was for these things, but this was
• 32:05
  Speaker 2: at the height of Doge. And unfortunately, the third text
• 32:09
  Speaker 2: I sent went through his green and I thought I
• 32:11
  Speaker 2: blocked the number and I checked in with my contact
• 32:13
  Speaker 2: and they said, oh, well, he had to change it
• 32:15
  Speaker 2: because of the secret baby thing. And that was when the...
• 32:18
  Speaker 2: I said, what? And then I checked the news. And
• 32:21
  Speaker 2: that was when the woman, I think it worked for
• 32:24
  Speaker 2: the communications department of the White House or Fox News,
• 32:27
  Speaker 2: had come forward and said, oh, yeah, I had a
• 32:29
  Speaker 2: lens baby, et cetera, et cetera. And he had to
• 32:31
  Speaker 2: change his number up. So I don't have it anymore.
• 32:34
  Speaker 1: Wow, Elon Musk is just like everybody else trying to
• 32:37
  Speaker 1: shake off a baby mama that he wasn't trying to have.
• 32:40
  Speaker 2: Allegedly, allegedly.
• 32:41
  Speaker 1: Hey, man, Billie Jean is not my lover, said Elon Musk.
• 32:46
  Speaker 1: But what did you... get from talking to these people
• 32:51
  Speaker 1: now obviously they're not all the same person but i
• 32:53
  Speaker 1: imagine there had to be some common threads that ran
• 32:55
  Speaker 1: between the people that are running these.
• 32:57
  Speaker 2: Companies yes yeah i think that it's it's really interesting
• 33:00
  Speaker 2: so so the process of getting them uh that they
• 33:04
  Speaker 2: were all my white whales as it were and i
• 33:07
  Speaker 2: i believed that we could use the fact that they
• 33:09
  Speaker 2: are in this race they are in a race to
• 33:11
  Speaker 2: deploy this to develop and deploy this technology And they
• 33:14
  Speaker 2: believe that if they win it, they will essentially control
• 33:17
  Speaker 2: the most powerful tool that humanity has ever built in
• 33:19
  Speaker 2: the form of artificial general intelligence. So then you have
• 33:21
  Speaker 2: to think, okay, well, if those are their incentives, how
• 33:24
  Speaker 2: can I create a circumstance in which they'd want to
• 33:26
  Speaker 2: participate in a documentary like this? Because there is no
• 33:28
  Speaker 2: real incentive for them to expose themselves to really difficult questions,
• 33:32
  Speaker 2: et cetera. So my belief was if I got one,
• 33:35
  Speaker 2: I would probably get more than one and I could
• 33:37
  Speaker 2: maybe get all of them. So what that meant was
• 33:39
  Speaker 2: for those two and a half years, I read and
• 33:42
  Speaker 2: listened to every single thing that any of them had
• 33:44
  Speaker 2: ever written. And I wrote down the names of every
• 33:46
  Speaker 2: person they ever name checked, good or bad. And I
• 33:49
  Speaker 2: reached out to the people that could get me to
• 33:50
  Speaker 2: those people. And those were a lot of the people
• 33:52
  Speaker 2: that we interviewed for the film. So once we had
• 33:55
  Speaker 2: amassed between 40 to 60 of those names, when we
• 33:59
  Speaker 2: actually made the ask, that that list would have at
• 34:03
  Speaker 2: least 20 people that they really love and respect and
• 34:06
  Speaker 2: a couple of people that they were very afraid of
• 34:08
  Speaker 2: what they would say. So that way they would want
• 34:10
  Speaker 2: to go on record to be able to.
• 34:12
  Speaker 1: Well done.
• 34:13
  Speaker 2: Yeah. So, you know, it's, it's a, it's a long,
• 34:16
  Speaker 2: it was a long, long, long con, long play. And
• 34:20
  Speaker 2: to that effect, like one of the things I want
• 34:23
  Speaker 2: to give them credit for, which is very different is from,
• 34:25
  Speaker 2: from previous CEOs is, you know, General Motors to my
• 34:30
  Speaker 2: knowledge knew about the harms of leaded gasoline, you know,
• 34:34
  Speaker 2: as soon as the thirties or forties, you know, The
• 34:38
  Speaker 2: oil industry knew about climate change 40 years ago. They
• 34:41
  Speaker 2: kept a lid on it. These CEOs have been very
• 34:44
  Speaker 2: proactive in saying that they believe that this technology could
• 34:48
  Speaker 2: result in the extinction of the human race. I'm sorry,
• 34:52
  Speaker 2: I made you laugh. No, I know, but I mean,
• 34:53
  Speaker 2: it is what it is. They've been saying it for,
• 34:55
  Speaker 2: I mean, Sam Altman's blog that he's been writing since
• 34:58
  Speaker 2: 2016 or 15, I think, is still online. You can
• 35:00
  Speaker 2: read every word that he's ever said. And then it
• 35:03
  Speaker 2: becomes a question about whether or not they're using that
• 35:05
  Speaker 2: to hype the power and the capabilities of these products.
