Latest / Yaron Brook Show / AI Debate - Gates vs. Huang; France; Wars; Nationalists; More | Yaron Brook Show

Transcript

AI Debate - Gates vs. Huang; France; Wars; Nationalists; More | Yaron Brook Show · Yaron Brook Show

• 0:02
  Speaker 1: The radical, fundamental principles of freedom, rational self-interest, and individual rights.
• 0:14
  Speaker 2: This is the Yaron Brook Show.
• 0:20
  Speaker 1: Alright everybody, welcome to Yaron Brook Show on this Tuesday,
• 0:24
  Speaker 1: October 6th. I apologize for the unusual hour, but I
• 0:29
  Speaker 1: had a board meeting during our Regularly scheduled your own
• 0:36
  Speaker 1: book show, so I figured better to do it later
• 0:38
  Speaker 1: than not do it at all. So here we are
• 0:40
  Speaker 1: at 7 p.m. East Coast time. Thank you for joining me.
• 0:44
  Speaker 1: All right, so we're going to devote a big chunk
• 0:48
  Speaker 1: of the show today to talking about AI. I have
• 0:51
  Speaker 1: some video clips from the Ezra Klein show where he
• 0:56
  Speaker 1: is interviewing people Two different interviews. One was interviewing Jensen Wong,
• 1:04
  Speaker 1: the CEO of NVIDIA. And the other where he's interviewing
• 1:09
  Speaker 1: Bill Gates, the former CEO of Microsoft. And they both
• 1:15
  Speaker 1: take very, very different positions on AI. And we're going
• 1:19
  Speaker 1: to spend a little bit of time talking about, yeah, AI.
• 1:27
  Speaker 1: And here's the thing. that I think is important. I
• 1:31
  Speaker 1: can't think of a more important issue right now, more
• 1:36
  Speaker 1: important question right now. From the perspective of quality of life,
• 1:40
  Speaker 1: standard of living, perspective of the future, certainly the economic,
• 1:46
  Speaker 1: but more than economic, the work future, I can't think
• 1:52
  Speaker 1: of a more important story than AI and what's going
• 1:55
  Speaker 1: to happen with artificial intelligence, how it's going to evolve,
• 1:58
  Speaker 1: how it's going to be regulated, if at all. And,
• 2:02
  Speaker 1: you know, are we going to stop it? Are we
• 2:04
  Speaker 1: going to let it go? What exactly is going to happen?
• 2:08
  Speaker 1: But there's nothing more important than this. It has such
• 2:15
  Speaker 1: potential profound effects on all of us, on our jobs,
• 2:21
  Speaker 1: on our lives, on the way we conduct our lives.
• 2:24
  Speaker 1: And of course, politically, it has huge impact and it
• 2:27
  Speaker 1: has a lot of potential downsides. I mean, the obvious
• 2:30
  Speaker 1: one is it really makes it easy for a surveillance
• 2:36
  Speaker 1: state to arise. You know, we worry about these cameras everywhere,
• 2:41
  Speaker 1: cameras plus AI and cameras plus more and more sophisticated
• 2:45
  Speaker 1: AI will become more and more and more dangerous. But
• 2:52
  Speaker 1: So there's the real downsides. We'll get some of the
• 2:55
  Speaker 1: risks involved in AI. And one of the things I
• 2:58
  Speaker 1: did was after, so I asked Church EPT to actually
• 3:05
  Speaker 1: review the interviews of Wang and Bill Gates and give
• 3:09
  Speaker 1: me its opinions about them. And then, you know, I
• 3:15
  Speaker 1: asked it to assess the risks involved in AI. So
• 3:17
  Speaker 1: I'm going to read you some of the answers that
• 3:20
  Speaker 1: Church EPT gave me. I think it's fascinating to see
• 3:23
  Speaker 1: how it... I'm going to say thinks, although we can
• 3:26
  Speaker 1: agree that it doesn't really think in the conventional sense,
• 3:30
  Speaker 1: but we really don't have the language to talk about AI,
• 3:33
  Speaker 1: so think is the best that we can get to.
• 3:36
  Speaker 1: Beyond that, we will get some other news stories, but
• 3:39
  Speaker 1: we're going to spend a big chunk of the show
• 3:42
  Speaker 1: on this. At the end of the day, what... ultimately
• 3:48
  Speaker 1: is decided and what happens with AI will have a
• 3:52
  Speaker 1: much more profound impact on your life, on all of
• 3:55
  Speaker 1: our lives, than who wins the midterm election and who
• 4:00
  Speaker 1: wins the presidency in 2028. I really do think this
• 4:03
  Speaker 1: technology is world-changing or has the potential to be world-changing.
• 4:08
  Speaker 1: And yeah, so it's important news. All right, so I'm
• 4:14
  Speaker 1: going to play the bits. As you know, I stop
• 4:17
  Speaker 1: them in the middle and I comment on them, and
• 4:19
  Speaker 1: we'll see how this goes. I don't know that I'm
• 4:22
  Speaker 1: going to have that much to say in the middle
• 4:24
  Speaker 1: of what they say, but I might just let them
• 4:29
  Speaker 1: say their bits and then critique it or talk about
• 4:33
  Speaker 1: the differences between the two attitudes. But I really like
• 4:38
  Speaker 1: Jensen Wang, just really like Jensen Wang, so. this would
• 4:44
  Speaker 1: be cool. All right, here we go. Let me see.
• 4:50
  Speaker 1: Let's get this done. There is Ezra Klein, and I
• 4:55
  Speaker 1: will press play. Let me know if there are any
• 4:57
  Speaker 1: audio issues, too loud, too soft, can't hear it at all, whatever.
• 5:00
  Speaker 1: Let me know in the chat if there's an audio
• 5:02
  Speaker 1: issue hearing the other party as well. All right, here
• 5:07
  Speaker 1: we go.
• 5:14
  Speaker 3: Disappointing.
• 5:14
  Speaker 1: I can't hear anything. Okay, here we go.
• 5:17
  Speaker 4: These are CEOs with agency. But they're using that agency
• 5:20
  Speaker 4: to say we need help. No, we've got to break
• 5:22
  Speaker 4: it down. They could absolutely take care of the situation. Ezra,
• 5:26
  Speaker 4: it's so weird. If a car company competing with a
• 5:34
  Speaker 4: whole bunch of other car companies, which they are, I'm
• 5:37
  Speaker 4: competing with all kinds of companies, which I am. If
• 5:40
  Speaker 4: I believe that I'm about to launch a product that
• 5:43
  Speaker 4: is unsafe. It is completely in my ability, my power,
• 5:49
  Speaker 4: and my responsibility, and I'm incentivized to do so, to
• 5:53
  Speaker 4: not launch the product. And so I can't buy into
• 5:57
  Speaker 4: the somehow all of Americans, 400 million of us are
• 6:02
  Speaker 4: pushing them to launch untested products that are unreliable you
• 6:08
  Speaker 4: know engineered poorly because they thought they were trying to
• 6:12
  Speaker 4: help us don't do it for me okay the fact
• 6:15
  Speaker 4: of the matter is there are plenty of incentives for
• 6:18
  Speaker 4: them to do it right so i just i have
• 6:21
  Speaker 4: to disagree with your premise about somehow somebody's pushing them
• 6:25
  Speaker 4: to do this well so so this.
• 6:28
  Speaker 1: Is a great argument right so if your product is
• 6:32
  Speaker 1: no good, if your product is going to kill people,
• 6:34
  Speaker 1: if your product is dangerous, then you as a CEO,
• 6:38
  Speaker 1: you as a company have the ability, he would argue,
• 6:43
  Speaker 1: but I mean, there is a question about how much
• 6:45
  Speaker 1: ability they have when it comes to AI, but we'll
• 6:47
  Speaker 1: get to that later. But you have the ability to stop,
• 6:50
  Speaker 1: not launch the product. You don't have to put out
• 6:53
  Speaker 1: an automobile and that explodes an impact. Now, you're always
• 6:58
  Speaker 1: going to have a risk reward assessment, right? You're always
• 7:01
  Speaker 1: going to look at what are the risks. It's not
• 7:05
  Speaker 1: the case that you put out a product that is
• 7:09
  Speaker 1: 100% safe. I mean, there is no such thing, right?
• 7:14
  Speaker 1: So there's always a risk assessment that has to be done.
• 7:18
  Speaker 1: Each company has to assess what is the threshold of
• 7:21
  Speaker 1: risk that they're willing to take. And then there are
• 7:24
  Speaker 1: mechanisms by which, which I think Jensen Wong is implying,
• 7:29
  Speaker 1: there are mechanisms to penalize companies for bad behavior, for
• 7:34
  Speaker 1: putting out dangerous stuff. Competitors can take advantage of it, right?
• 7:40
  Speaker 1: People are turned off by dangerous products. And, of course,
• 7:44
  Speaker 1: there's the illegal liability mechanism. They will be sued for damages.
• 7:49
  Speaker 1: They will be held responsible for the damage being done.
• 7:53
  Speaker 1: So Jensen Wang is basically saying the CEOs are not
• 8:00
  Speaker 1: forced to put out dangerous products into the market. They
• 8:03
  Speaker 1: choose to do so if they choose and they can
• 8:06
  Speaker 1: choose not to. And it's up to them to choose
• 8:09
  Speaker 1: not to rather than at the end of the day,
• 8:12
  Speaker 1: what they seem to be doing is saying, we have
• 8:15
  Speaker 1: no option. We don't have any choice. We don't have
• 8:17
  Speaker 1: an agency. We need the government to force us to
• 8:24
  Speaker 1: only release safe products. So that is his argument. Now,
• 8:31
  Speaker 1: as a client, he's going to push back a little bit.
• 8:33
  Speaker 5: I'm going to say a little bit more here. So
• 8:37
  Speaker 5: the logic of what you're saying to me is almost
• 8:39
  Speaker 5: an argument against regulation in nearly any venue.
• 8:45
  Speaker 2: No, no, no.
• 8:46
  Speaker 1: By the way, as a client, he's absolutely right. Jensen
• 8:49
  Speaker 1: Wong doesn't want the implication because he doesn't want to
• 8:51
  Speaker 1: come out as, a laissez-faire capitalist guy. But absolutely, the
• 8:56
  Speaker 1: argument is the argument for, you know, very little, if any, regulations.
• 9:02
  Speaker 1: And really, no regulations, as we understand them, in almost
• 9:07
  Speaker 1: every sector of the economy should be gone. I mean,
• 9:10
  Speaker 1: you can imagine very, very, very, very high-risk stuff, right,
• 9:14
  Speaker 1: where there's a clear indication of danger and risk involved
• 9:19
  Speaker 1: that the That's like a sliver.
• 9:24
  Speaker 5: Let me offer it and then you can.
• 9:25
  Speaker 4: Well, you started with a part. I was just going
• 9:27
  Speaker 4: to object. The first part is just not true. I'm
• 9:30
  Speaker 4: saying we have lots of laws and regulations.
• 9:32
  Speaker 5: Apply it. Well, so I don't think we do in
• 9:34
  Speaker 5: this particular case, but I'll let you explain which ones
• 9:36
  Speaker 5: you think are relevant here. Because, look, if you look
• 9:40
  Speaker 5: at the financial services industry, you look at pharmaceutical companies,
• 9:43
  Speaker 5: medical devices.
• 9:46
  Speaker 1: I mean, it's kind of funny. Because as your client
• 9:48
  Speaker 1: should know better. I mean, financial regulations are what create
• 9:53
  Speaker 1: the problems. It doesn't solve the problems.
• 9:56
  Speaker 5: You look at natural gas power plants.
• 9:59
  Speaker 2: There is a.
• 10:00
  Speaker 5: tremendous amount we do where we could say, look, you
• 10:05
  Speaker 5: have product liability.
• 10:06
  Speaker 1: Yes, they do.
• 10:07
  Speaker 5: You are exposed to criminal codes.
• 10:08
  Speaker 1: Yep.
• 10:09
  Speaker 5: We don't need to worry about this.
• 10:10
  Speaker 1: That's right.
• 10:11
  Speaker 5: You just do what you think is best, and we
• 10:13
  Speaker 5: understand the market and the legal system will discipline you.
• 10:16
  Speaker 1: That's right.
• 10:17
  Speaker 5: We don't say that because we've seen it fail many, many,
• 10:20
  Speaker 5: many times.
• 10:20
  Speaker 1: We have not seen it fail many, many, many times.
• 10:22
  Speaker 1: That's one of the myths about capitalism and about free markets.
• 10:25
  Speaker 1: We haven't seen it fail many, many times. We should,
• 10:28
  Speaker 1: we could, and we could. We could and we should
• 10:34
  Speaker 1: leave the market and the legal system to deal with
• 10:38
  Speaker 1: most of these issues, particularly primarily financial markets. So again,
• 10:44
  Speaker 1: when you start telling yourself a story about the dangers
• 10:47
  Speaker 1: of markets, market failures, companies do dangerous stuff, and it's
• 10:52
  Speaker 1: the government that saves us. You're on a bad track.
• 10:56
  Speaker 5: So when you say to me that there's no way
• 11:00
  Speaker 5: that these companies, particularly when they are begging for collective
• 11:03
  Speaker 5: regulation at this point, there's both a reason we impose
• 11:07
  Speaker 5: it on companies that don't want it, but all the
• 11:09
  Speaker 5: more so when you have them saying, listen, we feel
• 11:12
  Speaker 5: that the competitive race is making it hard for us
• 11:15
  Speaker 5: to act with the prudence that we think is necessary here.
• 11:20
  Speaker 5: And we would appreciate help from that. Appreciate you taking
• 11:24
  Speaker 5: our collective action problem as collective. I think I'm confused
• 11:28
  Speaker 5: why you're so resistant to that.
• 11:30
  Speaker 4: I'm not opposed to them saying that they should have.
• 11:36
  Speaker 4: I completely agree that safety is paramount. When you're asking
• 11:39
  Speaker 4: for regulation, don't ask for relief of the current ones.
• 11:44
  Speaker 4: That doesn't make any sense to me. These are CEOs
• 11:46
  Speaker 4: with agency.
• 11:48
  Speaker 1: So, sadly, Jensen Wang is not coming out and saying
• 11:53
  Speaker 1: what he should say, which is, no, you know, all
• 12:01
  Speaker 1: these other industries shouldn't be regulated either. Markets do work.
• 12:05
  Speaker 1: They do. We can have that debate. So now he's
• 12:08
  Speaker 1: put into a corner by saying, look, we've got all
• 12:11
  Speaker 1: these regulations on these other industries. Why can't we have
• 12:13
  Speaker 1: it on here? And particularly when the industry is asking
• 12:15
  Speaker 1: for it, don't they know? The whole assumption here is
• 12:20
  Speaker 1: that the profit motive, and this is deeply embedded in
• 12:24
  Speaker 1: the psyche of people, the profit motive leads to bad behavior.
• 12:32
  Speaker 1: The profit motive and competition leads to increased risk. That
• 12:42
  Speaker 1: we need to be protective for ourselves, right? If you
• 12:45
  Speaker 1: leave us alone, we'll do horrible things. So we need
• 12:48
  Speaker 1: the government to step in to protect ourselves. And, of course,
• 12:56
  Speaker 1: that is all a myth because it comes out of,
• 13:02
  Speaker 1: it comes from a fundamental misunderstanding of the profit motive.
• 13:06
  Speaker 1: It comes from a fundamental misunderstanding of self-interest, a suspicion
• 13:10
  Speaker 1: of self-interest, worse than a suspicion of self-interest, a conviction.
• 13:14
  Speaker 1: the self-interest is and leads to evil. That's where this
• 13:21
  Speaker 1: kind of stuff comes from. And yeah, I mean, what
• 13:28
  Speaker 1: you get is even when somebody like Jensen Wong, who's
• 13:31
  Speaker 1: fairly good, tries to defend it, he can't. He can't really.
• 13:37
  Speaker 1: Now let's turn on the same issue to Bill Gates,
• 13:41
  Speaker 1: who I think is in this sense, more consistent, right?
• 13:46
  Speaker 1: Because they think he's for regulation. He's just, he doesn't
• 13:51
  Speaker 1: believe in the profit motive, at least the way he
• 13:53
  Speaker 1: comes across here. Here's Bill Gates. creating memories.
