Transcription
We're trying to capture all the benefits of AI while maintaining human values and dignity, right? And I think we, speaking as the local humans, we clearly care about that. In the book, we say there's a scenario where we are the dogs to the AI human equivalents. Um, and so we've, we, as the top animals, if you will, in our 70 years of 70,000 years of existence, whatever it is, have never had, you know, like the Neanderthals were, uh, we're not quite at our level, shall we say, we've never had this arrival. And, and we spend a fair amount of time talking about what happens when that thing, whatever you want to call it, discovers things that we don't understand.
So, former Google CEO Eric Schmidt just gave another banger interview, and this time he's called AI a ticking time bomb. He talked about where we, as humans, stand with the upcoming AI of 2025, which is going to be a game-changer. So, let's dive in and listen to all of that.
Um, Eric, before we dive into your book, um, it's been about 5 years since you left Google and Alphabet to begin focusing more on your philanthropic work. Um, you've launched a number of initiatives focused on finding, identifying, and funding the next generation of scientists and innovators and leaders. And so my question to you is, what is your philosophy around this work and what do you hope to accomplish?
Well, thank you, and thank you guys for coming. Uh, the quick answer is, when you think about my success, it's fundamentally because of science. It's fundamentally because I was lucky enough to be born into the right family and the right educational system with the right scientific advantages, and everything sort of took off after that. So, and so I'm now, I think, second or third largest science funder in the United States, and science can be understood as talent plus things, right? So, talent is obvious, and America is a magnet for the best scientists, the best mathematicians, so forth, in the world. It's very important for American strength that we continue to do that. Um, and I fund a lot of human activities, various awards, various practitioners. We're talking about thousands of people. Um, and I have no idea what kind of science they will ultimately do, but they're a hell of a lot smarter than I was. So, each of these generations is just so much more impressive for whatever reason than the previous ones, and I'm, I'm, I know what I was and I know how much better they are.
The other thing I've been doing is trying to fund things in climate change, various physics problems, and so forth. And there I discovered I don't really like being a philanthropist. I really like building things. I'm sort of a builder. So I decided to build interesting things. So, for example, I'm trying, I have a rule that I'll do anything that I can do that's 10 times cheaper than what the government can do. Okay? Not 10%. Yeah. By the way, there's a lot. So I'm not talking about 10%. I'm talking about 10 times. And so I'm currently working on a design of an outer space telescope that would be launched from the local rocket company. And off we go. And, um, again, using modern techniques, modern scientists, you can build them relatively inexpensively compared to what the government does. I can give you example after example. I've built an a climate change monitoring system. One of the core problems with climate change is we don't have the accurate math and science around the real interactions at the hyper-local level. And so all of these decisions that the world will make about, you know, solar versus wind and so forth and so on are based on either poor science or no science. We're working on that.
Um, I mean, there's just a long list. I got extremely interested in, uh, the rate at which science will be impacted by AI. So, remember that science does not change based on the faculty members. It changes based on graduate students. So, the key idea is to fund the graduate students to apply AI to some problem. And I didn't understand this, but in most of science, they know the math because somebody 50 years ago did the formulas, and it, and the formula is named after them. But the data computation is too complicated even for today's computers. But using AI, you can build approximations. With those approximations, you can solve the problem. Once the system understands it, it can also generate new candidates. So, the basic rule is, once you understand using AI a system of organization like biology, which is poorly understood, we still don't even have a digital model of how a cell works, you can generate them. And so the, the power once you can understand it and then generate it is enormous.
Now, in the upcoming part, Dr. Eric Schmidt says that AI is going to surpass humans in every capability. Watch this. So this leads, and this is sort of the AI speech. Um, do you want drugs faster? Do you want cures to every known human disease? I've got one project which claims, you know, these are scientists, God knows, that they can find all human druggable targets within two years if, if we just fund a little bit more of their AI and GPUs. Like that's a big deal, right? And we'll see, right? But the, the return for that kind of discovery is profound in medicine. There's lots and lots and lots of those.