• 35:07
  Speaker 2: And depending on who you are, like depending on who
• 35:11
  Speaker 2: is writing that, I believe that each of them believe
• 35:13
  Speaker 2: different things about that. And I think people need to
• 35:16
  Speaker 2: make up their minds about those people. I've not actually
• 35:19
  Speaker 2: read the Ronan Farrow piece about Sam Altman, but there's
• 35:22
  Speaker 2: one from, I think, 2019 when he took over Y
• 35:26
  Speaker 2: Combinator that was written in The New Yorker. And Paul Graham,
• 35:30
  Speaker 2: who was Sam Altman's mentor, said of him, and he
• 35:35
  Speaker 2: meant this as a compliment, I believe the quote is,
• 35:38
  Speaker 2: if you dropped Sam on an island of cannibals and
• 35:40
  Speaker 2: came back in two years, he would be the president
• 35:42
  Speaker 2: of Cannibal Island. And I think that if you know
• 35:46
  Speaker 2: much about the way that he was fired and then
• 35:48
  Speaker 2: brought back to OpenAI, Sam seems to be the kind
• 35:52
  Speaker 2: of person who is really interested in telling you exactly
• 35:54
  Speaker 2: what he believes that you want to hear and using
• 35:57
  Speaker 2: that as a way to maintain power over whatever fishbowl
• 36:01
  Speaker 2: he's in. And I think that my personal opinion is
• 36:05
  Speaker 2: that I think that Dario and Demis are a little
• 36:07
  Speaker 2: bit different than that. I think Elon has different incentives.
• 36:10
  Speaker 2: I think that Mark Zuckerberg has different incentives. But regardless,
• 36:13
  Speaker 2: the most important thing to take away is if we
• 36:17
  Speaker 2: are trying to live in a world where the future
• 36:21
  Speaker 2: of humanity depends on whether or not one person is
• 36:24
  Speaker 2: a good guy, we're not ready. We have not built
• 36:28
  Speaker 2: the systems that can coexist with that world. But unfortunately,
• 36:34
  Speaker 2: you know, The rise of authoritarianism makes perfect sense given
• 36:38
  Speaker 2: that you have people in our society and parties in
• 36:40
  Speaker 2: our society that are trying to prove that the systems
• 36:42
  Speaker 2: that exist don't work and need to be dismantled. I
• 36:45
  Speaker 2: disagree with that. I think they need to be upgraded.
• 36:47
  Speaker 2: And I think that people need to engage in good
• 36:48
  Speaker 2: faith in upgrading them so that we don't evolve to
• 36:50
  Speaker 2: a place where we have one dude who we really
• 36:54
  Speaker 2: hope is a nice fella in control of that kind
• 36:57
  Speaker 2: of power and that kind of technology.
• 36:59
  Speaker 1: Well, like the good guy dependence thing becomes interesting when
• 37:01
  Speaker 1: talking about Altman and Musk. And for people who don't know,
• 37:03
  Speaker 1: Altman and Musk were working together and OpenAI and the
• 37:06
  Speaker 1: whole claim was we're going to do this. Basically, they
• 37:09
  Speaker 1: were going to be the morally upstanding people, right? Like
• 37:11
  Speaker 1: they were going to be the ones you could trust.
• 37:13
  Speaker 1: And then they broke up. Basically, I believe each claiming
• 37:17
  Speaker 1: that the other one really wasn't that invested in the
• 37:20
  Speaker 1: first place in doing the right thing. And now they're
• 37:23
  Speaker 1: in the race. And I mean, Elon Musk is the
• 37:25
  Speaker 1: guy to do the right thing is somewhat of a
• 37:27
  Speaker 1: laughable notion, but it wasn't as laughable in 2015. No.
• 37:31
  Speaker 2: And that's the thing that's really important. And actually, I'm
• 37:34
  Speaker 2: so glad you bring that up because we do need
• 37:35
  Speaker 2: to remember that there was a point where Elon Musk
• 37:38
  Speaker 2: was regarded to be the Thomas Edison of our time,
• 37:40
  Speaker 2: that he was going to lead us to Mars. He
• 37:41
  Speaker 2: was going to Tesla was going to replace every gas car.
• 37:46
  Speaker 2: We were going to fix global warming, et cetera. But
• 37:48
  Speaker 2: then you look at the ways that that and mirroring
• 37:51
  Speaker 2: it against social media, social media was, you know, Arab Spring.
• 37:54
  Speaker 2: We were going to be there was going to be
• 37:55
  Speaker 2: a flourishing of democracy. It was going to be this
• 37:58
  Speaker 2: tool that led to bringing out all of our best natures.
• 38:01
  Speaker 2: And look what happened. The incentives that are based around
• 38:04
  Speaker 2: maximizing for attention, maximizing for outrage point toward the society
• 38:08
  Speaker 2: that we got. And you need to look and have
• 38:11
  Speaker 2: the same understanding of what the tech industry is trying
• 38:13
  Speaker 2: to do with the proliferation of A.I. if the goal
• 38:16
  Speaker 2: is to be able to replace human beings so that
• 38:18
  Speaker 2: you can control the totality of the economy. Like when
• 38:21
  Speaker 2: Elon Musk says that AI is going to be a
• 38:23
  Speaker 2: multi-trillion dollar market, it's because he believes that's where we're headed.