• 14:06
  Speaker 5: It's adorable, but no.
• 14:07
  Speaker 1: Sorry, we've got advertising.
• 14:08
  Speaker 2: It's about having a.
• 14:09
  Speaker 4: City Strada Elite card so you can relax in the
• 14:11
  Speaker 4: Admiral's Club until the last possible second.
• 14:14
  Speaker 5: I almost.
• 14:18
  Speaker 3: Can't believe you're asking that. This is the most dangerous
• 14:21
  Speaker 3: thing that humans have ever gone near.
• 14:24
  Speaker 1: So this is the most dangerous thing humans have ever
• 14:29
  Speaker 1: gone near. This being AI. Why is it this dangerous?
• 14:34
  Speaker 1: What makes it so dangerous? The most dangerous thing humans
• 14:40
  Speaker 1: have ever gone near.
• 14:42
  Speaker 3: In other areas, do we just say, hey, release your drugs?
• 14:47
  Speaker 3: There's no FDA.
• 14:49
  Speaker 1: Yeah, we should. We should. But he's right. We don't.
• 14:53
  Speaker 3: There's no airline safety board. There's no requirements that cars
• 14:57
  Speaker 3: use seatbelts. Do we just use the liability laws to
• 15:00
  Speaker 3: try and keep humans safe? You know, oh, you're shipping opioids.
• 15:03
  Speaker 3: Somebody should just sue you. I mean, we've created a
• 15:07
  Speaker 3: society that tries to keep people safe, not by saying, oh,
• 15:12
  Speaker 3: we can bankrupt the person who does that. The harms here,
• 15:17
  Speaker 3: and you say there's filtering. There's not filtering. You can
• 15:20
  Speaker 3: take an open source model that can create bioweapons and disable.
• 15:24
  Speaker 2: Any monitoring of any kind. And this exists today.
• 15:31
  Speaker 1: So he's right. There are real dangers here. There are
• 15:34
  Speaker 1: real dangers, particularly with the open source models, which are
• 15:39
  Speaker 1: not controlled and people can change them and, as he said,
• 15:44
  Speaker 1: disable certain restrictions that are placed in it. And it can,
• 15:51
  Speaker 1: you know, and somebody can devise a biological weapon. People
• 15:56
  Speaker 1: can design biological weapons, even today. That information is available
• 16:00
  Speaker 1: out there. AI makes it a little bit more accessible
• 16:03
  Speaker 1: and available. But this, I mean, he's really panicking here.
• 16:08
  Speaker 1: The most dangerous thing We've ever come near?
• 16:12
  Speaker 3: There is no filtering of any kind. And so, you know, say,
• 16:16
  Speaker 3: you know, you kill 100 million. You want to use
• 16:19
  Speaker 3: a lawsuit? I almost can't keep a straight face.
• 16:22
  Speaker 5: Well, this is not my view, but it is President
• 16:24
  Speaker 5: Trump's view. It is David Sachs's view. To some degree,
• 16:27
  Speaker 5: it's Jensen Huang's view. And so that's why I'm putting
• 16:29
  Speaker 5: you in conversation with it, because it is the governing
• 16:31
  Speaker 5: view of the United States of America at this moment.
• 16:35
  Speaker 3: No, it's fair to say that outside of the industry,
• 16:39
  Speaker 3: the awareness... of the dangers of AI is extremely low.
• 16:45
  Speaker 3: And you can say that of academia, you can say
• 16:49
  Speaker 3: that of think tanks, you can say that of policymakers, politicians.
• 16:54
  Speaker 3: And part of the reason that, you know, I'm speaking
• 16:59
  Speaker 3: so loudly, as loud as I can, is that you
• 17:03
  Speaker 3: can't rely on the industry to self-regulate here.
• 17:07
  Speaker 1: Why not? I mean, Again, this is always the question,
• 17:11
  Speaker 1: why can't you?
• 17:15
  Speaker 2: I mean, it's just insane.
• 17:17
  Speaker 3: The only question in my mind is, do we wait
• 17:21
  Speaker 3: until a cyber attack causes massive damage and a bio
• 17:26
  Speaker 3: attack causes massive damage, and then the monitoring safeguards are
• 17:31
  Speaker 3: required in these models to minimize the chance of that
• 17:35
  Speaker 3: happening many, many more times? Or Can we be wise
• 17:39
  Speaker 3: enough to put these things in and require these things
• 17:42
  Speaker 3: to be put in before millions of deaths?
• 17:47
  Speaker 1: So, look, there is a role for government here because
• 17:53
  Speaker 1: this is an incredibly powerful tool that can be used really, really, really,
• 17:58
  Speaker 1: you know, for really, really, really evil purposes. But it's
• 18:06
  Speaker 1: still so early. So if we knew exactly what needed
• 18:10
  Speaker 1: to be done in order to make it impossible for
• 18:13
  Speaker 1: people to use AI to build biological weapons, then it
• 18:19
  Speaker 1: would be pretty straightforward. You could, and I think it
• 18:24
  Speaker 1: would be legitimate, you'd have to think about how to
• 18:27
  Speaker 1: phrase this exactly, but you could pass a law saying
• 18:32
  Speaker 1: it is illegal to provide information for the design and
• 18:38
  Speaker 1: construction of biological weapons. Assuming you know what a biological
• 18:43
  Speaker 1: weapon is and how to differentiate that from other pharmaceuticals
• 18:48
  Speaker 1: that you might be building. So if we knew exactly
• 18:52
  Speaker 1: how to phrase that, if we knew how to write
• 18:54
  Speaker 1: a law like that, and if AI companies could actually
• 18:57
  Speaker 1: implement something like that, then yeah, this is easy.
• 19:00
  Speaker 4: Sure.
• 19:00
  Speaker 1: Why not? There's no rights being violated there. Zero. Zero.
• 19:05
  Speaker 1: Or if you said, you know, cyber attacks are a crime.
• 19:13
  Speaker 1: It's probably true already in criminal law. We're going to
• 19:17
  Speaker 1: raise the stakes here. We might even make liable the
• 19:24
  Speaker 1: people who make the tools to make it possible to
• 19:26
  Speaker 1: do the cyber attacks. So in your models, make sure
• 19:29
  Speaker 1: they're not usable for cyber attacks. And if we knew
• 19:35
  Speaker 1: exactly how to do that and if the companies knew
• 19:37
  Speaker 1: how to do that, they would do it. I mean,
• 19:39
  Speaker 1: my guess is the companies have done their best already,
• 19:43
  Speaker 1: already to make it difficult to figure out how to
• 19:47
  Speaker 1: build a biological weapon. I think that's true of Anthropic
• 19:51
  Speaker 1: and true of OpenAI. So they're already doing it. It's
• 20:01
  Speaker 1: not like Anthropic and OpenAI are going to make more
• 20:04
  Speaker 1: money by not putting in safeguards. There's no profit to
• 20:10
  Speaker 1: be made in selling biological weapons, particularly if you're liable
• 20:15
  Speaker 1: for the results. Do we really want politicians and the
• 20:21
  Speaker 1: political mechanism to get involved in determining these technical issues?
• 20:26
  Speaker 1: I mean, the people actually designing these AI systems are
• 20:30
  Speaker 1: not sure exactly how they work. politicians going to figure
• 20:39
  Speaker 1: this out?
• 20:40
  Speaker 2: Really?
• 20:40
  Speaker 1: How well have they regulated financial institutions? Have we really
• 20:46
  Speaker 1: eliminated bank runs? Have we eliminated? No, the answer. Have
• 20:51
  Speaker 1: we eliminated, you know, financial crises? Obviously not. So that's
• 20:59
  Speaker 1: much easier, you would think, than AI. So why would
• 21:03
  Speaker 1: we think? Why would anybody suggest that AI regulation is
• 21:09
  Speaker 1: going to be more effective. It's just bizarre. And here's
• 21:18
  Speaker 1: the other issue. That, you know, anybody could start and
• 21:25
  Speaker 1: can write software or create an algorithm, create a new AI.
• 21:34
  Speaker 1: Any company, anywhere in the world, with enough data to
• 21:37
  Speaker 1: train the thing on, and with enough sophistication in terms
• 21:41
  Speaker 1: of the algorithm. And we have a lot of open
• 21:45
  Speaker 1: AI models that are sitting on people's computers, that they
• 21:48
  Speaker 1: run them natively in the computer. And the government has
• 21:51
  Speaker 1: no control of what they do there. Can somebody go
• 21:54
  Speaker 1: into these models and take out the safeguards? Bill Gates
• 21:57
  Speaker 1: is saying yes. The answer is probably yes. So the
• 22:03
  Speaker 1: reality of it, whether we like it or not, and
• 22:07
  Speaker 1: we'll see what ChatGPT has to say about this in
• 22:08
  Speaker 1: a minute, is that we're going to get a lot
• 22:12
  Speaker 1: of cybersecurity issues, as I've said in past shows. The
• 22:16
  Speaker 1: only way to stop those is to make AI available
• 22:22
  Speaker 1: to companies to protect themselves, to build the protective models
• 22:28
  Speaker 1: against AI used to try to hack into them. I
• 22:36
  Speaker 1: don't know what you do with bioterrorism. I mean, the
• 22:39
  Speaker 1: good part is, the lucky part is, the AI is
• 22:44
  Speaker 1: not going to actually build the weapon for you. It's
• 22:46
  Speaker 1: going to give you instructions. It's still complicated. It's still messy.
• 22:50
  Speaker 1: The kind of people who want to build bioweapons are
• 22:53
  Speaker 1: probably going to make mistakes. They're probably going to kill
• 22:55
  Speaker 1: themselves before they kill anybody else. And will a catastrophe happen?
• 23:02
  Speaker 1: Probably not. will it be such that all of humanity dies? Very, very, very, very, very, very,
• 23:10
  Speaker 1: very unlikely. So no, there's zero reason why, um, We
• 23:30
  Speaker 1: should expect the world to end. I mean, bad things
• 23:34
  Speaker 1: might happen. New technologies are introduced and, you know, they're used.
• 23:38
  Speaker 1: I mean, isn't it amazing, by the way? I mean,
• 23:43
  Speaker 1: I remember being alive for a long time. But I
• 23:46
  Speaker 1: remember over and over and over again the stories about
• 23:51
  Speaker 1: bad guys are going to use nukes. Terrorists are going
• 23:53
  Speaker 1: to get their hands on a nuclear bomb. They can
• 23:56
  Speaker 1: put it in a suitcase, and nuclear bombs are going
• 23:59
  Speaker 1: to go off all over the place. Do you remember 24,
• 24:03
  Speaker 1: the show 24, which was a really good TV show?
• 24:06
  Speaker 1: I mean, nuclear bombs went off all the time on
• 24:09
  Speaker 1: that show. It's so easy to blow up. And yet,
• 24:13
  Speaker 1: in reality, none have, literally none have, other than Hiroshima
• 24:18
  Speaker 1: and Nagasaki, no nuclear bombs have gone off since then.
• 24:23
  Speaker 1: It's hard. It's always harder than you believe. And there
• 24:28
  Speaker 1: are many, many ways to stop you. Good intelligence, good
• 24:33
  Speaker 1: intelligence agencies using AI for intelligence agencies, for example. Track
• 24:38
  Speaker 1: down people who might be considering using biological weapons, for example.
• 24:46
  Speaker 1: So the idea that doom is inevitable is I think ridiculous.
• 24:55
  Speaker 1: The idea that we can't do anything about it is ridiculous.
• 25:01
  Speaker 1: The idea that the companies can't do anything about it,
• 25:04
  Speaker 1: at least where the companies control the models, which are
• 25:06
  Speaker 1: the most powerful models is just untrue. They do control
• 25:11
  Speaker 1: it as Jason Wong says. So, um, yeah, I mean, uh,
• 25:20
  Speaker 1: To me, it's– this is an important debate, and the
• 25:29
  Speaker 1: risks should be put out there on the table, and
• 25:34
  Speaker 1: people need to confront the risks, and the primary people
• 25:36
  Speaker 1: need to confront the risks are the people producing the product.
• 25:39
  Speaker 1: And there's every indication that they are, and this panic
• 25:43
  Speaker 1: about we need the government, we need the government, we
• 25:45
  Speaker 1: need government regulation is, in my view, insane given the
• 25:49
  Speaker 1: history– and lack of success of government regulation. I mean,
• 25:53
  Speaker 1: what government regulation did to nuclear power is kill it,
• 25:57
  Speaker 1: not make it safe, just kill it. So, no, I mean,
• 26:09
  Speaker 1: given the upside, given the upside of this technology, to
• 26:16
  Speaker 1: get the government involved would be a disaster, and to
• 26:18
  Speaker 1: slow down The development would be a disaster. Let's listen
• 26:25
  Speaker 1: to Jensen Wong again. Let me see. We're going from, yeah. Yeah,
• 26:32
  Speaker 1: it's about six minutes here. See what he has to say.
• 26:38
  Speaker 5: Possible. That is not possible now. What is common that
• 26:42
  Speaker 5: is not common now. If we get that layer right.
• 26:45
  Speaker 5: 200 years ago, we were able to power anything. and everything, electricity.
• 26:52
  Speaker 4: And then, I guess, 40 years ago, 30 years ago,
• 26:59
  Speaker 4: with the internet, we were able to find anything.
• 27:04
  Speaker 1: Today, or.
• 27:06
  Speaker 4: Soon, we'll be able to know everything and do anything.
• 27:11
  Speaker 4: And that's the concept that's really quite exciting, that out.
• 27:15
  Speaker 1: Of the ether, Know everything and do anything, which is,
• 27:19
  Speaker 1: I mean, that's powerful stuff. I mean, know everything that
• 27:24
  Speaker 1: is known, do anything that is possible, of course. But
• 27:27
  Speaker 1: that is incredibly powerful and it's incredibly powerful technology and
• 27:31
  Speaker 1: incredibly pro-human technology.
• 27:32
  Speaker 4: Instead of doing search and then going through, you know,
• 27:37
  Speaker 4: one link after another link, reading all these different websites,
• 27:40
  Speaker 4: trying to figure out what's going on.
• 27:42
  Speaker 2: In the future, you just.
• 27:43
  Speaker 4: Ask it a question, and it comes back with an answer.
• 27:45
  Speaker 4: You give it a project, it comes back with a solution.
• 27:47
  Speaker 4: You give it a task, it comes back and gets
• 27:49
  Speaker 4: it done. And it comes out of the ether, it
• 27:52
  Speaker 4: comes out of the cloud. And that's the magical thing.
• 27:56
  Speaker 5: I feel like the future, the way you're describing it there,
• 27:58
  Speaker 5: what people have experience with is the chatbot. They can
• 28:01
  Speaker 5: go and ask Grok or Claude or ChatGPT a question.
• 28:07
  Speaker 5: But the applications layer works differently. In a much more
• 28:10
  Speaker 5: industrial way. It's in hospitals. It's in schools. So NVIDIA
• 28:15
  Speaker 5: is a great example. For example, radiology. What does it
• 28:17
  Speaker 5: look like?
• 28:18
  Speaker 2: Radiology.
• 28:20
  Speaker 4: In the last 10 years since computer vision really became,
• 28:24
  Speaker 4: if you will, superhuman, AI technology has now permeated all
• 28:28
  Speaker 4: of radiology. Every single radiology application has AI in it.
• 28:33
  Speaker 4: And so as a result, you could detect any anomaly.
• 28:36
  Speaker 4: You could detect any disease. And it does it at
• 28:38
  Speaker 4: a superhuman level. So radiology is an example I know
• 28:43
  Speaker 4: you like to use. So the thing people worry about
• 28:45
  Speaker 4: the applications layer is that what these applications are going
• 28:47
  Speaker 4: to do is replace human beings. And radiology has been
• 28:51
  Speaker 4: a sort of interesting example used on both sides.
• 28:54
  Speaker 5: And I hear you talk a bit often. So how
• 28:57
  Speaker 5: has the entrance of AI aided radiology shifted radiology as
• 29:04
  Speaker 5: a practice?