Um, thank you for that. And you also find time to write books. Um, the current book that we're going to talk about tonight for a little while, um, just reached the bestseller list of the New York Times, the LA Times, and USA Today last week. So, congratulations on that. And we'll, and we'll talk about your, your co-authors in, in a few minutes. But perhaps the most provocative, um, assertion that happens in the book, Genesis, from my perspective, was the notion that since the dawn of human civilization, humans have pretty much found themselves at the top of the food chain. And perhaps that's changing, and changing pretty rapidly. So, Eric, explain to us what you mean when you say that humans will no longer be at the top of the food chain.
So, when I think about this book, the way I like to think about it is, uh, and this is my interpretation. We're trying to capture all the benefits of AI while maintaining human values and dignity, right? And I think we, speaking as the local humans, we clearly care about that. In the book, we say there's a scenario where we are the dogs to the AI human equivalents. Um, and so we've, we, as the top animals, if you will, in our 70 years of 70,000 years of existence, whatever it is, have never had, you know, like the Neanderthals were, uh, we're not quite at our level, shall we say, we've never had this arrival. And it, and we spend a fair amount of time talking about what happens when that thing, whatever you want to call it, discovers things that we don't understand, right? And, and I'll give you some examples soon. And what does it mean to coexist? So, this is not a book about how robots are going to take over your jobs and then, as you know, in the movie, the female scientist eventually destroys the robot and wins the, you know, wins the movie, right? That's not what we're talking about here, although it's a very good movie. Um, what we're talking about is our coexistence with this, and the coexistence is very interesting. And before we talk about the negatives, which are much more interesting, let's talk about the positives. Now, you're in a production business, you do a lot of research, you wrote, write a lot. AI should double your productivity, right? It should allow you to run twice as fast, generate twice as many memos, do as many conferences, reach people, get twice as much reach. You personally should be able to do all of that for each one of you. I can give you an example of a factor of two, right? So, I'm historically a programmer. Some of my code is still being used, which is sort of shocking. It's time to replace my code. But the important point is, um, if I were a programmer today, half my code would be written by the computer. It would finish my code because I'm sort of lazy. I was a lazy programmer. Um, so let's think about doctors. You can have twice as good outcomes or more. You can have treat twice as many patients. If your goal is efficiency or outcomes, you can choose. Um, what about lawyers? Well, you can sue people twice as much, you know. I mean, it seems like everybody wins, right? Um, at least at the level of information workers. Now, there are jobs that are displaced, and everyone always says, "Well, what about me in my job?" Like, I'm a programmer. My programming job is going away. That's not how automation works. Automation takes away the jobs that are the least human, the least interesting, or the most dangerous. That's always been true. This is just the same wave. So, the important thing is that, um, this underlying technological revolution is going to happen, and what I will say to this group is that this, this wave is happening much, much faster than you understand, than I understand, than governments understand. Our systems are not ready for this, right? It's, it's simple things like copyright, right? How do you, you know, the copyright, everyone, all the lawyers are worried about that. I'm not worried about it. They are. It's a legitimate concern. But what happens when your child's best friend is non-human? Those of you who have kids, you have a 12-year-old, 13-year-old, how do you feel about that? It's coming and coming very quickly. Um, that sounds great. Um, clearly you do not have a 13-year-old. I have a 12-year-old. You have a 12-year-old. You want, yeah. You're you out with their friends and in with the new friend. Out with the old, name of the new.
Um, so in terms of what you might consider to be existential threats, um, you know, your co-author, and by the way, we're bringing Craig Monday here, uh, next month to, uh, Fiana. So that will be very exciting. But let's talk about your other co-author, who knew quite a lot about existential threats. Um, tell us, Eric, the origin story of how you met Henry Kissinger and began to develop a relationship with him.