• 38:27
  Speaker 2: And we are really getting to a point where if
• 38:31
  Speaker 2: the capabilities scale and they're able to start paying back
• 38:35
  Speaker 2: investors at the rate that they need to, they might
• 38:38
  Speaker 2: be right because it may become too tantalizing a proposition
• 38:41
  Speaker 2: for business to continue to exist without it. And at
• 38:44
  Speaker 2: that point, if we have eroded the social support structure
• 38:49
  Speaker 2: that the very middling ones that exist in our society already,
• 38:53
  Speaker 2: we're in a world where there's no transition plan between
• 38:56
  Speaker 2: double digit, you know, 40% unemployment. And if you, if
• 39:00
  Speaker 2: you've done your history lessons, you realize that Hitler rose
• 39:02
  Speaker 2: to power in Germany because of three consecutive years of 20% unemployment.
• 39:06
  Speaker 1: Yeah.
• 39:07
  Speaker 2: That's not good. And so I think that that's, yeah.
• 39:11
  Speaker 2: I mean, to your point, boy, I don't know if
• 39:14
  Speaker 2: Elon was different then. I don't know whether or not
• 39:17
  Speaker 2: the ketamine and spending too much time on Twitter rotted
• 39:20
  Speaker 2: his brain out. But like I said, either way, even
• 39:24
  Speaker 2: if it was Elon 20 you know 2008 uh i
• 39:29
  Speaker 2: wouldn't trust him with the technology because he's a human
• 39:30
  Speaker 2: being and that's that's not that's not the kind of
• 39:33
  Speaker 2: world we need to live in.
• 39:34
  Speaker 1: And as we said before and i think it bears
• 39:36
  Speaker 1: repeating these apps are designed to be customer friendly. Like,
• 39:40
  Speaker 1: I don't think we've really got into that, but that's
• 39:41
  Speaker 1: an important thing to note about the way this stuff
• 39:43
  Speaker 1: is used is that it is designed to give you
• 39:47
  Speaker 1: answers to make you use it more because it's as
• 39:50
  Speaker 1: a product, its goal is to make you use it.
• 39:53
  Speaker 1: They may have these other intentions about what it could
• 39:55
  Speaker 1: do like for the world, but the bottom line is
• 39:57
  Speaker 1: in order for it to catch on, They're going to
• 39:59
  Speaker 1: do the things in there that make you use it.
• 40:02
  Speaker 1: That's why I talk about when it lied to me
• 40:03
  Speaker 1: that time because it didn't want to tell me it
• 40:05
  Speaker 1: couldn't do something. It's just like people in your regular life.
• 40:09
  Speaker 2: Well, and down to a design level, there's a great
• 40:11
  Speaker 2: professor out of Oxford whose name I forget is one
• 40:14
  Speaker 2: of the lines that people use. Oh, it's just a tool.
• 40:16
  Speaker 2: It's just a tool. You know, a knife can be
• 40:18
  Speaker 2: used to stab somebody or it can be used to
• 40:20
  Speaker 2: spread butter. And this professor brought up, you need to
• 40:24
  Speaker 2: think about the intention of how you're designing it because
• 40:27
  Speaker 2: believe it or not, a gun, can also be used
• 40:30
  Speaker 2: to spread butter if you want it to, but it's
• 40:32
  Speaker 2: designed to kill people. And when you talk about sycophancy,
• 40:35
  Speaker 2: which is the notion of AI lying to you, like
• 40:38
  Speaker 2: the reason why when you put a query into ChatGPT,
• 40:40
  Speaker 2: the little three dots appears on the left side of
• 40:42
  Speaker 2: the screen, like it's sending you a text, that's for
• 40:45
  Speaker 2: you to believe that it's a person behind there. There's
• 40:48
  Speaker 2: a reason why Anthropic's model has a name and that
• 40:51
  Speaker 2: name is Claude. Like all of these things are intentions
• 40:54
  Speaker 2: of design. And it's interesting, there's a very... There's a
• 40:57
  Speaker 2: country called China that made it illegal to anthropomorphize models
• 41:01
  Speaker 2: because they understand that they don't want their populace, they
• 41:04
  Speaker 2: don't want to cede power and control to this ostensibly
• 41:09
  Speaker 2: omnipotent thing because then you start getting in these psychotic
• 41:12
  Speaker 2: states where you are believing that it shouldn't be making
• 41:16
  Speaker 2: every decision for you, et cetera. And children obviously are
• 41:19
  Speaker 2: the most vulnerable to those kinds of impressions where you
• 41:22
  Speaker 2: just want to give away all of your decisions to that.
• 41:24
  Speaker 2: And I'm certainly guilty of it based off of travel.
• 41:27
  Speaker 2: algorithm has been doing this to people for, I don't
• 41:29
  Speaker 2: know why I do pottery, Bo. I don't know why,
• 41:33
  Speaker 2: but I love pottery. And I don't know whether or
• 41:35
  Speaker 2: not the algorithm fed that to me because it knew
• 41:37
  Speaker 2: that it, that I would like it or whether or
• 41:39
  Speaker 2: not I just looked at one Reddit post for too long.