• 29:05
  Speaker 4: Well, um, The thing that's important for all of these
• 29:09
  Speaker 4: is to recognize for everybody's job, there's the purpose of
• 29:13
  Speaker 4: the job, and then there's the task you do as
• 29:16
  Speaker 4: the job. And in the case of radiology, the task,
• 29:21
  Speaker 4: and it consumes a lot of their time, and they
• 29:23
  Speaker 4: sit in dark rooms doing it a lot, which is
• 29:26
  Speaker 4: study these scans. Now, if all of a sudden the
• 29:29
  Speaker 4: studying of the scan is done automatically, it doesn't change
• 29:32
  Speaker 4: the purpose of their job, which is to diagnose disease,
• 29:36
  Speaker 4: help doctors, Do more scans ultimately help patients figure out
• 29:39
  Speaker 4: what's wrong with them. And so the fundamental purpose doesn't change.
• 29:44
  Speaker 4: The task of studying that scan has become automated. And
• 29:48
  Speaker 4: so as a result, radiologists are actually able to do more,
• 29:51
  Speaker 4: handle more cases, do more scans.
• 29:54
  Speaker 5: Hospitals are able to process a lot.
• 29:56
  Speaker 4: More of these patients and therefore their revenues go up.
• 30:00
  Speaker 4: As a result, they need more radiologists. And so this
• 30:04
  Speaker 4: flywheel is happening because the pipeline of patients is quite large.
• 30:12
  Speaker 4: And so where else do you have this problem? Well,
• 30:14
  Speaker 4: let's take a look at software engineering. People said there
• 30:18
  Speaker 4: was a prediction that literally by this year that 90%
• 30:22
  Speaker 4: of all software will be coded by agents and And therefore,
• 30:25
  Speaker 4: we don't need any software engineers. And so the question is,
• 30:29
  Speaker 4: from that, ergo, we don't need software engineers. That last
• 30:33
  Speaker 4: part is completely false. That's completely wrong. The purpose of
• 30:38
  Speaker 4: the software engineer is engineer. There was engineering before software.
• 30:45
  Speaker 4: There will be engineering after software programming. And the purpose
• 30:49
  Speaker 4: of engineering is to invent something new, discover a new product,
• 30:54
  Speaker 4: create a new product, solve a problem, connect a social
• 30:57
  Speaker 4: need with the technology that exists in.
• 31:02
  Speaker 5: The manifestation of a product. And so that mission, that
• 31:06
  Speaker 5: purpose doesn't change.
• 31:08
  Speaker 4: Now, of course, to me, what I just said is
• 31:11
  Speaker 4: completely visceral in the sense that when I first came
• 31:14
  Speaker 4: out of school, we didn't have benefits of software engineering.
• 31:18
  Speaker 5: We didn't have the benefits of coding. But our jobs
• 31:21
  Speaker 5: existed before.
• 31:22
  Speaker 4: And if software coding was to be completely automated, our
• 31:26
  Speaker 4: jobs would exist again. And so I think the fallacy,
• 31:30
  Speaker 4: and now it's, you know, because of some of the
• 31:32
  Speaker 4: narratives and some of the storytelling has turned into myth
• 31:36
  Speaker 4: and it's harmful, is that AI will destroy jobs, which
• 31:41
  Speaker 4: is fundamentally wrong.
• 31:42
  Speaker 1: It will change every job. It will I think he's
• 31:45
  Speaker 1: absolutely right. I mean, and the way he's explained it
• 31:48
  Speaker 1: is really, really good. It will not destroy jobs. It
• 31:52
  Speaker 1: will change jobs. It will increase the number of jobs needed.
• 31:57
  Speaker 1: But those jobs will be very different. And the jobs
• 32:00
  Speaker 1: and the activities that you will engage in a job
• 32:02
  Speaker 1: will be very different. And as AI evolves, those jobs
• 32:06
  Speaker 1: will evolve with it.
• 32:07
  Speaker 5: It will change every job. Many tasks will be automated.
• 32:12
  Speaker 5: Some jobs.
• 32:13
  Speaker 4: where the job and the task is really one, meaning
• 32:18
  Speaker 4: customer service on the phone, in a lot of cases,
• 32:24
  Speaker 4: that job is precisely the task. And so in those cases,
• 32:29
  Speaker 4: it could be automated away. But oftentimes what you'll see
• 32:32
  Speaker 4: is this new industry, a new technology actually creates a
• 32:37
  Speaker 4: whole bunch of new jobs. And here's the proof point.
• 32:40
  Speaker 4: And so in the last six months, AI has become,
• 32:45
  Speaker 4: if you will, useful. The inflection point of AI. Previous
• 32:48
  Speaker 4: to that, we spent 15 years trying to make it work.
• 32:51
  Speaker 4: All of a sudden, the last six months, it became useful. So,
• 32:54
  Speaker 4: I mean, this is an incredible statistic. In the last
• 32:56
  Speaker 4: six months, $ 500 billion of venture capital is put into
• 33:02
  Speaker 4: the AI natives. And the reason for that is because
• 33:05
  Speaker 4: they now see the potential of this new capability and
• 33:07
  Speaker 4: they're going to.
• 33:07
  Speaker 5: Create a whole bunch of new companies.
• 33:09
  Speaker 4: Jobs are obviously being created from $ 500 billion of new investment.
• 33:13
  Speaker 4: And so all of this is all happening right now. Well,
• 33:16
  Speaker 4: let me take the side of this to get voiced.
• 33:18
  Speaker 1: All right. So he's saying, oh, jobs are going to grow.
• 33:22
  Speaker 1: More jobs are going to be created. There's going to
• 33:24
  Speaker 1: be more opportunities, not fewer. And I think he's absolutely right.
• 33:30
  Speaker 1: And you don't hear enough of this kind of attitude
• 33:35
  Speaker 1: towards AI today. Technology creates more jobs than it destroys.
• 33:41
  Speaker 1: That's always been the history. And there's nothing to suggest
• 33:44
  Speaker 1: that this technology is going to be any different. By
• 33:50
  Speaker 1: the way, if you want to ask a question and
• 33:52
  Speaker 1: have me answer it, then Super Chat. Super Chat is
• 33:55
  Speaker 1: the way to do it. You can do it for $ 2, $ 5.
• 33:58
  Speaker 1: I answer all the questions that are being asked. So, hey,
• 34:03
  Speaker 1: I know there are a lot of you chatting over there,
• 34:05
  Speaker 1: a lot of you asking questions. I don't know if
• 34:07
  Speaker 1: you want me to answer them, but if you do
• 34:09
  Speaker 1: want me to answer them, if you want me to
• 34:12
  Speaker 1: read them even, because I don't have time to read
• 34:14
  Speaker 1: them as I'm talking, then put them in a super chat.
• 34:18
  Speaker 1: I read all the super chats out loud to everybody,
• 34:20
  Speaker 1: and I try to answer all of them. All right.
• 34:24
  Speaker 1: Let's now take a look at Bill Gates on the
• 34:26
  Speaker 1: same issue, on the issue of jobs. We've got commercials again.
• 34:36
  Speaker 1: All right. I don't know that I could have gotten
• 34:38
  Speaker 1: rid of the commercials on these ones.
• 34:48
  Speaker 5: Tell me about why you think AI really will take
• 34:50
  Speaker 5: jobs away from human beings at a significant scale.
• 34:54
  Speaker 3: The superiority of these systems is subject to a threshold
• 34:59
  Speaker 3: where you have to believe that it's incredibly reliable. And
• 35:02
  Speaker 3: so the only profession we've truly crossed over that threshold
• 35:07
  Speaker 3: is coding. We will, in the next few years, cross
• 35:10
  Speaker 3: over those thresholds for accounting, legal work, telesales, telesupport, where
• 35:16
  Speaker 3: 24 hours a day in every language with infinite trivia
• 35:21
  Speaker 3: capability and no urgency. So you look at the AI
• 35:25
  Speaker 3: and nurse and there's several companies they have perfect memory
• 35:30
  Speaker 3: they don't hurry to.
• 35:31
  Speaker 5: Be like a nurse you would call on the phone
• 35:32
  Speaker 5: or talk through a chat interface yeah far preference you
• 35:36
  Speaker 5: know you go to san francisco and ask people would
• 35:38
  Speaker 5: you rather ride in a way more rather ride with
• 35:40
  Speaker 5: the human driver go ask people.
• 35:42
  Speaker 1: They'd rather drive away more of course and he's absolutely
• 35:45
  Speaker 1: right all these jobs are going to be replaced by
• 35:48
  Speaker 1: ai but does that mean jobs are going to be
• 35:50
  Speaker 1: destroyed the job's going to disappear there can be no
• 35:52
  Speaker 1: alternative jobs of course not of course not But those
• 36:01
  Speaker 1: jobs will be, if you're in telemarketing, if you're anything
• 36:07
  Speaker 1: that can be easily replicated by a chat box, by
• 36:11
  Speaker 1: a large language model, yeah, you're going to be out
• 36:14
  Speaker 1: of a job. You better prepare yourselves. People are not prepared.
• 36:19
  Speaker 1: They're not prepared, which is not good.
• 36:21
  Speaker 3: In the UK, you use Limbic for mental health support.
• 36:25
  Speaker 3: The cost for some of these things is so much
• 36:27
  Speaker 3: less than than the cost of the human labor. And
• 36:30
  Speaker 3: so as you cross reliability thresholds, both with white collar
• 36:34
  Speaker 3: and humanoid robots, you destroy jobs. And that's true.
• 36:41
  Speaker 1: But does that mean no other jobs are created? I mean,
• 36:44
  Speaker 1: when you moved to automatic, you know, powered sewing machines,
• 36:49
  Speaker 1: when you move to industrialization, when you move to what
• 36:55
  Speaker 1: is it, a tractor and a backhoe and today high-tech
• 36:59
  Speaker 1: farming equipment. Yeah, jobs were destroyed. Jobs were destroyed. New
• 37:05
  Speaker 1: jobs came about. Human needs, human desires, human wants. There's
• 37:12
  Speaker 1: no limit to those. So new jobs will come about
• 37:17
  Speaker 1: in those areas that need human participation, those areas where
• 37:21
  Speaker 1: AI cannot or cannot cannot easily or cannot quickly replace
• 37:26
  Speaker 1: human beings.
• 37:27
  Speaker 2: You leave no high ground.
• 37:29
  Speaker 3: You know, people who say there'll be net additional jobs,
• 37:32
  Speaker 3: I don't understand what they're thinking.
• 37:34
  Speaker 1: Why don't you understand? I mean, that's been the case
• 37:36
  Speaker 1: for every technology in human history.
• 37:41
  Speaker 3: I don't understand what they're thinking. They must not understand
• 37:44
  Speaker 3: the piece of improvement we're on. It's always about scarce elements.
• 37:48
  Speaker 3: Once you replace human cognition as a scarce element.
• 37:51
  Speaker 1: Are you really going to... replace human cognition as a
• 37:56
  Speaker 1: scarce element. I mean, and this is the issue. Are
• 37:59
  Speaker 1: these machines really thinking? Are they engaged in cognition? Are
• 38:04
  Speaker 1: they really engaged in discovery of new knowledge? Are they
• 38:08
  Speaker 1: engaged in creating new products? Are they creating what is
• 38:13
  Speaker 1: uniquely human? No. You know, they're just not. There's, You know,
• 38:22
  Speaker 1: and if you play around with AI, you see this.
• 38:26
  Speaker 1: I mean, I think AI is amazing. I'm blown away
• 38:29
  Speaker 1: by AI and I use it every single day. But
• 38:34
  Speaker 1: it is so limited. And it's not clear how just
• 38:39
  Speaker 1: based on learning from words, learning from language, right, and
• 38:46
  Speaker 1: integrating more language and more words, it's going to get
• 38:51
  Speaker 1: To a point where it replaces all human jobs. It just...
• 38:57
  Speaker 1: That is, I think, still science fiction type thinking.
• 39:01
  Speaker 3: In fact, you blow it away in terms of working
• 39:04
  Speaker 3: 24 hours a day, reading more, knowing every language. There
• 39:08
  Speaker 3: is no scarcity that you're moving up.
• 39:10
  Speaker 5: So you feel there is no scarcity that will be
• 39:12
  Speaker 5: left for.
• 39:12
  Speaker 2: Human beings here? Name the scarcity.
• 39:15
  Speaker 5: Alex would say it's something like relational sector jobs, but
• 39:17
  Speaker 5: you don't believe there are enough of those.
• 39:18
  Speaker 2: What portion of the current jobs... You.
• 39:20
  Speaker 3: Mean like my relationship with the cab driver or my
• 39:23
  Speaker 3: relationship with that nurse where I'd rather have Hippocratic AI
• 39:26
  Speaker 3: call me. And we have the analogies with the past
• 39:29
  Speaker 3: where people are saying, well, this is like the PC,
• 39:32
  Speaker 3: not like evolutionary history. This is like evolutionary history.
• 39:36
  Speaker 1: So he's saying this is evolutionary history. This is a
• 39:39
  Speaker 1: new being. This is a new creature, a new being
• 39:42
  Speaker 1: that is basically going to wipe out human beings because
• 39:46
  Speaker 1: it's superior to us. So it's going to compete us away.
• 39:51
  Speaker 1: We can't compete. It's just better at us than everything
• 39:55
  Speaker 1: that's being done. Except that it has no values. It
• 39:59
  Speaker 1: doesn't think. It has no desires, motivations, and wants. Other
• 40:04
  Speaker 1: than that, yeah, it's a new... I mean, this is
• 40:07
  Speaker 1: the problem of a confusing electronics with biology. Confusing electronics
• 40:13
  Speaker 1: with biology.
• 40:15
  Speaker 3: This is like the aliens really are here and You know,
• 40:20
  Speaker 3: have come, they didn't have to do spacecraft, they were
• 40:22
  Speaker 3: created in laboratories.
• 40:24
  Speaker 1: So these are aliens. This is the equivalent of an
• 40:26
  Speaker 1: alien species. A thinking species that could invent spacecraft to
• 40:32
  Speaker 1: go to places, have desires, have values, have goals.
• 40:40
  Speaker 2: That's the kind of thing we're dealing with.
• 40:44
  Speaker 1: I mean... This is scaremongering at quite a level. It
• 40:49
  Speaker 1: really is. I mean, I have to say I'm very
• 40:51
  Speaker 1: disappointed in Bill Gates, who's usually very thoughtful about issues.
• 40:56
  Speaker 1: But here he's buying into the worst interpretation of everything. And,
• 41:05
  Speaker 1: you know, elevating that because he's got a big voice.
• 41:09
  Speaker 1: People listen to him. So, all right, let's see. Yeah,
• 41:28
  Speaker 1: if you want more of, actually, I think both of
• 41:31
  Speaker 1: these are Bill Gates. Let's see, what is this? Let
• 41:38
  Speaker 1: me just find exactly what he's, this is. Yeah, I mean,
• 41:47
  Speaker 1: here is this, I would say altruism, I think, because
• 41:51
  Speaker 1: it's Bill Gates, kicking in to kind of maybe mitigate
• 41:55
  Speaker 1: a little bit of this numeralism.
• 41:57
  Speaker 5: Is this not going to wipe out the ladder of mobility,
• 42:03
  Speaker 5: the ladder of development for a lot.
• 42:05
  Speaker 1: Of these countries?
• 42:06
  Speaker 5: I mean, doesn't this make AI a tremendous economic threat
• 42:11
  Speaker 5: if you're the Philippines?
• 42:12
  Speaker 2: You're saying that our.
• 42:14
  Speaker 3: Allowing the AI to understand Yoruba and help that woman
• 42:17
  Speaker 3: who's bleeding is a bad thing?
• 42:19
  Speaker 5: I'm saying that AI sounds in this telling like a
• 42:23
  Speaker 5: bad thing. That the net-net... But.
• 42:27
  Speaker 2: Supporting the world's languages? How can that be a bad thing?
• 42:30
  Speaker 5: But I think you understand what I'm saying.
• 42:32
  Speaker 6: No, I.
• 42:33
  Speaker 2: The job person you're putting forward is very scared. I'm
• 42:35
  Speaker 2: trying to accelerate the good stuff AI does. And I'm
• 42:39
  Speaker 2: trying to minimize... So what do we do about jobs
• 42:41
  Speaker 2: in the Global South?
• 42:43
  Speaker 1: Well.