So, Henry was my best friend, and he, as you know, died a year ago. And to work with somebody of his brilliance is a special privilege. And my strong advice to you is, if you know somebody who's that brilliant, especially if they're slightly weird, they need some help. Talk to them, right? You will, you will benefit so much from the mentorship and so forth. In Henry's case, I met him when he was, uh, 85 or so, and it was at a, a Bilderberg conference, which can be understood as a conference for old people. And, uh, and he was obviously old, and he, um, uh, I knew who he was, obviously, and chatted with him, and he said, "I really have a problem with Google." And I said, "What?" He said, "Well, look at the results." I said, "Well, I did. They look accurate to me." Which is the wrong thing to say to Henry Kissinger. Um, and so I invited him to Google, where he showed up and walked on stage and gave a speech. And he began by saying he was convinced that Google was a threat to civilization, and the Googlers loved it because it made them feel important. Um, so that began the process, and of course, Henry had to learn how Google worked. He didn't really understand how algorithms worked and so forth. He's, you know, not his background, but he's a quick study. And what I learned, which is why the book is so relevant, is when, before any of us were alive, um, right after the war, he wrote his undergraduate thesis, which is the longest undergraduate history thesis in the history of Harvard, um, on the relationship of reality and Kant. And he's very, very interested in human perception and how we perceive the worlds around us. And he became convinced 15 years ago that this online thing was going to do to us what it has now done. So, for example, I meet all sorts of people who seem to be confused between the difference between the online world and the real world. They get confused. They, they seem to be living in the wrong one and behaving in the wrong way. Um, and this is going to become, this is going to be exacerbated by AI to a level that's hard for all of us to understand. So that's how it started for the books. Um, he wrote an article on basically the future of AI with my assistants, and then that got enough press, and he wrote another book on leaders, and his his favorite leaders were Mao Zedong, Nixon, Anwar Sadat, Golda Meir, you know, this is who he's writing about. And all anyone wants to talk to him about is his AI article. So he got the market feedback, he's like 95, and said, "I want to work on this." So he wrote the first book, and now the second book. He finished this, to be honest, a week before he died. He couldn't even get out of bed. That's how important this was to him.
I read that he died about a year after ChatGPT came out. Yeah.
Um, Eric, on the, on the, not on the subject of Henry, and, and one of the other things that, as we know, Dr. Kissinger is known for, is opening up China. And you're thinking a lot about, um, the, the AI race between the US and China, and you've been very clear on, on the fact that you want the US to win in the battle against, the battle for AI hegemony. So, you, you and I were at a speaking of conferences, you and I were at a conference together in Beijing in 2019. It was hosted by Mike Bloomberg. Y, and Henry was there. And what occurred to me at that event was that there's so much entanglement, right? It was meant to be actually a diplomatic back channel between US and Chinese CEOs. And my question to you is, how does the US position itself for victory in this battle when there's so much collaboration going on scientifically and in business between the US and China?
So, um, so Henry was very concerned that China and the US would get, uh, in opposite opposition on these issues. So he put together what is called a track two dialogue, which Craig and I are now doing, uh, in his, in after Henry's death. And it's hilarious because we have like these meetings, and you have all the Americans on Zoom, and they're all dressed like us, and on the other side is a table, and the big guy is in the middle. They all are wearing the same suits, and they're all reading from their same scripts. So, and by the way, there are typically nine to 11 of them. So it takes quite a bit of time to get through the meeting, and they don't say very much. And I said to him, "This is what diplomacy is?" He says, "Yes." I said, "How did you get through a hundred years of this?" Right? It's mind-boggling. But the stakes are huge. And, and let me give you a little bit of the background and why, if it's okay, I want to explain why this is so important. Um, when Henry and I, his last visit to China was, uh, two summers ago, and I was there with him for the, and we met with Xi, and all the usual, you know, people running the country with motorcades and the whole bit. Um, it was pretty clear to me at the time that they were two years behind. The way you find out what a country is doing is you don't ask the politicians because they're, they're either confused or won't tell you the truth, especially in China. But if you ask engineers a very precise question, they will answer your question. So, the question to the engineers in China was, "What are the scale of the training runs for foundation models?"