• 41:42
  Speaker 2: And now I've spent a thousand dollars on pottery lessons
• 41:44
  Speaker 2: this year, dude. This is such a tripper thing. I
• 41:48
  Speaker 2: think I like it. I think I like it. But
• 41:51
  Speaker 2: you know, it's, it's this notion, the center of humane
• 41:53
  Speaker 2: technology brings up this notion that every human being wants
• 41:56
  Speaker 2: to believe we're living above the algorithm we want to
• 41:59
  Speaker 2: believe that we are above the control mechanisms that uh
• 42:04
  Speaker 2: you know all of the different ways that algorithms are
• 42:06
  Speaker 2: designed to monopolize our thought and our attention we all
• 42:09
  Speaker 2: want to believe we're above that because we're human beings
• 42:12
  Speaker 2: but guess what study after study just shows we're very
• 42:15
  Speaker 2: impressionable and it's not our fault but if we account
• 42:18
  Speaker 2: for it we can put different guardrails in place so
• 42:21
  Speaker 2: that we don't fall prey to uh AI sycophancy, AI psychosis,
• 42:26
  Speaker 2: et cetera, et cetera. We just have to know it.
• 42:28
  Speaker 1: Now, you mentioned China, and I think that's an important
• 42:30
  Speaker 1: part of this and for the tail end of what
• 42:33
  Speaker 1: we're doing here, because this is a geopolitical competition superpower
• 42:39
  Speaker 1: sort of issue. So one thing that is interesting, and
• 42:41
  Speaker 1: I think it was Giannis who wrote a book that
• 42:42
  Speaker 1: mentioned this, is that economically speaking, like Europe is in
• 42:45
  Speaker 1: a very precarious position because there is no European that
• 42:49
  Speaker 1: owns anything like Facebook, like Twitter, these sites that just
• 42:54
  Speaker 1: dominate Amazon, right? They get all your information and have
• 42:57
  Speaker 1: all this stuff, right? Like those guys that are turning
• 42:59
  Speaker 1: into trillionaires, none of them are from Europe because Europe
• 43:02
  Speaker 1: does not, nobody in Europe possesses that level of infrastructure.
• 43:06
  Speaker 1: China is in a different situation, right? They've got some
• 43:09
  Speaker 1: of those things and they got that TikTok thing, which
• 43:11
  Speaker 1: becomes its own discussion, the power that that algorithm has
• 43:14
  Speaker 1: and everything else. But the thing that's important about this
• 43:19
  Speaker 1: is there's a space race involved here. And the Chinese
• 43:24
  Speaker 1: are trying to figure out how they're going to do it.
• 43:26
  Speaker 1: And a lot of it has to do with microchips
• 43:28
  Speaker 1: and who produces them and who can then get possession
• 43:30
  Speaker 1: of them and everything else. But in the end, this
• 43:33
  Speaker 1: is not just a race about economics and like the
• 43:37
  Speaker 1: stock market and who can get the riches. it's kind
• 43:39
  Speaker 1: of sort of the way that the way the countries
• 43:42
  Speaker 1: view this at a national level is that this is
• 43:44
  Speaker 1: about the fate of the world. So how much like
• 43:46
  Speaker 1: the government will only do with so much curbing of
• 43:49
  Speaker 1: this industry, because if you curb it here and the
• 43:51
  Speaker 1: Chinese run away, then what do we have?
• 43:55
  Speaker 2: And that's, you know, the three main, the three main
• 43:57
  Speaker 2: things that allow for capabilities to scale our power data
• 44:01
  Speaker 2: and compute, which is the chips. And when you look
• 44:04
  Speaker 2: at the infrastructure that China has built over the last
• 44:06
  Speaker 2: 20 years, uh, Most people don't realize that they are
• 44:09
  Speaker 2: not just the world's leader in renewable energy. They are
• 44:12
  Speaker 2: like many, many, many, many times second place. And they've
• 44:15
  Speaker 2: been doing this for decades at this point. Then you
• 44:18
  Speaker 2: think about data. When you live in a society where
• 44:21
  Speaker 2: people don't have any rights at all for their medical records,
• 44:24
  Speaker 2: et cetera, you can just feed all that data in there,
• 44:26
  Speaker 2: not to mention all the stuff that they're able to steal.
• 44:29
  Speaker 2: And then you talk about compute. We have the CHIPS Act,
• 44:32
  Speaker 2: which was trying to curtail the amount of the state-of-the-art
• 44:36
  Speaker 2: chips that were going to China. That seems, this all
• 44:39
  Speaker 2: changes such and such. Even if they just needed to
• 44:41
  Speaker 2: buy them in the black market, which they were doing anyway,
• 44:43
  Speaker 2: they could still maintain access to those. But the thing
• 44:46
  Speaker 2: that looks like it's happening now with the birth of these, like,
• 44:50
  Speaker 2: you know, we had another deep seek moment last week
• 44:53
  Speaker 2: or two weeks ago with Kimi K3 and the new
• 44:55
  Speaker 2: deep seek model where their capabilities are outpacing.