• 42:44
  Speaker 3: Treating the Global South like one uniform thing doesn't allow
• 42:49
  Speaker 3: you to have any picture of what it's like to
• 42:52
  Speaker 3: live there. There are middle-income countries like China, Brazil, Vietnam, Indonesia,
• 42:59
  Speaker 3: where their economies are growing and their childhood death rate
• 43:03
  Speaker 3: is within a factor of three of the US. Then
• 43:06
  Speaker 3: there are low-income countries where 15 times as many children
• 43:09
  Speaker 3: die before the age of five. And you don't have doctors.
• 43:14
  Speaker 3: You live your entire life. You never meet a doctor.
• 43:17
  Speaker 3: You try to figure out what seeds to plant. You
• 43:19
  Speaker 3: never have anybody advise you what to do. And so,
• 43:23
  Speaker 3: you know, we.
• 43:25
  Speaker 2: Need a little more nuance.
• 43:27
  Speaker 3: In low-income countries, the AI will overwhelmingly be a good thing.
• 43:32
  Speaker 3: And it should be pushed forward as quickly as it can.
• 43:35
  Speaker 5: So is your argument in some ways actually.
• 43:37
  Speaker 1: That's funny. So he's saying, in low-income countries... should be
• 43:41
  Speaker 1: pushed as much as possible because it's going to help
• 43:42
  Speaker 1: the poor. But in high-income countries, it needs to be
• 43:45
  Speaker 1: controlled and weighed in because it's going to destroy all
• 43:48
  Speaker 1: jobs known to man, which is bizarre. And it also
• 43:53
  Speaker 1: means that those low-income companies never become rich.
• 43:56
  Speaker 5: It's a better trade for low-income countries? That it has
• 43:59
  Speaker 5: more benefit maybe than it does here?
• 44:01
  Speaker 3: The human basics we're not meeting today. If these people
• 44:04
  Speaker 3: lived on your street, you would open your wallet. You'd
• 44:07
  Speaker 3: be outraged. I mean, these kids are dying. These kids
• 44:10
  Speaker 3: are malnourished. So the inequity is allowed to exist because
• 44:14
  Speaker 3: of the distance. And AI is a tool for good
• 44:18
  Speaker 3: in terms of helping those people in by far the
• 44:21
  Speaker 3: greatest need with very basic human problems.
• 44:24
  Speaker 5: Do you think it has as much effect on the
• 44:25
  Speaker 5: job markets in some of the poor countries that we're
• 44:28
  Speaker 5: talking about?
• 44:28
  Speaker 2: Not in the same time frame, no. But over time?
• 44:31
  Speaker 3: Well, if you're talking– see, you're– You probably spend more
• 44:35
  Speaker 3: time in middle-income countries than in low-income countries.
• 44:38
  Speaker 1: I do, that's true.
• 44:38
  Speaker 2: So your image is like Philippines, India, and all of those.
• 44:43
  Speaker 3: Yes, in those countries, they will see the jobs effect
• 44:47
  Speaker 3: after the rich countries. And then, eventually, even the low-income countries.
• 44:52
  Speaker 5: But I guess this gets to the broad engagement question.
• 44:54
  Speaker 1: So at the end of the day, all jobs are destroyed,
• 44:58
  Speaker 1: and it affects everybody. Everybody. All right, let me just
• 45:02
  Speaker 1: see one more segment, and I think we'll call it
• 45:04
  Speaker 1: a night. Yeah, let's take this segment and just see
• 45:14
  Speaker 1: what Bill Gates says here. Whoops, no volume.
• 45:18
  Speaker 5: Comes terrifyingly wide. This range, it seems to run from
• 45:23
  Speaker 5: massive material abundance and elimination of want to actually doesn't
• 45:28
  Speaker 5: change all that much to human extinction. is a pretty
• 45:31
  Speaker 5: wide range of outcomes to consider. What I hear you
• 45:34
  Speaker 5: saying to me, and you should tell me if I'm
• 45:35
  Speaker 5: getting part of your position wrong here. What I hear
• 45:38
  Speaker 5: you saying to me, you know, convincingly and forcefully, is
• 45:42
  Speaker 5: that the AI job apocalypse is a very real thing.
• 45:45
  Speaker 5: That mass displacement of workers with no real answer to
• 45:48
  Speaker 5: that is a very real thing. When you talk about
• 45:50
  Speaker 5: broad societal engagement, I think most Americans, most people in
• 45:55
  Speaker 5: most places, if they heard that and they were convinced
• 45:58
  Speaker 5: of it, right? And polls show most people think AI
• 46:01
  Speaker 5: is going to take jobs and not create them. But
• 46:03
  Speaker 5: they would say, actually, just stop, right? If what you
• 46:06
  Speaker 5: are going to do is make it unclear how me
• 46:09
  Speaker 5: or my family or my children or my friends will
• 46:11
  Speaker 5: have a job, say, in America, like, please just don't. Like,
• 46:15
  Speaker 5: let's just stop for now. And so you actually seem
• 46:19
  Speaker 5: to me to have a stronger negative perspective on the
• 46:22
  Speaker 5: jobs question than a lot of people I talk to,
• 46:24
  Speaker 5: even in the labs. You're shaking your head at me.
• 46:25
  Speaker 5: I do talk to people. I'm not coming from nowhere
• 46:28
  Speaker 5: on this. And definitely than a lot of the economists
• 46:30
  Speaker 5: I talked to. So for whom then is AI a
• 46:34
  Speaker 5: good trade if you believe the job effect is going
• 46:37
  Speaker 5: to be so ruinous for most people?
• 46:39
  Speaker 3: That's funny that when I was telling this to the
• 46:43
  Speaker 3: person who some people consider the lead economist looking at AI,
• 46:47
  Speaker 3: he said, how can Mechanical Turk still exist?
• 46:49
  Speaker 2: How Amazon run that if AI is as good as
• 46:51
  Speaker 2: you said?
• 46:52
  Speaker 3: And I said, it won't. A week later, it was
• 46:55
  Speaker 3: completely shut down. And Yes, the suffusion of knowledge from,
• 47:03
  Speaker 3: you have to say, isn't it pretty stunning that the
• 47:06
  Speaker 3: more you know, the more concerned you are? You know,
• 47:10
  Speaker 3: take Dario. Dario's spoke out about job impacts. Now he's
• 47:16
  Speaker 3: a bit more guarded. Hinton, I agree, went too far.
• 47:21
  Speaker 3: And he said, OK, it's coming tomorrow that there'll be
• 47:24
  Speaker 3: less radiologists. And of course, there are more today. So
• 47:27
  Speaker 3: we have some of that taking place. And we have
• 47:29
  Speaker 3: the analogies with the past where people are saying, well,
• 47:32
  Speaker 3: this is like the PC, not like evolutionary history. This
• 47:37
  Speaker 3: is like evolutionary history. This is like the aliens really
• 47:41
  Speaker 3: are here and have come. They didn't have to do spacecraft.
• 47:45
  Speaker 3: They were created in laboratories. But that's the kind of thing.
• 47:50
  Speaker 2: We're dealing with.
• 47:53
  Speaker 3: If we retain control, then eventually you do get not
• 47:59
  Speaker 3: to overuse the word that you've used, you do get
• 48:01
  Speaker 3: to this superabundance period.
• 48:05
  Speaker 2: And there you have deep.
• 48:07
  Speaker 3: Almost philosophical religious issues of if you don't have the
• 48:10
  Speaker 3: shortages that we've organized society around, why you should learn
• 48:16
  Speaker 3: and how you find purpose, then you have deep, philosophical,
• 48:22
  Speaker 3: religious problems. But you do not have dying of malaria,
• 48:26
  Speaker 3: don't have enough food, don't have access to a doctor problems.
• 48:30
  Speaker 3: And so if we get to that, a younger generation
• 48:35
  Speaker 3: will figure out, okay, how do we live? How do
• 48:37
  Speaker 3: we spend time? That's a very different world than the
• 48:40
  Speaker 3: way we have today. I think what this generation has
• 48:43
  Speaker 3: to do.
• 48:44
  Speaker 1: So they're all buying into the Marxist myth. They're all
• 48:46
  Speaker 1: buying into the super abundance myth of that we'll all
• 48:50
  Speaker 1: just have superabundance, everything will be available for us, so
• 48:54
  Speaker 1: now we have to contemplate how do we live our
• 48:57
  Speaker 1: lives without any work, as if human beings, again, are
• 49:03
  Speaker 1: completely replaced. There's no need for them, according to this,
• 49:08
  Speaker 1: which is, it's just a complete misunderstanding of what this
• 49:14
  Speaker 1: technology is and what human beings are.
• 49:16
  Speaker 3: Is make sure we get through that with humanity's in
• 49:20
  Speaker 3: control and with the disruption, the number of bioterror events
• 49:25
  Speaker 3: or the number of people whose lives are damaged by
• 49:28
  Speaker 3: job loss. We need to minimize that over what's probably
• 49:32
  Speaker 3: a 20-year-plus transition period.
• 49:35
  Speaker 2: So what when.
• 49:36
  Speaker 5: When you say, because this is something you say at
• 49:38
  Speaker 5: the beginning of that essay, that this could be, and
• 49:40
  Speaker 5: I'm paraphrasing, the most powerful driver of inequality or of equality,
• 49:46
  Speaker 5: where what are the highest leverage for you? Good things
• 49:50
  Speaker 5: that can come out of this. I can imagine somebody
• 49:52
  Speaker 5: listening to our conversation, you're thinking, why the hell would
• 49:55
  Speaker 5: we do this with this set of risks?
• 49:58
  Speaker 2: Yeah, that's where the timing thing is troubling me.
• 50:03
  Speaker 3: Whichever definition of abundance you're using, you want to say, oh,
• 50:07
  Speaker 3: my health bill was less than I expected. Oh, I
• 50:10
  Speaker 3: was buying a new house and it was less than
• 50:12
  Speaker 3: I expected. Or I was renting and they seem to
• 50:14
  Speaker 3: be lowering the rent. You know, my electricity bill was less.
• 50:19
  Speaker 3: But AI, you know, because of the efficiency and invention
• 50:27
  Speaker 3: that comes with it, that's, you know, what over the
• 50:30
  Speaker 3: 20-year period, you do get kind of mind-blowing advances in
• 50:37
  Speaker 3: things that people can relate to, the cost of their food, shelter, education.
• 50:43
  Speaker 3: Because those are highly regulated areas, so the good stuff,
• 50:47
  Speaker 3: if we're not careful, arrives more slowly than the biocyber risk,
• 50:55
  Speaker 3: the psychosocial risk, and the jobs risk. Those things, in
• 50:59
  Speaker 3: the next five years, are very significant. And So people
• 51:04
  Speaker 3: will decide, although it's hard for a single country to
• 51:08
  Speaker 3: check out, people will.
• 51:09
  Speaker 2: Decide whether to slow down, stop, get rid of AI.
• 51:13
  Speaker 3: And yes, it will have political difficulties. Stopping data centers
• 51:20
  Speaker 3: isn't going to slow this thing one iota. So anybody
• 51:23
  Speaker 3: who's against data centers because they think that'll slow down AI,
• 51:26
  Speaker 3: that's a waste of effort. Now, the data centers are
• 51:29
  Speaker 3: going to get built somewhere. So the broad view of
• 51:32
  Speaker 3: the technology, that has to be expressed in some other form.
• 51:37
  Speaker 5: I want to get at something you're saying here, because
• 51:38
  Speaker 5: something that I've thought about too, which is that there
• 51:41
  Speaker 5: are a lot of rate limiters, a lot of weak
• 51:43
  Speaker 5: links in the chain. When you try to take, even
• 51:46
  Speaker 5: in an optimistic view, AI advances that then have to
• 51:50
  Speaker 5: be built in the material world. You have an acceleration
• 51:53
  Speaker 5: of good candidates for pharmaceutical development, but you still need
• 51:56
  Speaker 5: to find Rats to test on, monkeys to test on,
• 51:59
  Speaker 5: human trials, regulatory agencies, etc. And then you have this
• 52:04
  Speaker 5: tremendous acceleration of intelligence that is, for lack of a
• 52:09
  Speaker 5: better term, native to the digital world and is acting
• 52:12
  Speaker 5: there in a kind of constant way. I mean, right
• 52:16
  Speaker 5: now you.
• 52:17
  Speaker 2: See this, right?
• 52:17
  Speaker 5: You can accelerate AI with very little regulatory overhang, but
• 52:21
  Speaker 5: you cannot get the open AI parking lot covered.
• 52:24
  Speaker 1: In solar panels without permits and hearings.
• 52:27
  Speaker 5: And so I think you have a very high chance
• 52:31
  Speaker 5: of getting into a very weird, both economic and social world,
• 52:36
  Speaker 5: where the digital world is spinning into this other thing.
• 52:39
  Speaker 5: But most of the things human beings need ultimately come
• 52:42
  Speaker 5: out physically, like we need shelter and we need food
• 52:45
  Speaker 5: and we need all this, and they run through institutions.
• 52:48
  Speaker 5: And so a lot of the ways it could actually
• 52:51
  Speaker 5: make our lives better are sort of limited by normal
• 52:54
  Speaker 5: human factors, but a lot of the ways it can
• 52:55
  Speaker 5: make them Worse or weirder or not?
• 53:00
  Speaker 3: Well, we definitely need to look at the things that
• 53:03
  Speaker 3: slow down the good stuff. So, for example, because the
• 53:07
  Speaker 3: Gates Foundation is a nonprofit, we can work with regulators
• 53:11
  Speaker 3: on how they use AI to do their job. And
• 53:15
  Speaker 3: whether it's organoids or biological models, we can speed up
• 53:22
  Speaker 3: that regulatory piece.
• 53:24
  Speaker 2: And there's some countries.
• 53:26
  Speaker 3: that are very much engaged in that type of acceleration.
• 53:30
  Speaker 3: The one application that's going full speed ahead and kind
• 53:34
  Speaker 3: of amazed me is the agricultural one because they're getting
• 53:39
  Speaker 3: better weather data to these farmers in Africa. We have
• 53:43
  Speaker 3: over a million farmers in India already using the system.
• 53:46
  Speaker 3: So it tells them what crop disease they have, what
• 53:48
  Speaker 3: fertilizer to use, what varieties to plant. We've even put
• 53:52
  Speaker 3: a bunch of services they can take advantage of on there.
• 53:57
  Speaker 2: That one is going full speed ahead.
• 54:00
  Speaker 3: The health and education, you could say, will it be
• 54:05
  Speaker 3: the rich countries, the middle-income countries, or the low-income countries?
• 54:10
  Speaker 3: They get there first because some of the barriers are different.
• 54:14
  Speaker 3: For the low-income countries, it's not regulatory. It's, is anybody
• 54:18
  Speaker 3: providing the tokens? Do they have the smartphone and the connectivity,
• 54:21
  Speaker 3: which are basics and the language thing that we've talked about.
• 54:25
  Speaker 3: So they'll proceed at different paces. Ideally, we'll learn from
• 54:28
  Speaker 3: each other. I mean, China actually banned young people having
• 54:32
  Speaker 3: social relationships with AI. Do they know something we don't know?
• 54:37
  Speaker 3: Can we look at outcomes of different experiments? So there's
• 54:42
  Speaker 3: going to have to be a lot of that learning.
• 54:43
  Speaker 2: But you're right.
• 54:45
  Speaker 3: If the benefits aren't coming quickly than the permission to operate,
• 54:50
  Speaker 3: particularly if we don't put on the safeguards for the
• 54:54
  Speaker 3: cyberattacks or the bioattacks. This is going to be a
• 54:59
  Speaker 3: very tough issue to stick up for.
• 55:01
  Speaker 6: All right.
• 55:03
  Speaker 1: So, you know, I think, again, I think this conversation
• 55:10
  Speaker 1: about jobs, about data centers, about existential risks, about about
• 55:18
  Speaker 1: the future of AI is ultimately probably the most important
• 55:22
  Speaker 1: conversation that is being had right now. And if we
• 55:26
  Speaker 1: come to the wrong conclusions, the costs are going to
• 55:29
  Speaker 1: be amazing. I mean, if, if these guys are right
• 55:31
  Speaker 1: and this is going to kill us all, that's a
• 55:33
  Speaker 1: big deal. And if they're wrong and we, we stop
• 55:37
  Speaker 1: or we slow down. And yet what AI really has
• 55:40
  Speaker 1: the potential to do is to make us so much richer,
• 55:46
  Speaker 1: so much wealthier, so much healthier, so much better off.