Now, in the upcoming part of the interview, Dr. Eric Schmidt emphasizes the US need to win the AI race against China. If the US loses to China, it's going to affect all of us. And the answer is, there weren't any, and there weren't just one getting started. Fast forward to now, China has produced three models, um, something called DeepSeek, something called Quen, and something called Hongyang, and I apologize for the mispronunciations, which are at the same level as OpenAI's O1 or close enough. That is a remarkable achievement. And in China, what happens is they have civil-military fusion. They have a 2025 plan. They've identified AI. They're going to dominate by 2030. They're going to do it. There is no other country that can amass the money, the people, the knowledge, and focus that China can. None. And the Europeans are largely hopeless, having spent lots of time working with them on this. They just don't have the money. They don't have the people. They don't have the energy. They can't do the financings and so forth. But China can do it. So, it's going to be a US-China race. There's a separate question which Henry and I were never able to resolve, which is what happens to the other 197 countries. But for those two, it's that fight. Now, why is the fight so important? I'll give you a bit of background. So, you all understand ChatGPT as language to language. If you want to know the ChatGPT moment for me this year, it was NotebookLM. And if you haven't played with it, it's a Google product. And if you're confused by what it is, there's a, just type in your favorite search engine, NotebookLM, not human, and you'll find out why it's so extraordinary. It's essentially a way of looking at documents and media and analyzing them with a really fun way of presenting it. Um, that's last year's. What's next year's are agents, large context windows, and automatic coding. Agents mean essentially a system with memory. So, it gets input. It learns something by generating candidates. It tests things. So, turn on and off the lights in the room. Eventually, it figures out that turns off that light, that turns, and remembers it learns in a microsecond, and then it produces its outcome. You can concatenate agents, and so almost every business workflow is an agent. So, all the businesses that you're in will find themselves with agentic solutions where the agents will all be assigned to get the customer, bill the customer, service the customer, complain about the customer, rate the customer, so forth. All of that stuff, do the accounting. There's, and this, there's are generally called remote process automation or robotic process automation. That's going to be a huge category. Um, the issue of context windows is the ability to use use the systems for short-term memory. You can basically take what it said, feed it back in, and say, "What do you think next?" And the third one is automatic programming. One of the key things that's, and I'll just give me 30 more seconds to explain because of those things, the following two facts are likely true. The first is, in the next year or two, we're going to have breakthrough super programmers, that is AI systems that are programmers, and we're probably going to have breakthrough AI mathematicians. Why? There are many, many areas that people are pursuing. But in software, it's essentially they're called scale-free. If you're doing software, you can just keep making more software. You don't need to go consult something. And furthermore, you can run the software, and you get a signal as to whether it works. So, you try, you try, you try it, you learn it, you're done. You try, try, try, learn something new. Try, try, learn something new. These computers are supercomputers that can do that in small numbers of minutes. And you have a polymath programmer. So, so much for my job. For a mathematician, the same is true because both mathematics and programming have a small number of language items and a clear objective function. And these systems that you hear about are built around next-word prediction. It's simpler if you have a smaller vocabulary, and it's simpler if you have a training fund. It's easy if you understand it that way. There's a gazillion startups that are doing both. They're going to be successful. When that happens, what do you need? You need programmers and math. Now, this is the slope of improvement in the US. This is the slope of improvement in China. The first one that gets to the ability to do the AI research agents, the slope changes because instead of having a thousand programmers, which you have at Google or AI, OpenAI, or Microsoft, or what have you, you have millions. You just launch them and say, "Try this, try that," and so forth. That slope. So, you sit there and you go, "Why is he talking about the slope?" In my world, if you're growing this quickly, your competitor cannot catch up because by the time the competitor gets next to you, you're already up here. It's crucial that America be the leader. It's going to be difficult enough with China as a fast follower. They work harder, their products are better, they're cheaper. There are all sorts of reasons why the Chinese are dominating this space. They have a different economic model. But if we don't get there first, we lose the fundamental leverage of our educational system, our scientists, and so forth. Sorry for the speech. This is, I'm trying, I'm trying to get everyone in our government to understand this. And they all kind of go, "Okay, and then what do you propose?" And I say, "Well, here's an example. There's a series of meetings last week on this in the government. Um, maybe you could do a deal with the tech companies. They need energy, and you need access to what they're doing. Maybe you can figure out a way to get them more energy because these things are huge energy hogs, in return for some national security help." There's a deal to be done in the American system. They don't need to do a deal in China because they've already been brutalized to do it. They've been given billions of dollars, and they're going to do it or not, you know, but, you know, they're on that path. We need to be on the same path.