• 44:58
  Speaker 1: And for people to know, deep seek is the Chinese
• 45:01
  Speaker 1: version of these five companies that we're talking about. And
• 45:04
  Speaker 1: they shook the world up. when they were able to
• 45:06
  Speaker 1: come up with one of these models, they do it
• 45:07
  Speaker 1: a lot cheaper, which may be very fearful for these
• 45:10
  Speaker 1: five companies. Yes.
• 45:12
  Speaker 2: Oh my God.
• 45:13
  Speaker 1: The Chinese bootleg version is here.
• 45:16
  Speaker 2: And the joke, the funny thing about that is obviously
• 45:18
  Speaker 2: like the American frontier models were built off of stolen
• 45:23
  Speaker 2: and unlicensed data that the amount, like really when you
• 45:27
  Speaker 2: think about the amount of theft and folding in of
• 45:30
  Speaker 2: unlicensed data into these models, you can't really conceive of it.
• 45:33
  Speaker 2: It's every email you've ever written. It's every, every Reddit post,
• 45:36
  Speaker 2: it's every, your children's vacation, like photos of them on vacation,
• 45:39
  Speaker 2: it's everything, goes into these models. And I'm being general,
• 45:42
  Speaker 2: it's not literally everything, but it is a staggering amount
• 45:44
  Speaker 2: of data. The thing that always makes me laugh a
• 45:49
  Speaker 2: little bit is the Chinese models invented this process of
• 45:52
  Speaker 2: distillation where they essentially query the American models enough times
• 45:56
  Speaker 2: they can essentially pull the brain out of it. And
• 45:58
  Speaker 2: so for a few million dollars as opposed to a
• 46:00
  Speaker 2: few billion dollars, you can essentially pull out all of
• 46:03
  Speaker 2: the different ways it works and build this Chinese model.
• 46:05
  Speaker 2: And then the frontier labs say, well, hey, you stole
• 46:08
  Speaker 2: all of our data. And I'm over here sitting like, dude,
• 46:11
  Speaker 2: you invented the game. Don't hate this new player for
• 46:16
  Speaker 2: playing the game that you set up. And the thing
• 46:19
  Speaker 2: that's also really important, like there are ways in which
• 46:22
  Speaker 2: people will paint this as, oh, well, you know, this
• 46:25
  Speaker 2: AI race is made up. The thing you have to
• 46:26
  Speaker 2: understand is it's many different domains. So you touched upon it.
• 46:29
  Speaker 2: There's the military domain. There's the surveillance domain, there's the
• 46:33
  Speaker 2: economic domain, where right now, understanding that from the American side,
• 46:38
  Speaker 2: when you have the vast percentage of the total growth
• 46:42
  Speaker 2: of the stock market over the last several years has
• 46:43
  Speaker 2: been wedded to these different companies, you're in a world
• 46:46
  Speaker 2: where a trillion dollars of capital is sort of betting
• 46:50
  Speaker 2: on this existing. At a certain point, that doesn't become
• 46:53
  Speaker 2: a bet that you want to win, that becomes a
• 46:55
  Speaker 2: bet that you need to win. So if China is
• 46:57
  Speaker 2: able to flood our market and the global marketplace with
• 47:00
  Speaker 2: models that are of similar or even higher capabilities that
• 47:03
  Speaker 2: they're giving to you for free, number one, you're now
• 47:06
  Speaker 2: in a place where, okay, if I'm a business, I
• 47:10
  Speaker 2: have to choose, do I want China getting all my data?
• 47:12
  Speaker 2: And do I want to have my compute costs that
• 47:15
  Speaker 2: I'm spending on AI to be $ 20, 000 less? I don't
• 47:19
  Speaker 2: care if China has my data, they're going to get
• 47:20
  Speaker 2: it anyway. I may as well go with a cheaper model.
• 47:22
  Speaker 2: Then you're in a Huawei situation where basically China is
• 47:24
  Speaker 2: able to go to Africa, to Europe, et cetera, and
• 47:27
  Speaker 2: be able to corner the market at giving these free
• 47:29
  Speaker 2: models away And the thing that really is challenging is
• 47:33
  Speaker 2: institutional trust in America as a country. I think it's
• 47:35
  Speaker 2: probably worldwide about at China level at this point. So
• 47:39
  Speaker 2: the idea of us being these trustworthy partners that we
• 47:42
  Speaker 2: can develop technology with, when that's gone and we're more expensive,
• 47:46
  Speaker 2: there's no incentive for Europe or for Africa to want
• 47:49
  Speaker 2: to get in bed with America anymore. So it's really
• 47:51
  Speaker 2: a perfect storm for... you know, right now, what I
• 47:54
  Speaker 2: think we're seeing is China's playbook is going to be
• 47:56
  Speaker 2: to continue flooding America and the world market with absolutely
• 47:59
  Speaker 2: free open source, open way models that you can customize
• 48:02
  Speaker 2: because that's their competitive advantage because all the frontier models,
• 48:05
  Speaker 2: closed source, you can't manipulate them, et cetera, et cetera.