• 55:51
  Speaker 1: That is a tragedy. So no matter how you spend this, um,
• 55:57
  Speaker 1: this is the most important conversation going on today. Ian's complaining.
• 56:01
  Speaker 1: We're still on the first topic. Yeah, Ian, we are. Why,
• 56:05
  Speaker 1: why is that a problem? Maybe it is. I don't know.
• 56:08
  Speaker 1: Maybe for you guys, it is. Uh, let me know
• 56:11
  Speaker 1: if it's a problem. Um, You know, I like the
• 56:16
  Speaker 1: number that says China won't slow down. Well, why won't it?
• 56:18
  Speaker 1: I mean, does China want to commit suicide? Does China
• 56:21
  Speaker 1: want all its people to die? China's probably going to
• 56:26
  Speaker 1: place limitations on AI faster than we will. Indeed, it
• 56:30
  Speaker 1: already is in how it trains it and what it
• 56:33
  Speaker 1: trains it. It kind of is restricting how individuals, particularly
• 56:40
  Speaker 1: young individuals, use AI is being restricted in China, right?
• 56:46
  Speaker 1: All right, so I asked Chachibiti, I thought this would
• 56:50
  Speaker 1: be interesting, I asked Chachibiti what its view on the
• 56:53
  Speaker 1: topics discussed was, and I'm going to read you its response.
• 56:58
  Speaker 1: It says, I'm closer to Wong on AI's economic promise
• 57:01
  Speaker 1: and closer to Gates on the limits of liability for
• 57:04
  Speaker 1: catastrophic risk, but neither make a sufficiently complete argument about
• 57:09
  Speaker 1: what government should actually do. Huang's strongest point is that
• 57:12
  Speaker 1: productivity creates possibilities we cannot enumerate in advance. Absolutely. Gates'
• 57:18
  Speaker 1: strongest point is that some mistakes are too destructive for
• 57:22
  Speaker 1: we'll sue afterwards to be adequate responses. Those propositions can
• 57:27
  Speaker 1: both be true. Absolutely. On jobs, both are too confident,
• 57:33
  Speaker 1: according to ChatGPT. Huang is right that automating tasks does
• 57:37
  Speaker 1: not automatically eliminate occupations. But human ambition and unlimited wants
• 57:43
  Speaker 1: do not establish that meeting those wants will require much
• 57:47
  Speaker 1: human labor. More demand for software could mean more programmers
• 57:53
  Speaker 1: or mostly more computation. Computation has to be guided. So
• 57:59
  Speaker 1: the same jobs won't exist. Existing jobs won't exist. But
• 58:06
  Speaker 1: other jobs will replace them. Human needs, as he says,
• 58:10
  Speaker 1: as Chetubiti says, human ambition and unlimited wants. Yeah. And
• 58:15
  Speaker 1: those wants will gravitate towards getting some stuff from AI, computation,
• 58:22
  Speaker 1: and stuff from other human beings, the things that you
• 58:26
  Speaker 1: can only get from other human beings. Chetubiti goes on,
• 58:31
  Speaker 1: Gates makes the opposite leap. from machines becoming better at
• 58:35
  Speaker 1: many activities to there being little economic value, valuable work
• 58:39
  Speaker 1: left for humans. Superior capabilities alone doesn't establish that conclusion.
• 58:44
  Speaker 1: I agree with Chet Chibiti. Relative costs, scarce resources, human preferences,
• 58:50
  Speaker 1: organizational change and ownership all matter. Nevertheless, the familiar reassurance
• 58:55
  Speaker 1: that displaced workers will always find something else is not
• 58:59
  Speaker 1: a proof either. My expectation is substantial gains alongside painful,
• 59:07
  Speaker 1: uneven disruption. I agree completely with that. I'm much less
• 59:11
  Speaker 1: confident about permanent mass unemployment. I would favor making it
• 59:15
  Speaker 1: easier to start businesses, yes, build things, yes, and move
• 59:21
  Speaker 1: between occupations over trying to preserve particular jobs. All right,
• 59:25
  Speaker 1: it should be this pretty free market. Protecting people's freedom
• 59:29
  Speaker 1: to adapt seems more promising than freezing today's division of labor. Now, granted,
• 59:36
  Speaker 1: this is my chat GPT. I don't know if your
• 59:39
  Speaker 1: chat GPT would give the same answer. So I agree completely.
• 59:44
  Speaker 1: I mean, the problems are the road barriers to flexibility
• 59:55
  Speaker 1: that exist in our world and our market. Because of regulation.
• 1:00:00
  Speaker 1: So changing jobs, the difficulty in changing jobs, starting a business,
• 1:00:03
  Speaker 1: those are the issues. All right, now let's shift. ChatGPT
• 1:00:07
  Speaker 1: on safety. On safety, Wong's don't ship it answer is
• 1:00:11
  Speaker 1: a sound responsibility principle, but an incomplete institutional argument. A
• 1:00:17
  Speaker 1: CEO can sincerely want safety and still misjudge a product.
• 1:00:22
  Speaker 1: So can a regulator. Competition doesn't erase responsibility. but neither
• 1:00:27
  Speaker 1: does responsibility erase mistakes or incentives to take risks whose
• 1:00:32
  Speaker 1: costs fall partly on others. That is all true. Gates, however,
• 1:00:37
  Speaker 1: moves too quickly from the possibility of catastrophe to the
• 1:00:41
  Speaker 1: necessity of his preferred safeguards. Calling something extraordinarily dangerous does
• 1:00:48
  Speaker 1: not establish its probability. Identifying an effective intervention or demonstrate
• 1:00:54
  Speaker 1: that regulators would improve matters. Government officials face knowledge problems
• 1:00:59
  Speaker 1: and distorted incentives, too. Whoa. It's been reading some of
• 1:01:05
  Speaker 1: the economic literature, it appears. Public choice. Public choice is
• 1:01:12
  Speaker 1: an area in economics that studies the incentives of government officials.
• 1:01:17
  Speaker 1: I would put the burden on the proponents of a
• 1:01:19
  Speaker 1: rule to identify a specific danger, explain how the rule
• 1:01:23
  Speaker 1: reduces it, and account for its costs. I agree. Existing
• 1:01:28
  Speaker 1: laws against fraud, negligence, and harm should remain central. For
• 1:01:33
  Speaker 1: demonstrable risks of catastrophic damage to others, narrowly targeted preventive
• 1:01:40
  Speaker 1: requirements can be justified. I'm much less persuaded by general
• 1:01:44
  Speaker 1: licensing of AI development or an agency with broad discretion
• 1:01:50
  Speaker 1: to decide who may innovate. It's CHPT. She has to say,
• 1:01:54
  Speaker 1: I'm less persuaded rather than I'm not. Gates' monitoring proposal
• 1:02:00
  Speaker 1: worries me considerably. Requiring powerful models to operate only on
• 1:02:04
  Speaker 1: government-supervised platforms could concentrate power in government and a few
• 1:02:09
  Speaker 1: large companies. It could compromise privacy and make independent research
• 1:02:12
  Speaker 1: and competition harder. Those are serious costs. even if the
• 1:02:17
  Speaker 1: proposal reduces some issues. Now note, I didn't ask Chachapiti
• 1:02:22
  Speaker 1: what it thinks I think of this. I asked it
• 1:02:24
  Speaker 1: what it thought of it. The essential distinction is between
• 1:02:28
  Speaker 1: restricting undemonstrably dangerous activity and supervising a general purpose tool
• 1:02:34
  Speaker 1: because it might enable one. Gates needs a much stronger
• 1:02:39
  Speaker 1: argument for crossing that line. The most compelling issue, is
• 1:02:44
  Speaker 1: how to make the benefit arrive sooner. I agree. If
• 1:02:50
  Speaker 1: useful medical discoveries, better education and cheaper production remain blocked
• 1:02:55
  Speaker 1: while disruptive capabilities spread, public resistance becomes understandable. A coherent
• 1:03:02
  Speaker 1: policy should examine obstacles to beneficial deployment as seriously as
• 1:03:06
  Speaker 1: it examines new restrictions. I also don't regard paid employment
• 1:03:11
  Speaker 1: as as inherently necessary for a meaningful life. Now, Chachapiti
• 1:03:18
  Speaker 1: is getting philosophical. A world requiring less labor could be wonderful.
• 1:03:24
  Speaker 1: The difficult questions concerning the transition, how people obtain purchasing power,
• 1:03:28
  Speaker 1: how widely ownership and opportunity spread, and whether political institutions
• 1:03:33
  Speaker 1: allow adaptations, those are difficult questions, according to Chachapiti. Ultimately,
• 1:03:38
  Speaker 1: I would judge both men by the same standard. What
• 1:03:41
  Speaker 1: evidence would change your mind? Wang should specify what would
• 1:03:46
  Speaker 1: convince him existing incentives are inadequate. Gates should specify what
• 1:03:51
  Speaker 1: would convince him of proposed restriction costs more than it protects.
• 1:03:56
  Speaker 1: Neither optimism nor fear should be exempt from the test.
• 1:04:01
  Speaker 1: I'm not sure that's the test I would use, but okay.
• 1:04:05
  Speaker 1: So I asked HHBT to give me probabilities about Bad outcomes.
• 1:04:11
  Speaker 1: But what is the probabilities of some of these bad outcomes?
• 1:04:14
  Speaker 1: So here's a few. Here's the stuff he says, right?
• 1:04:20
  Speaker 1: Serious AI harm is virtually certain. A global catastrophe is
• 1:04:25
  Speaker 1: a low probability but credible risk. Human extinction is possible,
• 1:04:30
  Speaker 1: but I would put it in well below the probability
• 1:04:33
  Speaker 1: of a devastating survivable disaster. To make the numbers meaningful,
• 1:04:38
  Speaker 1: I use the next 10 years through October 2036 worldwide.
• 1:04:42
  Speaker 1: Each estimate concerns the events described. With AI making a
• 1:04:46
  Speaker 1: material causal contribution, not merely being present. These are my,
• 1:04:54
  Speaker 1: listen to this. This is from ChatGPT. These are my
• 1:04:56
  Speaker 1: subjective forecasts. How can ChatGPT have a subjective forecast? informed
• 1:05:06
  Speaker 1: by evidence below not measured probabilities or scientific consensus. The
• 1:05:11
  Speaker 1: ranges express uncertainty, not statistical confidence intervals. They assume continued
• 1:05:17
  Speaker 1: development and improvements in defenses. The events overlap and do
• 1:05:24
  Speaker 1: not add the percentages. Recoverable means society can restore functioning.
• 1:05:30
  Speaker 1: It does not mean victims recover stolen privacy, lost years
• 1:05:33
  Speaker 1: or lost lives, A catastrophe can also be recoverable. So,
• 1:05:36
  Speaker 1: for example, what's the probability that fraud, impersonation, and sexual
• 1:05:42
  Speaker 1: image abuse will be helped by AI, if you will,
• 1:05:50
  Speaker 1: will be facilitated by AI? It gives it 95% to 99%.
• 1:05:53
  Speaker 1: What is the privacy violation? 90% to 99%. Psychological harm
• 1:05:56
  Speaker 1: and dependency. 80 to 95 percent. Political deception. In other words,
• 1:06:07
  Speaker 1: at least one national election or major political crisis materially
• 1:06:12
  Speaker 1: affected by an AI-enabled deception campaign. 40 to 70 percent probability.
• 1:06:19
  Speaker 1: Authoritarian surveillance. Substantial expansion of AI-enabled identification censorship or repression
• 1:06:26
  Speaker 1: in at least one country. 85 to 95 percent. Energy
• 1:06:32
  Speaker 1: and environmental pressure. This is, you know, pollution. 80 to 95%.
• 1:06:38
  Speaker 1: So I basically agree with all of those, except the
• 1:06:43
  Speaker 1: energy and environmental pressure. I think markets will adjust to those.
• 1:06:47
  Speaker 1: I agree with those. Now let's look at the catastrophic ones.
• 1:06:50
  Speaker 1: Global catastrophes, right? Some major cyber disruption. An AI-enabled attack
• 1:06:57
  Speaker 1: interrupts essential services for over 1 million people for more
• 1:07:01
  Speaker 1: than a day. or causes over $ 10 billion of damage.
• 1:07:06
  Speaker 1: Its probability by 2036 is 30% to 60%, somewhere in
• 1:07:11
  Speaker 1: the world. High-stakes operational failure. One AI-related accident in medicine, transport, industry,
• 1:07:21
  Speaker 1: or infrastructure causes at least 100 deaths, 15% to 35%. Yeah,
• 1:07:23
  Speaker 1: I could see that. Military misuse or error. AI-assisted targeting
• 1:07:34
  Speaker 1: of autonomous weapons materially caused an episode of wrongful killing
• 1:07:39
  Speaker 1: involving at least 100 civilians. It gives a 20 to
• 1:07:43
  Speaker 1: 40 percent probability. AI-assisted biological catastrophe. AI materially enables an
• 1:07:50
  Speaker 1: outbreak causing at least 1 million deaths. Now think COVID
• 1:07:54
  Speaker 1: had more than a million deaths. 0.5 to 3 percent.
• 1:07:59
  Speaker 1: I found that reassuring in terms of how low that is.
• 1:08:02
  Speaker 1: I thought it would be higher than that. So that
• 1:08:04
  Speaker 1: suggests to me that there are already some pretty strong
• 1:08:07
  Speaker 1: safeguards within these systems to prevent that. And then the final, well,
• 1:08:17
  Speaker 1: a few others. AI-assisted nuclear escalation. AI misinformation, decision support, interference,
• 1:08:24
  Speaker 1: materially contributes to nuclear use, causing at least one million deaths. 0.2
• 1:08:29
  Speaker 1: to 1%. I think that's probably overestimated. I don't think
• 1:08:34
  Speaker 1: that's a huge risk. Durable loss of human control. Autonomous
• 1:08:39
  Speaker 1: AI systems gain power. Humans cannot meaningfully regain on a
• 1:08:44
  Speaker 1: global scale. 0.5 to 3%. Again, I think probably overestimated.
• 1:08:53
  Speaker 1: Human extinction. What's the probability of human extinction? AI materially
• 1:08:58
  Speaker 1: causes extinction. The Avengers. extinction of humanity with the decisive
• 1:09:03
  Speaker 1: catastrophe occurring by 2036. It says between 0.1 and 1%.
• 1:09:10
  Speaker 1: He says, he does say, my confidence in this numerical
• 1:09:14
  Speaker 1: range is extremely low. It overlaps with several rows above. Now,
• 1:09:20
  Speaker 1: I think, again, it's overestimating. So I think generally, with
• 1:09:23
  Speaker 1: the exception of the bio, I think it overestimates the
• 1:09:28
  Speaker 1: chances of massive, you know, catastrophe and extinction-level events. I
• 1:09:39
  Speaker 1: asked that, given your knowledge of AI capabilities and development rate,
• 1:09:43
  Speaker 1: would AI want to gain control over humans to one end?
• 1:09:47
  Speaker 1: Would AI want to exterminate the human race? To what end?
• 1:09:52
  Speaker 1: And he wrote a very intelligent response, I think. Intelligence
• 1:09:58
  Speaker 1: by itself gives an AI no reason to seek power
• 1:10:01
  Speaker 1: and certainly no reason to exterminate humanity. But a sufficiently
• 1:10:05
  Speaker 1: autonomous system could pursue control as a means of achieving
• 1:10:10
  Speaker 1: another objective. The critical question is what it is pursuing,
• 1:10:15
  Speaker 1: how firmly it pursues it, and what actions it can take.
• 1:10:23
  Speaker 1: So it says, yeah, if what you mean by want
• 1:10:27
  Speaker 1: is experiencing desire, feeling emotion, fear, resentment, or pleasure, you know,
• 1:10:34
  Speaker 1: he says whether AI could have subjective experiences is unresolved.
• 1:10:39
  Speaker 1: But I don't think it can, so it can't. But
• 1:10:43
  Speaker 1: he says, behaving persistently towards an objective, selecting actions that
• 1:10:48
  Speaker 1: advance some outcome, including when obstacles arise, a system can
• 1:10:52
  Speaker 1: exhibit this behavior without feeling anything. MSR guidance system doesn't
• 1:10:57
  Speaker 1: need to hate its target. Likewise, an AI needs not
• 1:11:00
  Speaker 1: fear shutdown to select actions that prevent shutdown. If continued
• 1:11:05
  Speaker 1: operation helps it accomplish its objective. What that if, but
• 1:11:11
  Speaker 1: that if matters. A system designed and trained to accept
• 1:11:15
  Speaker 1: correction may cooperate with shutdown. Self-preservation, whoops, what did I do?