Well, let me press you a little bit further on that. So, um, there's this tension then, um, between the need for very fast-paced innovation, um, to make sure that the US wins the war on AI. Um, there's also a need for regulation around some of the existential threats. And just indulge me while I tee this up, um, for just a minute. Eric Schmidt was awarded the Department of Defense Medal for Distinguished Public Service for your work as chairman of the Innovation Board for the Department of Defense. You served as chairman of the US National Security Commission for Artificial Intelligence. You served as a member of NASA's National Space Council. You are the founder and chair of the Special Competitive Studies Project, which makes recommendations for how to strengthen America's long-term global competition in the age of AI. And you served as the National Security Commission on Emerging Biotechnology. So, as someone who is underqualified to really be making these, um, kinds of assertions, hypothetically, if the president-elect had chosen you instead of Elon Musk to be a senior technology advisor, what would be, and you've alluded to this, but what would be your first two very tactical recommendations?
Um, well, there's a long list, but in order for us to be successful as a nation, we have a deal between business, the pro, the universities, and the government. The government comes up with essentially at-risk money in various forms, and they provide a, a sort of a structure to operate in. The universities produce this incredible talent, and the businesses, through venture capital and so forth, acquire it. That's how our national security works. By the way, the military doesn't actually build anything. They contract with private firms to build it. I know this because I spent a lot of time on it. So, the first thing I would do is ensure that that model is strong. So, what does it need? It needs more research money for universities. It needs more foreign visas for high-skills immigration. It is, of the many stupid things our government does, the single stupidest thing we do is take somebody who's a math PhD and kick them out of the country. Now, by the way, you may not like them. You can, there's plenty of states you can put them in if you don't like them, but keep them in a state in America. You see my point? It's insane. So, for example, the quantum initiative that China has was developed by a Chinese national who was trained in America. We wouldn't give him a visa. So, he went back to China, and guess what? He's building systems whose jobs are to decrypt all of your communications using quantum effects. They don't work yet, thank goodness, right? They've also launched satellites for quantum entanglement that the US hasn't done. That's not okay. It's just not okay, you know, wake up, guys. So, I would, I would first focus on that to making that work.
I think the second thing is that, uh, today, economic growth and shareholder growth and so forth is coming from these very large tech companies. Now, there's many things they do wrong, but at the end of the day, they are national treasures. And I, for one, having been the CEO of and chairman of one of them, I'm, I'm tired of people not recognizing. So, for example, when during the, uh, COVID, I was largely here in Miami, and how did I get food? Amazon. Thank you, Amazon. Okay. Thank you for, for making it work. They, they really were there, right? So, so you see my point. There's kind of this weird attitude. So, I'm not suggesting that you give large companies unfettered attitude, unfettered access, but I will tell you that these are scale businesses, and the, the companies that win are going to be big, and you're going to have to deal with how to regulate them. I'm sorry to sound so, so corporatist, but I think that's true.