• 48:08
  Speaker 2: So it's, it's really, there's a whole other side of
• 48:12
  Speaker 2: it where the attitude that the Chinese people have towards
• 48:15
  Speaker 2: AI is wildly different than the United States because they've
• 48:18
  Speaker 2: been deploying it in really specific, what are called narrow
• 48:21
  Speaker 2: use categories. So it's, they have an aging population. So
• 48:25
  Speaker 2: it's how are we gonna make this replace jobs that
• 48:29
  Speaker 2: people are aging out of that we don't need to
• 48:31
  Speaker 2: replace with a human being because their population is shrinking
• 48:33
  Speaker 2: and it can then support us. That's really great. In
• 48:36
  Speaker 2: the United States, we're being told, we're gonna replace you
• 48:39
  Speaker 2: at the height of your career. That's very bad. So
• 48:42
  Speaker 2: these differing opinions of the way that AI is gonna go,
• 48:45
  Speaker 2: they've already won the culture war that their country's having
• 48:47
  Speaker 2: internally with how they feel about AI. And for the
• 48:50
  Speaker 2: United States to be competitive in that, There's a great
• 48:52
  Speaker 2: Cat Power song I love called Living Proof. We need
• 48:56
  Speaker 2: an era of living proof. For every time that they
• 48:58
  Speaker 2: said it's going to steal your job, they need to
• 49:00
  Speaker 2: show you 20x, 30x that it's actually going to make
• 49:02
  Speaker 2: your life better. And they're not doing that right now.
• 49:05
  Speaker 2: And if we want to win any AI race, it's
• 49:08
  Speaker 2: predicated on living proof. And I think that hopefully they're
• 49:12
  Speaker 2: starting to wake up for that. I don't know that
• 49:15
  Speaker 2: they will. I don't think they will in time, but
• 49:18
  Speaker 2: we'll see.
• 49:19
  Speaker 1: Now I'm going to take that living proof idea and
• 49:21
  Speaker 1: get us to where I want to close on this,
• 49:23
  Speaker 1: which is what I mentioned people earlier. which is the
• 49:26
  Speaker 1: Ellison article in the New York Times. Did you see that?
• 49:28
  Speaker 2: I have not seen it yet, so you've got to
• 49:29
  Speaker 2: educate me too.
• 49:30
  Speaker 1: Okay, okay. I think when I tell you a little,
• 49:33
  Speaker 1: you'll get it, right? So for those who have not
• 49:34
  Speaker 1: seen it, Larry Ellison, who has once been the richest
• 49:37
  Speaker 1: man in the world, the man behind Oracle, the man
• 49:40
  Speaker 1: who owns his own Hawaiian island, the man whose son
• 49:42
  Speaker 1: is trying to buy Warner Brothers Discovery, basically bankrolled by
• 49:47
  Speaker 1: his father and heavily so in Oracle stock, And that's
• 49:51
  Speaker 1: where this gets interesting, because the article winds up being
• 49:54
  Speaker 1: a discussion about how tenuous his fortune is and how
• 49:57
  Speaker 1: not just his fortune, but also the American economy hinges
• 50:01
  Speaker 1: on the adoption rates of AI. I think they said
• 50:03
  Speaker 1: in the year 2025, 95% of GDP growth in this
• 50:07
  Speaker 1: country was on the building of data centers or just
• 50:09
  Speaker 1: what the AI companies were doing. So what we have
• 50:13
  Speaker 1: is a system or not a system is not the
• 50:15
  Speaker 1: way to put it but basically what's going on now
• 50:17
  Speaker 1: is these companies are doing all this ai building and
• 50:21
  Speaker 1: all these data centers and everything else and they're racking
• 50:24
  Speaker 1: up crazy levels of debt to do so because that
• 50:27
  Speaker 1: is what is going to be necessary in order for
• 50:29
  Speaker 1: them to win this battle but if you pay attention
• 50:32
  Speaker 1: to it it really only only one of them it
• 50:35
  Speaker 1: feels like can win by the notion of how this
• 50:37
  Speaker 1: goes but it's going and going and going and so
• 50:40
  Speaker 1: ellison and oracle have decided they want to be the
• 50:43
  Speaker 1: back end of all this ai stuff and that's where
• 50:46
  Speaker 1: their money comes through but in the process they've had
• 50:48
  Speaker 1: to do all the borrowing and now their bonds are
• 50:51
  Speaker 1: now rated at junk stats and the six so it
• 50:55
  Speaker 1: becomes a dilemma for ai that for people who observe
• 50:58
  Speaker 1: ai from this distance because on one hand we don't
• 51:01
  Speaker 1: want this to get out of control but this has
• 51:03
  Speaker 1: the potential to create an spectacular economic depression because it
• 51:09
  Speaker 1: is the growth right now that is propping up our
• 51:12
  Speaker 1: limp economy. It's this entire sector that is doing it.
• 51:16
  Speaker 1: And so for Ellison, he's now so debt-ridden that he's
• 51:21
  Speaker 1: like a metaphor for everybody else. He borrowed all the
• 51:24
  Speaker 1: money he possibly could, and now he could wind up
• 51:27
  Speaker 1: being ruined. It's partially holding up the Warner Brothers deal.