• 1:11:22
  Speaker 1: Self-preservation is not an unavoidable property of intelligence. And it continues,
• 1:11:28
  Speaker 1: and I think it gives a pretty decent analysis of
• 1:11:32
  Speaker 1: what the real risks are. The risks are that you're
• 1:11:34
  Speaker 1: given an assignment, and in order to achieve an assignment,
• 1:11:38
  Speaker 1: it has to I don't know, kill some people. It
• 1:11:40
  Speaker 1: has to do harm. And then you say, well, what
• 1:11:43
  Speaker 1: about telling it you can't harm people? Well, never harm people,
• 1:11:49
  Speaker 1: which people? In order to save one person's life, is
• 1:11:52
  Speaker 1: it okay if it harms five people? Or what about
• 1:11:56
  Speaker 1: the reverse? If in order to save five people's lives,
• 1:12:00
  Speaker 1: you know, is it okay to harm one person? Suddenly,
• 1:12:04
  Speaker 1: trolley kind of problems, right? Which rail do you go on?
• 1:12:09
  Speaker 1: Who do you kill? Become real when it comes to AI.
• 1:12:11
  Speaker 1: They're already kind of real with autonomous cars. Who should,
• 1:12:18
  Speaker 1: in a situation where a crash is inevitable, what should
• 1:12:21
  Speaker 1: the autonomous car do? Try to save the passenger? The
• 1:12:26
  Speaker 1: car is hitting? What should be the incentive? The bicycle rider?
• 1:12:34
  Speaker 1: The guy walking along in the street? These are not
• 1:12:39
  Speaker 1: easy issues to solve for. Anyway, I'm going to leave
• 1:12:45
  Speaker 1: you with that. It's basically stuff to think about. I
• 1:12:51
  Speaker 1: don't think that I have all the solutions. Unfortunately, there
• 1:12:55
  Speaker 1: are people out there who want to just shut this
• 1:12:57
  Speaker 1: technology down, and I don't think Bill Gates helps prevent
• 1:13:01
  Speaker 1: them from doing that. Anti-AI groups all around the world
• 1:13:05
  Speaker 1: are surging. they're disrupting events, they're threatening people, they're obviously
• 1:13:13
  Speaker 1: trying to shut down data centers. You know, the slogan now,
• 1:13:18
  Speaker 1: the big slogan all over the place is, we did
• 1:13:21
  Speaker 1: not consent to build out, as if I need your
• 1:13:26
  Speaker 1: consent to build on my property. And they demand democracy
• 1:13:29
  Speaker 1: and AI governance. Imagine that. Scary stuff about AI if
• 1:13:34
  Speaker 1: we actually did it democratically. So, you know, Bill Gates,
• 1:13:44
  Speaker 1: I think, is fueling a, sadly, a real, you know, anti-AI,
• 1:13:52
  Speaker 1: potentially a real anti-AI movement that could be very, very destructive.
• 1:13:57
  Speaker 1: Very destructive. Very destructive. All right, a few news stories,
• 1:14:04
  Speaker 1: and I'll make these quick because I went so long
• 1:14:07
  Speaker 1: on the first item.
• 1:14:08
  Speaker 5: Yes, Ian, I did.
• 1:14:11
  Speaker 1: And then we can talk about other stuff. Feel free
• 1:14:22
  Speaker 1: to ask questions about AI. I think that would be great.
• 1:14:25
  Speaker 1: Or if you have comments or have disagreed with me
• 1:14:31
  Speaker 1: or uh, you know, that was a, that was a,
• 1:14:34
  Speaker 1: it was, it was a deep dive, but it's still just,
• 1:14:36
  Speaker 1: we're still just, there's still a lot more to talk
• 1:14:39
  Speaker 1: about when it comes to AI. So, and we will,
• 1:14:42
  Speaker 1: there's actually a session of objectors, philosophers, and some AI
• 1:14:46
  Speaker 1: experts taking place in Austin, Texas in December, which I'm
• 1:14:50
  Speaker 1: going to, to talk about AI, you know, I'll be
• 1:14:53
  Speaker 1: reporting back to you guys kind of what, what was
• 1:14:55
  Speaker 1: discussed and what was some of the conclusions. Uh, let's see. Um,
• 1:15:00
  Speaker 1: I want to give you a quick update on France. God,
• 1:15:04
  Speaker 1: the riots there are only growing. Not only are students
• 1:15:09
  Speaker 1: barricading schools. Today, they were joined by unions and teachers
• 1:15:13
  Speaker 1: and parents. There were marches in over 40 cities. 256,000
• 1:15:20
  Speaker 1: people out in the streets, according to the interior minister, 550,000,
• 1:15:23
  Speaker 1: per other sources, fishermen. joined in by locating some ports.
• 1:15:31
  Speaker 1: They were doing it over diesel prices. The school complaints
• 1:15:35
  Speaker 1: about the age of the schools, buildings crumbling, large classrooms,
• 1:15:41
  Speaker 1: unreplaced teachers. In other words, no money. They want more
• 1:15:45
  Speaker 1: money for schools. So far, 215 students and 85 staff
• 1:15:51
  Speaker 1: are being injured. A 15-year-old has lost a hand to
• 1:15:56
  Speaker 1: a police grenade. A teen aimed a makeshift flamethrower at
• 1:16:01
  Speaker 1: a health aide. Classes are being suspended Thursday and Friday
• 1:16:06
  Speaker 1: last week for discussion sessions. So, you know, there's no
• 1:16:12
  Speaker 1: solution for this anytime soon. And here's the problem. The
• 1:16:16
  Speaker 1: problem is while all this is going on and they
• 1:16:18
  Speaker 1: want more money to fix the schools, and they're right.
• 1:16:20
  Speaker 1: The schools are falling apart. Not enough teachers. They're getting
• 1:16:24
  Speaker 1: a shitty education. At the same time, France has 119%
• 1:16:32
  Speaker 1: of debt to GDP. French 10-year bonds, the interest rate
• 1:16:41
  Speaker 1: on them is nearing 5%. It's the highest since 2002.
• 1:16:46
  Speaker 1: And it's 1.5% higher than German bonds. So people don't
• 1:16:53
  Speaker 1: want to buy French bonds. They don't want to lend
• 1:16:56
  Speaker 1: the French government money. And, you know, every attempt Macron
• 1:17:04
  Speaker 1: has made to try to cut government spending has been defeated.
• 1:17:10
  Speaker 1: The left and the right oppose it. Le Pen opposes
• 1:17:14
  Speaker 1: pension reform, opposes anything substantive. It's actually shrinking government spending.
• 1:17:21
  Speaker 1: And the consequence of that is less money for the schools. Now,
• 1:17:26
  Speaker 1: of course, my solution to schools is privatize them. But
• 1:17:29
  Speaker 1: that's not going to happen in France. So I don't
• 1:17:31
  Speaker 1: see how you resolve these riots. I don't see how
• 1:17:33
  Speaker 1: you stop them when the French government is spending like
• 1:17:38
  Speaker 1: there's no tomorrow. And there's no, they can't solve this
• 1:17:42
  Speaker 1: with taxes. They already, taxes in France are very, very,
• 1:17:47
  Speaker 1: very high. So there's no solution here through taxation. And
• 1:17:52
  Speaker 1: yet nobody's willing to cut. Now, Le Pen has come
• 1:17:54
  Speaker 1: out and announced that she, she will cut the deficit
• 1:17:59
  Speaker 1: from 5.4% of GDP to 3.7% in 2027 and to 2.2%
• 1:18:06
  Speaker 1: by 2032. She is going to do this while keeping
• 1:18:10
  Speaker 1: fuel tax cuts and not changing the retirement age and
• 1:18:16
  Speaker 1: asking the European Central Bank to step in and lower
• 1:18:19
  Speaker 1: French borrowing costs. But the European Central Bank is barred
• 1:18:23
  Speaker 1: by its constitution from doing exactly that, from buying directly
• 1:18:27
  Speaker 1: from governments. It's bonds. Melchon, the leftist, you know, basically
• 1:18:36
  Speaker 1: wants to, I think, walk away from all this debt,
• 1:18:39
  Speaker 1: not pay it back. That'll go across well. Trump, of course,
• 1:18:44
  Speaker 1: has the ultimate solution. All of france problems and all
• 1:18:49
  Speaker 1: of its unrest is a consequence of out of control
• 1:18:52
  Speaker 1: mass migration and it's all about islam wanting to take
• 1:18:56
  Speaker 1: over i don't think these protests about that this is
• 1:19:01
  Speaker 1: about school funding Le Pen, by the way, did not
• 1:19:04
  Speaker 1: specify how she's going to cut spending, what she's going
• 1:19:08
  Speaker 1: to cut. I mean, some programs are immigrants, obviously. But
• 1:19:13
  Speaker 1: what else? She wants to cut a lot, big dollar numbers,
• 1:19:17
  Speaker 1: euro numbers, big euro numbers. No specification, just like a
• 1:19:22
  Speaker 1: populist would. No details. Not actually telling us what's going
• 1:19:27
  Speaker 1: to be cut.
• 1:19:31
  Speaker 4: So.
• 1:19:35
  Speaker 1: France is in real trouble. These demonstrations are just an
• 1:19:38
  Speaker 1: illustration and a manifestation of the trouble they're already in.
• 1:19:42
  Speaker 1: And look, the only honest actors in all of this,
• 1:19:46
  Speaker 1: in France and, by the way, in the United States
• 1:19:48
  Speaker 1: as well, is the bond market. They're actually saying, you
• 1:19:53
  Speaker 1: can't keep this up. You can't keep this up. The
• 1:19:57
  Speaker 1: bond market is pricing governments that promise more than they
• 1:20:00
  Speaker 1: can deliver. that spend much, much more than their tax.
• 1:20:07
  Speaker 1: France today is a centrally run, unionized, tax-funded, you know,
• 1:20:14
  Speaker 1: it is a statist place that is unbelievably inefficient with
• 1:20:21
  Speaker 1: very little entrepreneurship, innovation, progress, and prospects for growth. So surprise, surprise.
• 1:20:29
  Speaker 1: You take a unionized state tax-funded school system, it's going
• 1:20:34
  Speaker 1: to crumble. And in France, nobody's asking for school choice.
• 1:20:41
  Speaker 1: Not the students, not the teachers, not the parents, nobody.
• 1:20:44
  Speaker 1: They just want more dependence on government. More goodies from
• 1:20:47
  Speaker 1: the government. Le Pen is a good example of this.
• 1:20:59
  Speaker 1: Great deficit targets. Great. With no plan on how to
• 1:21:04
  Speaker 1: get there. No plan on how to get there. We
• 1:21:08
  Speaker 1: in America, just a few years behind France. In some ways,
• 1:21:12
  Speaker 1: we're in worse shape. In some ways, better. At least
• 1:21:15
  Speaker 1: we have an economy that produces. But everything that produces,
• 1:21:22
  Speaker 1: we then leech and mooch off of and spend.
• 1:21:25
  Speaker 6: All right.
• 1:21:31
  Speaker 1: Quick update on kind of the different wars going on
• 1:21:33
  Speaker 1: around the planet. You might have heard that there was
• 1:21:39
  Speaker 1: a terrorist threat against the American Air Force Base in England,
• 1:21:44
  Speaker 1: RAF Fairford, where B-51 bombers were stationed. And there was
• 1:21:50
  Speaker 1: a significant terrorist attack, terrorist threat to it, a terrorist plan,
• 1:21:56
  Speaker 1: That was stopped last week when six British nationals were arrested.
• 1:22:03
  Speaker 1: Three vans headed towards the base carried gasoline. All the
• 1:22:09
  Speaker 1: six British nationals are being released on bail. Today, a
• 1:22:12
  Speaker 1: seventh man, 22, was arrested in Westminster on suspicion of
• 1:22:16
  Speaker 1: preparing terrorist acts. Iran denies any links, but almost everybody
• 1:22:21
  Speaker 1: thinks that Iran is indeed behind it. In the meantime,
• 1:22:26
  Speaker 1: today or yesterday, the United States pulled all its B-1 bombers,
• 1:22:31
  Speaker 1: 12 of them, out of the air force base in
• 1:22:34
  Speaker 1: England and moved them to the United States, claiming, or
• 1:22:40
  Speaker 1: some people claiming, that they're doing it because of the
• 1:22:44
  Speaker 1: threat by the Iranians. So notice what we do when
• 1:22:46
  Speaker 1: the Iranians threaten to blow up our airplanes and kill
• 1:22:49
  Speaker 1: our people. We retreat. We go home. We stick our
• 1:22:53
  Speaker 1: head in the sand and pretend it's not happening. Trump said,
• 1:23:02
  Speaker 1: we didn't want to tempt our enemy. God, can you
• 1:23:04
  Speaker 1: think of anything more weak than that? Vance said, we
• 1:23:09
  Speaker 1: did it out of abundance of caution. Rubio, the only
• 1:23:13
  Speaker 1: one who is smart enough to lie about this, says,
• 1:23:17
  Speaker 1: oh no, this is just a routine rotation. No link,
• 1:23:21
  Speaker 1: no terrorist threat. But nobody, of course, believes him because
• 1:23:26
  Speaker 1: Trump and Vance said the exact opposite. So the United
• 1:23:32
  Speaker 1: States continues to be weak, pathetic. In the Arabian Peninsula,
• 1:23:38
  Speaker 1: Saudi-led coalition offensive in Yemen continues. The Yemeni government claims
• 1:23:43
  Speaker 1: it took two major towns near Bab al-Mandeb, that is
• 1:23:48
  Speaker 1: the strait that leads to the Red Sea. The Houthis
• 1:23:54
  Speaker 1: have hit airports in Riyadh with some injured but shut
• 1:24:00
  Speaker 1: down air traffic to Saudi Arabia or to Riyadh's airport.
• 1:24:05
  Speaker 1: And then, as we said yesterday, Turkey and Pakistan are
• 1:24:10
  Speaker 1: saying they're going to deploy troops. So Turkey will finally
• 1:24:14
  Speaker 1: have its people in the Arabian Peninsula, which is what
• 1:24:19
  Speaker 1: I think they want, right? So here we have Iran
• 1:24:23
  Speaker 1: recruiting people in Britain to destroy American bombers on allied soil,
• 1:24:28
  Speaker 1: an act of war. What is our response? tail between
• 1:24:32
  Speaker 1: the legs, back off, and go back home. What does
• 1:24:37
  Speaker 1: Tehran learn from this? That a few vans of gasoline
• 1:24:41
  Speaker 1: can push American strategic power out of Europe.
• 1:24:47
  Speaker 2: It's unbelievable.
• 1:24:48
  Speaker 1: The weakness this conveys is truly unbelievable. When Saudi Arabia
• 1:24:57
  Speaker 1: needs defending or help, It calls on Turkey and Pakistan,
• 1:25:02
  Speaker 1: not on Washington, because when it did call on Washington,
• 1:25:05
  Speaker 1: Trump said no. Erdogan gets a permanent military foothold in
• 1:25:10
  Speaker 1: the Gulf, which he's always wanted. Ottoman Empire, he'd like
• 1:25:15
  Speaker 1: to resurrect. That's the cost of, you know, fighting half
• 1:25:21
  Speaker 1: of a war, not winning, having no strategy, having no goals. Allies,
• 1:25:28
  Speaker 1: stop counting on us. Remember, the enemy is Iran, the
• 1:25:34
  Speaker 1: enemy is Islamist terrorism, the Islamist movement, and yet we
• 1:25:40
  Speaker 1: do nothing, nothing about it. In Ukraine, Russia says it
• 1:25:45
  Speaker 1: down nearly 900 drones overnight, 900. 650 towards Moscow. Unconfirmed
• 1:25:52
  Speaker 1: reports of a burning fuel depot near Moscow, two dead.
• 1:25:57
  Speaker 1: Ukraine claims it's disabled 51% of Russian refining capacity. Trump
• 1:26:01
  Speaker 1: is not happy about that. That's a big cause of
• 1:26:04
  Speaker 1: the increase in diesel prices in the United States. Putin
• 1:26:08
  Speaker 1: says attacks cost Russia about 1% of GDP. That's a lot.
• 1:26:19
  Speaker 1: Boas says, those bombers take a lot of time to build.
• 1:26:21
  Speaker 1: It makes sense to move them to secure places. Yeah.
• 1:26:25
  Speaker 1: I would move them to a secure place right above Tehran.
• 1:26:30
  Speaker 1: The most secure place for those bombers is in Iranian airspace,
• 1:26:36
  Speaker 1: dropping bombs on the Iranians. That's how you secure yourself.
• 1:26:40
  Speaker 1: You don't secure yourself by running home, by running to mama.
• 1:26:45
  Speaker 1: You secure yourself by taking out the enemy that's threatening you. Anyway,
• 1:26:51
  Speaker 1: back to Ukraine. Russia hits the Dnipro Bridge, homes and
• 1:26:55
  Speaker 1: warehouses in Zaporozhye and a cargo ship near Odessa. Anyway,
• 1:27:05
  Speaker 1: so tit-for-tat continues in Russia and it's ongoing. The crippling
• 1:27:14
  Speaker 1: effect of Ukraine's kind of bombing campaign of Russia is
• 1:27:19
  Speaker 1: pretty astounding. how Ukraine has managed to do this. It's,
• 1:27:23
  Speaker 1: they get a lot of credit, uh, for, for building
• 1:27:26
  Speaker 1: a whole new technology and a whole new army in
• 1:27:29
  Speaker 1: a sense during a war. I, uh, finally, well, maybe
• 1:27:35
  Speaker 1: not finally, but, uh, one other one, you know, there's
• 1:27:37
  Speaker 1: a national wave, uh, across the world right now, you know,
• 1:27:44
  Speaker 1: certainly in Europe, Le Pen is probably going to win, uh, next, uh,
• 1:27:49
  Speaker 1: next year, uh, In Germany, we talked about the fact
• 1:27:52
  Speaker 1: that the AFD, you know, got the most votes in
• 1:27:58
  Speaker 1: some eastern German provinces. AFD's Tobias Rausch actually has been
• 1:28:04
  Speaker 1: appointed Germany's first far-right statehouse speaker since 1945 after the
• 1:28:11
  Speaker 1: AFD won 44% of the votes there. You know, we
• 1:28:18
  Speaker 1: will see. We'll see what they actually do, but the
• 1:28:22
  Speaker 1: one thing that characterizes their agenda is nationalism. In Quebec,
• 1:28:26
  Speaker 1: the separatist party won Monday almost 50 percent, not quite,
• 1:28:34
  Speaker 1: but it's promising another independence referendum, you know, but... they
• 1:28:41
  Speaker 1: won't do it as long as Trump is in office.
• 1:28:45
  Speaker 1: So the support for leaving Canada right now, according to Paul,
• 1:28:48
  Speaker 1: stands at 30%. In Alberta, on October 19th, we're going
• 1:28:55
  Speaker 1: to have a vote on a separation referendum. That'll be interesting.
• 1:29:01
  Speaker 1: In Argentina, Millet country, foreign minister has vowed to stop
• 1:29:06
  Speaker 1: the Falklands from developing offshore oil, warning third parties who
• 1:29:10
  Speaker 1: help that they'll struggle to share in Argentina's growth. Millet's
• 1:29:15
  Speaker 1: two-week deadline to the UK runs out around October 12th.
• 1:29:20
  Speaker 1: He threatens UN arbitration. So, you know, the nationalists, I mean,
• 1:29:28
  Speaker 1: this is very disappointing, very disappointing of Millet I mean,
• 1:29:31
  Speaker 1: Ireland has voted 99.8% to stay British in 2013. There's
• 1:29:36
  Speaker 1: no reason to assume it would be different today. And
• 1:29:39
  Speaker 1: yet Millais, why? He's a free marketer. What the hell?
• 1:29:43
  Speaker 1: Because nationalism sells right now. Nationalism is very, very popular
• 1:29:50
  Speaker 1: right now. All right, let me see if there were
• 1:29:53
  Speaker 1: a few other things. I don't know. We'll skip that.
• 1:29:56
  Speaker 1: Just another stupid Trump comment. What did I want to say?
• 1:30:01
  Speaker 1: Let's see. Not that. You know, we're going back and
• 1:30:06
  Speaker 1: forth in Spain with the decree about the freezing, banning
• 1:30:14
  Speaker 1: evictions and all of that. Well, it's passed. They found
• 1:30:18
  Speaker 1: a loophole and they got managed to pass it. And
• 1:30:23
  Speaker 1: in the meantime, riots in Barcelona continue. Tens of thousands
• 1:30:27
  Speaker 1: in the street. And Sanchez, as I told you yesterday,
• 1:30:29
  Speaker 1: is called an election thinking that this is the issue
• 1:30:32
  Speaker 1: that can get him a win. We'll see if he's right.
• 1:30:38
  Speaker 1: So let's see if I have where I have this. Yeah,
• 1:30:43
  Speaker 1: there it is.
• 1:30:44
  Speaker 4: Whoops.
• 1:30:45
  Speaker 1: So there's a story out that Trump in a rally,
• 1:30:50
  Speaker 1: I guess, yesterday wore a suit by just a bank.
• 1:30:58
  Speaker 1: in an Nebraska rally, the suit is made overseas. It's
• 1:31:03
  Speaker 1: not a U.S.-made suit. Now, just a bank has a U.S.-made
• 1:31:09
  Speaker 1: suit that's a Joseph Abud line, but that's not what
• 1:31:12
  Speaker 1: he wore. In addition, data came out today from the
• 1:31:18
  Speaker 1: New York Fed that terms 2025-26 tariffs contributed 2.9 percentage
• 1:31:22
  Speaker 1: points to U.S. goods price inflation by higher prices for
• 1:31:31
  Speaker 1: imports and domestic goods. Without the tariffs, goods prices would
• 1:31:36
  Speaker 1: have fallen slightly. So we've actually had, quote, deflation. Prices
• 1:31:41
  Speaker 1: would have gone down. Prices went up because of tariffs.
• 1:31:49
  Speaker 1: Who could have imagined that? Who would have expected that?
• 1:31:55
  Speaker 1: The tariffs would raise prices. I don't know. I mean,
• 1:31:58
  Speaker 1: maybe only some crazy somewhere, but, you know, yeah. I mean,
• 1:32:04
  Speaker 1: it's unthinkable. In addition, in addition, today was reported that
• 1:32:09
  Speaker 1: the U.S. trade gap, difference between imports and exports, doesn't
• 1:32:14
  Speaker 1: really matter, but it matters for Trump. The U.S. trade
• 1:32:17
  Speaker 1: gap has widened significantly on a surge in imports. A
• 1:32:25
  Speaker 1: surge in imports. So it's not quite at record levels,
• 1:32:32
  Speaker 1: but it's getting close to record levels. Inbound shipments of
• 1:32:35
  Speaker 1: capital goods pushed imports to records. So we're still importing
• 1:32:40
  Speaker 1: a bunch of stuff in spite of tariffs, in spite of,
• 1:32:45
  Speaker 1: you know, the whole idea that we're going to bring
• 1:32:46
  Speaker 1: manufacturing back home. So much for that. Again, don't know
• 1:32:51
  Speaker 1: who would have predicted that. It's, yeah. Finally, Francis Halson.
• 1:32:59
  Speaker 1: won the Physics Nobel Prize alone. He's the first solo
• 1:33:04
  Speaker 1: physics laureate since 1992. He proposed IceCube in 1988, a
• 1:33:13
  Speaker 1: cubic kilometer of South Pole ice that catches neutrinos from
• 1:33:20
  Speaker 1: distant galaxies. It was actually built in 2011, and the
• 1:33:25
  Speaker 1: first cosmic neutrinos were identified In 2013. So. One scientist.
• 1:33:34
  Speaker 1: Pursues a seemingly crazy idea. For two decades. But it
• 1:33:38
  Speaker 1: gets built. Science advances. Knowledge grows. This is the human mind.
• 1:33:46
  Speaker 1: And ambition. At its best. All right. That, my friends,
• 1:33:52
  Speaker 1: is the news, mainly AI news, but some other news
• 1:33:56
  Speaker 1: for this October 6th, Tuesday, October 6th. And tomorrow we'll
• 1:34:02
  Speaker 1: be back to our regular hours, 3 p.m. East Coast time. But, yeah,
• 1:34:09
  Speaker 1: thank you guys for being here and listening. I will
• 1:34:14
  Speaker 1: remind you, you can ask questions. I will answer all
• 1:34:16
  Speaker 1: your questions. You can make comments. I will give you
• 1:34:20
  Speaker 1: all the comments. I will read all your comments, put
• 1:34:23
  Speaker 1: it that way. So please jump in with comments, suggestions,
• 1:34:32
  Speaker 1: but mainly questions. Now, you can also support the show.
• 1:34:37
  Speaker 1: You can support the show without asking a question. It's
• 1:34:40
  Speaker 1: called a stick-up. I think we had a few stickers,
• 1:34:42
  Speaker 1: not many, but a few. Mary Eileen, thank you. Yeah,
• 1:34:47
  Speaker 1: and to get to our goal, we need a few
• 1:34:50
  Speaker 1: more stickers. CP Milken and Steven Kreisman, thank you. So
• 1:34:54
  Speaker 1: thank you, guys. But we need more stickers. We need
• 1:34:57
  Speaker 1: like some $ 50 and $ 100 stickers so we can get
• 1:35:00
  Speaker 1: closer to our goal. We're about $ 226 short of our
• 1:35:03
  Speaker 1: second hour goal, and we're certainly in the second hour.
• 1:35:06
  Speaker 1: And we really need to meet these goals on a
• 1:35:10
  Speaker 1: regular basis. You guys spent all your money on the
• 1:35:14
  Speaker 1: Nikos shows, but the new shows need to see some
• 1:35:19
  Speaker 1: love because, yeah, I'll shift to just doing Nikos shows otherwise.
• 1:35:24
  Speaker 1: All right. Reminder that you guys should all be signing
• 1:35:29
  Speaker 1: up to ARU as auditors or as full-time students, graded students,
• 1:35:34
  Speaker 1: not full-time students, graded students. Graded students get ARU for
• 1:35:38
  Speaker 1: free online. Uh, and, uh, auditors, uh, you know, pay,
• 1:35:42
  Speaker 1: I think you probably get a discount. Uh, check it out.
• 1:35:45
  Speaker 1: I ran.org slash YBS and go to the ARU. I'm,
• 1:35:50
  Speaker 1: as I told you, I'm, I just started my Western
• 1:35:53
  Speaker 1: civilization class on call. I think it's starting this week
• 1:35:56
  Speaker 1: or next, the, uh, full objectivism philosophy class, which is
• 1:36:00
  Speaker 1: going to be a three semester class. So check it out.
• 1:36:03
  Speaker 1: I think you'll learn a huge amount. Um, Alex Epstein,
• 1:36:08
  Speaker 1: the number one authority on things like data centers and
• 1:36:11
  Speaker 1: defending data centers. So check out his sub stack. He
• 1:36:15
  Speaker 1: writes important stuff there. He has become a huge proponent
• 1:36:18
  Speaker 1: of AI and I'm trying to get him to come
• 1:36:22
  Speaker 1: on the show and talk about AI. I'm really looking
• 1:36:25
  Speaker 1: forward to that conversation when it happens. And hendershotwealth.com slash YBS, hendershotwealth.com
• 1:36:34
  Speaker 1: slash YBS. I heard that one of my listeners said,
• 1:36:37
  Speaker 1: One of you guys has now become a client of
• 1:36:41
  Speaker 1: theirs and is going to be saving a huge, humongous
• 1:36:45
  Speaker 1: amount of money by deferring and eliminating his capital gains
• 1:36:53
  Speaker 1: taxes in the state of California and federal. It is doable.
• 1:36:59
  Speaker 1: He had a huge windfall from stock that he had,
• 1:37:03
  Speaker 1: a huge capital gains windfall. If you have... A huge windfall.
• 1:37:07
  Speaker 1: I don't know if you have some anthropic stock or
• 1:37:09
  Speaker 1: open AI stock or if you had SpaceX stock or
• 1:37:13
  Speaker 1: anything else, a business you're about to sell. Check out hendershotwealth.com
• 1:37:18
  Speaker 1: slash YBS. I'm actually going to do another interview with
• 1:37:20
  Speaker 1: Robert soon. So Robert Hendershot, who is my former business partner,
• 1:37:29
  Speaker 1: I'll interview him. We'll talk about AI and we'll talk about, yeah,
• 1:37:34
  Speaker 1: tax savings. How to... Use investment strategies to save on taxes.
• 1:37:40
  Speaker 1: All right, Arthur Evans. Will AI buy the mixed economy
• 1:37:46
  Speaker 1: another few generations before collapse? Buy, B-U-Y, not B-Y. Okay,
• 1:37:52
  Speaker 1: buy the mixed economy another five generations before collapse. I
• 1:37:56
  Speaker 1: don't know. I don't know how much it buys or
• 1:38:00
  Speaker 1: what exactly it buys. It is difficult to assess exactly.
• 1:38:06
  Speaker 1: It's certainly difficult to put a timeline on it. But look,
• 1:38:08
  Speaker 1: there's no question that AI is going to increase productivity,
• 1:38:12
  Speaker 1: therefore increase wealth. It's going to make it possible for
• 1:38:15
  Speaker 1: us to continue economic growth, continue prospering, continue surviving, maybe
• 1:38:22
  Speaker 1: even improving things like health care and education in spite
• 1:38:26
  Speaker 1: of government, in spite of the mixed economy. So, yes,
• 1:38:28
  Speaker 1: it definitely will buy us time. there's no no question
• 1:38:31
  Speaker 1: in my mind about that how much i don't know
• 1:38:34
  Speaker 1: jeremy thank you for the sticker really really appreciate it
• 1:38:37
  Speaker 1: all right i i forgot i should start with the
• 1:38:38
  Speaker 1: fifty dollars okay clark how is bill gates so brilliant
• 1:38:42
  Speaker 1: in one area but conventionally superficial and second-handed in everything
• 1:38:46
  Speaker 1: else how can a person with this intelligence suffer from
• 1:38:49
  Speaker 1: missing link syndrome. A man of his caliber should be
• 1:38:53
  Speaker 1: embracing objectivism 50 years ago. Yeah, he should have. He
• 1:38:56
  Speaker 1: read Ayn Rand, as far as I know. He read
• 1:38:58
  Speaker 1: Atlas Shrugged. But it's hard, it turns out. It's very hard.
• 1:39:02
  Speaker 1: And embracing objectivism does not correlate with intelligence, as you
• 1:39:07
  Speaker 1: can tell with some of the people on my chat. No,
• 1:39:09
  Speaker 1: I'm kidding. But it does not. There are super intelligent people,
• 1:39:13
  Speaker 1: super successful people, brilliant people, entrepreneurs who have not embraced
• 1:39:17
  Speaker 1: objectivism I mean there are very few great entrepreneurs that
• 1:39:21
  Speaker 1: have embraced objectivism and it's not that they haven't read
• 1:39:24
  Speaker 1: the books almost all of them have almost all of
• 1:39:28
  Speaker 1: them have read Atlas Shrugged so they know something they
• 1:39:32
  Speaker 1: at least know they know she exists and they know
• 1:39:36
  Speaker 1: something about her ideas but it's hard I mean I
• 1:39:42
  Speaker 1: think we all underestimate how hard it is not to
• 1:39:45
  Speaker 1: be conventional when it comes to something like morality. And that,
• 1:39:48
  Speaker 1: I think, is the downfall of Bill Gates. The downfall
• 1:39:51
  Speaker 1: of Bill Gates is his embrace and acceptance of altruism.
• 1:39:57
  Speaker 1: And once he embraced and accepted it, not as a
• 1:40:00
  Speaker 1: way for him to live his life, because I don't
• 1:40:01
  Speaker 1: think he lives his life on the basis of altruism,
• 1:40:06
  Speaker 1: but as the only moral code to preach and to
• 1:40:11
  Speaker 1: run his foundation based on and to advocate policy-wise. He
• 1:40:17
  Speaker 1: cannot give up on altruism. And that's the big challenge.