• 51:31
  Speaker 1: Because it's being backed by his money. It's not terribly
• 51:33
  Speaker 1: different than Elon Musk when he backed up the Twitter
• 51:37
  Speaker 1: purchase with all those shares of Tesla. And it's like, oh,
• 51:40
  Speaker 1: what exactly is going to happen here? And so people
• 51:44
  Speaker 1: are going to wind up in a dilemma about who
• 51:46
  Speaker 1: they root for in this because rooting for the world
• 51:49
  Speaker 1: can also lead to rooting for your own at least
• 51:52
  Speaker 1: short-term economic destruction. Mm-hmm.
• 51:55
  Speaker 2: Well, and the irony is here, I... I won't encourage
• 51:59
  Speaker 2: people to search for it, but there was the most
• 52:02
  Speaker 2: succinct description of the Jenga tower that has been built
• 52:06
  Speaker 2: with AI. Actually, I watched on Newsmax, believe it or not.
• 52:09
  Speaker 2: But they basically pointed out everything you're pointing out, but
• 52:11
  Speaker 2: it goes even further than that. Because once you start
• 52:15
  Speaker 2: seeing the economy dragging in certain sectors, for example, in
• 52:18
  Speaker 2: the entertainment industry, you start seeing people cash out on
• 52:21
  Speaker 2: their pensions. And then when you have people doing that
• 52:24
  Speaker 2: en masse, what that does to their capacity to make investments.
• 52:29
  Speaker 2: And it goes the other way around. If you're afraid
• 52:31
  Speaker 2: that the economy is gonna collapse, you start liquefying all
• 52:35
  Speaker 2: of your assets and then the stock market deflates. There's
• 52:37
  Speaker 2: so many different, there's such an enormous amount of volatility
• 52:41
  Speaker 2: that it's impossible to really feel like we can make
• 52:45
  Speaker 2: any bets on the future. The one intuition that I
• 52:48
  Speaker 2: want people to take away, there's two different things. One
• 52:51
  Speaker 2: is there are really two fallacies The first and most
• 52:55
  Speaker 2: insidious is this is inevitable, that the way that this
• 52:58
  Speaker 2: technology is being developed and deployed is inevitable, that we
• 53:00
  Speaker 2: have no control over it. That is nonsense. And it
• 53:03
  Speaker 2: has always been nonsense. There's many different examples of people
• 53:06
  Speaker 2: developing powerful technologies where my favorite example of this is
• 53:11
  Speaker 2: when they, well, yeah, I'll say the example that I,
• 53:16
  Speaker 2: how many human clones have you seen walking around? Like
• 53:18
  Speaker 2: the incentives to create super soldiers using gene line editing
• 53:21
  Speaker 2: after we had cloning technology is enormous. You know, Organ farms,
• 53:25
  Speaker 2: as far as I know, I don't know what Putin
• 53:27
  Speaker 2: and Xi have got cooking up. I saw their hot mic.
• 53:30
  Speaker 2: I don't know that we have organ farms anywhere in
• 53:32
  Speaker 2: the country. These are things where we as a society
• 53:35
  Speaker 2: decided that this was a bridge too far and we
• 53:36
  Speaker 2: shouldn't cross it. So that's one side. And then on
• 53:39
  Speaker 2: the regulatory side, you know, when we passed the nuclear
• 53:43
  Speaker 2: nonproliferation agreements, I believe in the 60s, we didn't have
• 53:47
  Speaker 2: the technology to even understand whether companies were complying with it.
• 53:51
  Speaker 2: But we passed, we signed those treaties and then people
• 53:53
  Speaker 2: built it. And then we were able to make sure
• 53:55
  Speaker 2: the countries were complying. We are able as a society
• 53:58
  Speaker 2: to steer where we want to go and force these
• 54:01
  Speaker 2: companies to react. So that's one thing. The other one,
• 54:04
  Speaker 2: which comes from more the artist side and the labor side,
• 54:07
  Speaker 2: is the it will never be able to fallacy. It's
• 54:09
  Speaker 2: very tempting to look at the way the capabilities are
• 54:11
  Speaker 2: moving and to say, you know, it's not going to
• 54:14
  Speaker 2: do that. It's not going to get better. It's going
• 54:15
  Speaker 2: to level off here. And being somebody who's watched for
• 54:18
  Speaker 2: three years, hearing people say it's never going to be
• 54:20
  Speaker 2: able to do the hands right. It's never gonna be
• 54:22
  Speaker 2: able to make a song that's indistinguishable from a human song.
• 54:25
  Speaker 2: And you keep seeing time after time after time, they
• 54:28
  Speaker 2: eventually figure it out. So what we have to understand
• 54:31
  Speaker 2: is we have this very narrow window where we are
• 54:34
  Speaker 2: able to rein the technology in. And the very good
• 54:36
  Speaker 2: news is people are aware. And there's a phrase, Beau,
• 54:40
  Speaker 2: that you introduced me to, which is the okey-doke. Yes.