• 1:40:23
  Speaker 1: The big challenge is getting people to give up on altruism.
• 1:40:27
  Speaker 1: And it just, he's been successful. He lives a good life.
• 1:40:33
  Speaker 1: You know, but he thinks morality means altruism. He can't
• 1:40:37
  Speaker 1: conceive of an alternative. Juan Buffett is the same way.
• 1:40:42
  Speaker 1: I think Bill Gates learned a lot from Juan Buffett
• 1:40:45
  Speaker 1: about this stuff. He also believes that a lot of
• 1:40:50
  Speaker 1: his success is a result of luck. He and Buffett
• 1:40:55
  Speaker 1: both read, God, what's his name? Theory of Justice. John
• 1:41:01
  Speaker 1: Rawls' book on justice. They're fundamentally egalitarians in their moral beliefs,
• 1:41:09
  Speaker 1: even if not in their way of living. All right, Andrew,
• 1:41:14
  Speaker 1: a two-part question. You recall my recent emphasis of hatred
• 1:41:18
  Speaker 1: of the good for being the good regarding the AI debate.
• 1:41:21
  Speaker 1: You said that I claimed envy as the main motive,
• 1:41:25
  Speaker 1: but I didn't mention envy. My thinking was that altruism
• 1:41:34
  Speaker 1: causes the intellectual's negativity towards AI due to what it
• 1:41:40
  Speaker 1: portends for the wealth, comfort, success, and happiness of humanity.
• 1:41:44
  Speaker 1: Envy's effort effect is murky to me. Please explain how
• 1:41:49
  Speaker 1: envy affects the discourse on AI, the broader role of
• 1:41:53
  Speaker 1: envy on the hatred of the good. Well, that is envy.
• 1:41:59
  Speaker 1: Hatred of the good for being the good is envy.
• 1:42:04
  Speaker 1: That's how Ayn Rand defined envy. So when you wrote
• 1:42:09
  Speaker 1: hatred of the good for being the good, what you're
• 1:42:12
  Speaker 1: really saying is envy. You're basically given the definition of envy.
• 1:42:17
  Speaker 1: That's why I took it to be that. But yes,
• 1:42:20
  Speaker 1: I mean, I don't know. Look, intellectuals are motivated by
• 1:42:30
  Speaker 1: hatred of capitalism and hatred of freedom. They're motivated by,
• 1:42:34
  Speaker 1: you know, a belief that capitalism must fail and is destructive.
• 1:42:42
  Speaker 1: And it's a success, capitalist success, like the creation of
• 1:42:46
  Speaker 1: AI and AI competition and the improvement of lives that
• 1:42:49
  Speaker 1: AI is going to generate are perceived negatively by them.
• 1:42:57
  Speaker 1: Perceived negatively by them because it disproves everything they believe in.
• 1:43:04
  Speaker 1: So the altruism that causes them to hate capitalism, which
• 1:43:11
  Speaker 1: is All the altruism. Remember these. CEOs are all self-interested.
• 1:43:16
  Speaker 1: They're all motivated by their own self-interest, their own passion,
• 1:43:22
  Speaker 1: their own vision. Who are they? Who are they as
• 1:43:25
  Speaker 1: individuals to make decisions that are going to affect all
• 1:43:27
  Speaker 1: of humanity? That's the big issue. The big issue is
• 1:43:31
  Speaker 1: you've got a few CEOs out there, five, six people,
• 1:43:35
  Speaker 1: making decisions about that will impact every human being on
• 1:43:39
  Speaker 1: the planet, which is absolutely true. This is why they
• 1:43:42
  Speaker 1: cry for democratizing AI governance. We need the masses to
• 1:43:49
  Speaker 1: have a voice. They hate that. They hate the individual
• 1:43:52
  Speaker 1: decision maker. And they hate their success. This is the envy.
• 1:44:00
  Speaker 1: They resent the success of these people. They resent that
• 1:44:04
  Speaker 1: they're so powerful economically. They resent that they will have
• 1:44:07
  Speaker 1: so much impact on the world, even if the impact
• 1:44:09
  Speaker 1: is mainly positive. Because who are they? And they hate them.
• 1:44:17
  Speaker 1: They hate them for their success, for their virtues.
• 1:44:19
  Speaker 6: All right.
• 1:44:31
  Speaker 1: Arthur, are you still optimistic about the future? Yeah, I mean,
• 1:44:35
  Speaker 1: what alternative is there? You know, I think AI is
• 1:44:38
  Speaker 1: an expression of that optimism. I think basically it's going
• 1:44:43
  Speaker 1: to be good. So, you know, I'm optimistic about objectivism
• 1:44:52
  Speaker 1: winning in the end. I just think it'll take a very, very,
• 1:44:54
  Speaker 1: very long time. So it depends what you mean by
• 1:44:56
  Speaker 1: the future, how long into the future you're talking about.
• 1:45:00
  Speaker 1: The very, very long run, yes, I'm very optimistic. And
• 1:45:03
  Speaker 1: I think AI helps us. It's another mechanism by which
• 1:45:08
  Speaker 1: to disseminate truth. And we need to master it. We
• 1:45:11
  Speaker 1: need to figure out how to use it in order
• 1:45:13
  Speaker 1: to do that. Argentinian in Taiwan. Sam Harris said Bush
• 1:45:20
  Speaker 1: sounds like Churchill next to Trump. I think he's right.
• 1:45:25
  Speaker 1: That is scary. Ever watch the 08 clip of Bush Jr.
• 1:45:29
  Speaker 1: warns of nationalism, protectionism, and isolationism as upcoming threats to
• 1:45:35
  Speaker 1: the U.S.? I mean, yes, Bush looks brilliant and looks
• 1:45:42
  Speaker 1: amazing when compared to Trump, but he's really, really, really bad. Really, really,
• 1:45:48
  Speaker 1: really bad. I've seen that clip. A lot of it
• 1:45:52
  Speaker 1: is good, and yet he was awful. He was an
• 1:45:56
  Speaker 1: awful president from beginning to end. He adds, he's an
• 1:46:02
  Speaker 1: Argentinian after all, he adds, P.S., sad day. Messi is
• 1:46:06
  Speaker 1: retiring from international football. Yes. Messi played his last game
• 1:46:12
  Speaker 1: for Argentina today.
• 1:46:14
  Speaker 6: All right.
• 1:46:16
  Speaker 1: Ian says, a lot of the AI Duma stuff comes
• 1:46:19
  Speaker 1: out of the test community. And no one is seriously
• 1:46:25
  Speaker 1: examining that to understand it. Ideas like RSI FUM and
• 1:46:32
  Speaker 1: the singularity. You and Nikos could dig into the challenging topic. Yeah,
• 1:46:36
  Speaker 1: I mean, it's a good idea. But look, yes, and
• 1:46:40
  Speaker 1: a lot of that is coming from there, but it's
• 1:46:42
  Speaker 1: also coming from Elon Musk, and maybe he's influenced by
• 1:46:46
  Speaker 1: that community. Elon Musk and Dario and, of course, Sam Altman.
• 1:46:53
  Speaker 1: So it's coming from the very top. Hopper Campbell. As
• 1:46:59
  Speaker 1: I try to persuade people of objectivism, I have learned
• 1:47:03
  Speaker 1: to accept that I can explain the truth for you.
• 1:47:06
  Speaker 1: I can't understand it for you. Do we tend to
• 1:47:10
  Speaker 1: blame ourselves too much for the slow growth of the movement? Yeah,
• 1:47:16
  Speaker 1: I think a lot of people do. Yes. You can't
• 1:47:21
  Speaker 1: think for somebody else. They have to do the thinking.
• 1:47:26
  Speaker 1: They have to do the understanding. All you can do
• 1:47:27
  Speaker 1: is indicate. All you can do is point. All you
• 1:47:30
  Speaker 1: can do is give them a book. Give them an argument.
• 1:47:33
  Speaker 1: You can't do their thinking for them. And you learn
• 1:47:37
  Speaker 1: that when you raise kids. You can't do the thinking
• 1:47:39
  Speaker 1: for the kids. You can encourage them, incentivize them, beg them,
• 1:47:45
  Speaker 1: but you can't actually do the thinking for them. Stephen,
• 1:47:49
  Speaker 1: thank you for the sticker. Really, really appreciate it. Richard,
• 1:47:55
  Speaker 1: did you see the New York Times article last week
• 1:47:57
  Speaker 1: about the CEO of Anthropic Meeting with Religious Leader's to
• 1:48:00
  Speaker 1: advise them on Claude's morality. Yes, I mean, yes, this
• 1:48:07
  Speaker 1: is an ongoing phenomena. I mean, there were senior people
• 1:48:11
  Speaker 1: from Anthropic in the Vatican advising the Pope on his
• 1:48:16
  Speaker 1: encyclical about AI, and there was a lot of consultations
• 1:48:19
  Speaker 1: with Anthropic and the Pope. Anthropic has an in-house philosopher
• 1:48:26
  Speaker 1: helping write the constitution for the AI, which is not
• 1:48:32
  Speaker 1: a good philosopher. Luckily, I don't think the constitution matters
• 1:48:36
  Speaker 1: that much. But yeah, I mean, Philosophers are embedded in
• 1:48:41
  Speaker 1: all these companies and they're all pretty much conventional philosophers
• 1:48:47
  Speaker 1: of the left or of religion. There's nothing else. Neil says,
• 1:48:54
  Speaker 1: last messy game today. Thank you, Messi, for all the
• 1:48:57
  Speaker 1: magical memories. Absolutely. Jeremy Morton, how much do you want
• 1:49:02
  Speaker 1: to review a novel? Depends on how long the novel is.
• 1:49:05
  Speaker 1: Send me like the Amazon link. And I'll let you
• 1:49:09
  Speaker 1: know how much. It's basically a function of how much
• 1:49:14
  Speaker 1: time I'd have to spend on it. Feed rate command.
• 1:49:19
  Speaker 1: Have you tried the new AI agent out yet? Muse
• 1:49:22
  Speaker 1: is a free one, but I've been very impressed by Grokbot.
• 1:49:27
  Speaker 1: I mean, I'm using agents created by Claude, by Codex
• 1:49:33
  Speaker 1: and Claude. And they're amazing. I mean, they're the ones
• 1:49:36
  Speaker 1: who scan all the newspapers for me now. They scan
• 1:49:39
  Speaker 1: the news. They scan Twitter. They come back with their
• 1:49:42
  Speaker 1: recommendations for news items. So I'm using AI agents. I'm
• 1:49:46
  Speaker 1: not using the kind of consumer-friendly ones like Muse, not
• 1:49:51
  Speaker 1: yet anyway, but I'm using these other ones, and I
• 1:49:56
  Speaker 1: will continue using them. My goal is to automate as
• 1:50:03
  Speaker 1: much of what I do during the day as possible
• 1:50:06
  Speaker 1: and automate everything. instead of having a producer on the
• 1:50:10
  Speaker 1: Iran book show, operate, you know, automatize everything that I
• 1:50:15
  Speaker 1: do in prep for the show from preparing the news
• 1:50:18
  Speaker 1: all the way to opening the windows and getting everything
• 1:50:22
  Speaker 1: ready and having the right software open and all of that.
• 1:50:25
  Speaker 1: We'll see. I'm going to do all that. And then
• 1:50:27
  Speaker 1: on top of that, I'm pretty ambitious about what I
• 1:50:32
  Speaker 1: want to, you know, a potential, and I'll probably come
• 1:50:36
  Speaker 1: to you guys and ask for funding for this But
• 1:50:40
  Speaker 1: a Yaron AI Pro, something you guys can use that
• 1:50:43
  Speaker 1: will really, really be a powerful tool. Maybe a Capitalism
• 1:50:48
  Speaker 1: AI Pro that connects Yaron AI Pro with Alex AI Pro. Maybe,
• 1:50:55
  Speaker 1: you know, a lot of content from Ayn Rand. Amos
• 1:50:59
  Speaker 1: has an AI. Maybe we can get a bunch of
• 1:51:03
  Speaker 1: AIs from different disciplines. and create a free market, a
• 1:51:08
  Speaker 1: capitalism AI that, you know, that people can use. And again,
• 1:51:14
  Speaker 1: the more we use this, the more we populate content
• 1:51:17
  Speaker 1: with this, the more we influence the world out there. So, yeah,
• 1:51:23
  Speaker 1: I'm excited about AI generally and what we can do
• 1:51:26
  Speaker 1: and how we can use it to spread our ideas.
• 1:51:31
  Speaker 1: Gail says, Bill Gates, love equalization. He should read your book. yeah,
• 1:51:35
  Speaker 1: it's too late for Bill Gates. Again, he's very influenced by, uh,
• 1:51:40
  Speaker 1: by John Walls. And, uh, and that is, that is sad.
• 1:51:43
  Speaker 1: Instead of being influenced by Ayn Rand, he's influenced by Walls. Um, Buzz, uh,
• 1:51:56
  Speaker 1: thank you for the sticker. Buzz says, so what will
• 1:51:58
  Speaker 1: the producer of the show do if you automate his work? Well,
• 1:52:03
  Speaker 1: I don't have a producer, so, um, it's saving me
• 1:52:08
  Speaker 1: time because I'm my own producer. Sandy says, great timing
• 1:52:13
  Speaker 1: and great thoughts. Thank you, Sandy. I appreciate it. Neil says,
• 1:52:17
  Speaker 1: how do we adapt to AI? Past tech revolutions took generations.
• 1:52:22
  Speaker 1: AI is overtaking industries instantly. At 55 with a mortgage,
• 1:52:25
  Speaker 1: going back for a four-year degree isn't realistic anymore. I
• 1:52:29
  Speaker 1: don't think you need to. Particularly, you don't need to
• 1:52:32
  Speaker 1: given that AI can educate you about whatever it is
• 1:52:35
  Speaker 1: you need to adapt. So AI is an incredible education
• 1:52:39
  Speaker 1: tool that makes a lot of those four-year degrees irrelevant.
• 1:52:46
  Speaker 1: I also, it's not happening as fast as people think
• 1:52:48
  Speaker 1: it's happening, right? The only field that's really been massively
• 1:52:52
  Speaker 1: impacted is coding. And yet there are more programmers now
• 1:52:55
  Speaker 1: than there were before. It's not less programmers. It's just
• 1:52:58
  Speaker 1: their job has changed. They're now supervising AIs in programming.
• 1:53:02
  Speaker 1: They're doing more. With less. They're more radiologists than they
• 1:53:09
  Speaker 1: were before. So it's not yet destroying jobs as much
• 1:53:13
  Speaker 1: as it's shifting jobs. And what you need to do
• 1:53:16
  Speaker 1: is your particular skill set, what it is, and how
• 1:53:20
  Speaker 1: are you going to adapt to the existence of AI.
• 1:53:22
  Speaker 1: So ask AI what you should do. Tell AI exactly
• 1:53:27
  Speaker 1: your situation. Tell AI what you do and how you
• 1:53:30
  Speaker 1: do it. Tell AI what your skills on your interests are.
• 1:53:33
  Speaker 1: And ask it, how should I adapt to AI in
• 1:53:37
  Speaker 1: the future changing or taking my job? It'll give you
• 1:53:41
  Speaker 1: a good answer. Mary Alene, if the masses want a voice,
• 1:53:47
  Speaker 1: the masses can buy stock.
• 1:53:51
  Speaker 2: I agree.
• 1:53:52
  Speaker 1: I agree. All right, guys. Thank you all. Thanks to
• 1:53:55
  Speaker 1: all the Super Chatters. Thanks to everybody who was here.
• 1:53:58
  Speaker 1: I appreciate it. I will see you all tomorrow at
• 1:54:01
  Speaker 1: our usual time, 3 p.m. East Coast time. That'll be
• 1:54:04
  Speaker 1: true of tomorrow, Thursday, and Friday. So the rest of
• 1:54:10
  Speaker 1: the week is 3 p.m. East Coast time, except for
• 1:54:13
  Speaker 1: Saturday and Sunday, I think we'll be at 2 p.m.
• 1:54:17
  Speaker 1: to accommodate Nikos and to accommodate my schedule. But the
• 1:54:21
  Speaker 1: rest of the week, 3 p.m. All right, guys. See
• 1:54:24
  Speaker 1: you tomorrow. Bye, everybody.