• 54:45
  Speaker 2: The okey-doke in this domain is if they can convince
• 54:50
  Speaker 2: you that you are replaceable and that you have no
• 54:53
  Speaker 2: self-determination and you have no power, then they are able
• 54:56
  Speaker 2: to do whatever they want. Where we find changes, just
• 55:00
  Speaker 2: talking to people that you know about what's going on,
• 55:03
  Speaker 2: people at your company and saying like, hey, I really
• 55:06
  Speaker 2: don't think we should be doing this because people hate
• 55:08
  Speaker 2: this shit. Just being able to show a person with
• 55:10
  Speaker 2: money or if you're working in the finance department saying,
• 55:13
  Speaker 2: we're wasting a shitload of money on these tokens. We're
• 55:15
  Speaker 2: not getting any return out of this. Maybe we should
• 55:17
  Speaker 2: pump the brakes. Or you know, being in your, with
• 55:21
  Speaker 2: your kid's school. Like when you realize that not only
• 55:24
  Speaker 2: are the teachers using GPT to grade the kid's work,
• 55:27
  Speaker 2: but they're using it with their communication. Like these are
• 55:30
  Speaker 2: different things that you can, you can change the way
• 55:33
  Speaker 2: that you're interacting with this technology, that the people in
• 55:36
  Speaker 2: your lives are interacting with it. And yeah, you can just,
• 55:42
  Speaker 2: the other metaphor I give is if you look at
• 55:44
  Speaker 2: your marriage as a series of problems to be solved,
• 55:46
  Speaker 2: you will never relax. You will always feel like you're
• 55:48
  Speaker 2: living in hell and in chaos. But if you set
• 55:50
  Speaker 2: up boundaries, that becomes intuitive. You know where you stand.
• 55:54
  Speaker 2: So when you see it, you know how to react.
• 55:56
  Speaker 2: And rather than feeling like you're playing whack-a-mole, just decide
• 55:59
  Speaker 2: how you want your life to be run vis-a-vis technology
• 56:02
  Speaker 2: and how you want to be deploying it and how
• 56:06
  Speaker 2: you want the people in your life to be deploying it.
• 56:08
  Speaker 2: Just keep having the conversations. And it doesn't have to
• 56:10
  Speaker 2: be stressful. It doesn't have to suck. You just have
• 56:12
  Speaker 2: to coordinate.
• 56:14
  Speaker 1: And maybe good old-fashioned capitalism will save us. This is
• 56:17
  Speaker 1: from a professor at Vanderbilt from that other story. He's like,
• 56:20
  Speaker 1: to me, it's a math problem. We are making trillions
• 56:22
  Speaker 1: of dollars in investments on the back of tens of
• 56:25
  Speaker 1: billions of dollars of revenue. That might be.
• 56:30
  Speaker 2: I'm not an economist, but let me say this. Yeah.
• 56:33
  Speaker 2: I'm an English major, but that makes sense to me.
• 56:35
  Speaker 2: I see it.
• 56:37
  Speaker 1: I see the vision. Yep.
• 56:39
  Speaker 2: Amen.
• 56:40
  Speaker 1: Ted Tripper, I want you guys to check out the
• 56:42
  Speaker 1: AI doc or how I became an apocalyptic optimist. I
• 56:46
  Speaker 1: saved that term for the movie. I think people can
• 56:48
  Speaker 1: figure it out. But I want to check it out.
• 56:49
  Speaker 1: I checked it out. It's absolutely worth your time. I
• 56:51
  Speaker 1: think you can get it on Hulu right now. But
• 56:53
  Speaker 1: if not, you know how it goes.
• 56:54
  Speaker 2: Find it on Peacock. You can rent it from Amazon. Peacock, Amazon,
• 56:59
  Speaker 2: or Apple, you can rent it. And just keep your
• 57:01
  Speaker 2: eyes out because I think it's going to Netflix relatively soon.
• 57:05
  Speaker 2: But just keep your eyes out for it. Or any plane.
• 57:06
  Speaker 2: I've gotten multiple texts last week that it's a great
• 57:09
  Speaker 2: plane movie. It'll make you feel okay if you crash.
• 57:13
  Speaker 2: It's good for you.
• 57:14
  Speaker 1: You are a gentleman and a scholar, sir. I thank
• 57:16
  Speaker 1: you so much.
• 57:17
  Speaker 2: It's so good to see you both.
• 57:19
  Speaker 1: Likewise, man. And ladies and gentlemen, thanks so much for
• 57:21
  Speaker 1: joining us here on The Right Time. We do this
• 57:22
  Speaker 1: a couple times a week during the summer. Mike McQuaid's
• 57:25
  Speaker 1: handling everything behind the scenes. Thank you, sir. Hit voicemail 323-596-7767. Remember,
• 57:32
  Speaker 1: follow The Right Time. Subscribe, like, rate us, review us.
• 57:35
  Speaker 1: Give us five stars. You only give us four stars.
• 57:37
  Speaker 1: I'm inclined to believe you are a hater. And we'll
• 57:39
  Speaker 1: talk to you guys in a couple of days. Take
• 57:41
  Speaker 1: it easy.