Today, the, um, uh, another thing I would do is, um, this is a much longer conversation. I, I spent a decade working with our military. I'm a huge fan of our military people. I'm not a huge fan of anything else in the military because their systems are 1980s. Um, there's an old joke that at the rate procurement is going, there'll be one airplane made. It will be used on Mondays by the Air Force. It will be used on Tuesday by the Marines. On Wednesday, it'll need repair. And on Friday, it'll be recertified. And on Saturday and Sunday, we take the day off. Um, so, the, there's a real problem in the way we do our national security. Um, I've been heavily involved with working with Ukraine and trying to understand autonomy and abundance, sort of what I focus on my world, and it just makes no sense to build all these huge aircraft carriers and tanks and so forth. Um, tanks are very dangerous now because of drones. We should have major, major drone programs in America. And by the way, all of the current drones are dependent upon Chinese subassemblies. That's a problem. We need to fix that. We need to build factories to build the things for drones for national security. I mean, I can go on.
Yeah. You had mentioned, and we're going to go to questions, um, in just a few minutes, you you had alluded to, um, the fact that there is a whole world out there besides the US and China, and what, uh, those countries are doing to innovate or not on AI is critical. And as Dr. Kissinger knows, the world is safe when everyone is doing well and everyone is, um, more prosperous. So, Foreign Policy magazine just ran a story today saying AI is going to be very bad for the Global South. What can the US do to ensure that there is an even distribution of prosperity from AI?
Um, first place, I think these technologies are not very liberal in the way they work. They don't naturally produce equality. They don't naturally redistribute the gains. The fact of the matter is that a relatively small number of people and a small number of countries who move quickly and can operate at scale, which is sort of what I'm talking about for our nation, we win, and everyone else can't keep up. They can't keep up because they don't have the economic systems. They are not smart enough. Their educational systems aren't good enough. They can't get the data, or in Europe's case, they have stupid regulations which no one seems to be able to fix. So, I spent a decade trying to get Europe to the table. Europe is not going to show up. And the reason they're not going to show up is they're stuck with, they have lots of problems. They're not growing. Everybody knows what's going on there. But they're not organized around innovation. So, as much as we complain about America, and everyone, you know, hates the politics, loves the politics, whatever, this is an innovation nation, and that innovation is what's driving it. It drives everybody crazy, but it also drives all of our growth.
If you look at the Global South, the core problem is most of the population is in the North. Um, do you include India in the Global South? Well, I mean, they would. So, India is an interesting case because India is a counterexample to this point. They're growing. I was just there for a week. Um, very close to a bunch of the senior tech people there. Um, and they have a shot because they have enormously brilliant educated, and Pakistanis, and that diaspora. So, I think India has a shot. I think India, China is going to win. Russia has essentially, because of the war, put itself into the Stone Age and gone back 20 or 30 years and done horrific damage. That's got to get fixed. But that's not recoverable in the short term. Africa doesn't have the oomph. So, one of the questions that I don't know the answer to is, let's pick a nice, pleasant country in Africa that's agrarian, well-meaning, reasonably democratic. When this stuff shows up, what do they do? Now, if you look at Europe, the Europeans tried to have it all ways, which makes sense. They wanted to sell luxury goods to China, and they wanted cheap energy from Russia, and they wanted security from America. And that structure is unraveling economically. It's a hard problem. Um, they're going to have to have a serious reckoning as to whether they want to grow or not. One of the other rules about countries is grow or die. And here's why. In a democracy, you can't redistribute the money. So, let's assume we have a fixed budget. So, you know, you guys are in charge of this, you charge in that, and you and I are the elected officials. We cannot fundamentally take money from you to give it to them and vice versa. However, when there is growing money coming in from over there, we can add money to our pet projects or reward our voters or corruption, or whatever it is. You, you would never do this, but if I were misbehaving, I could do that. But I can only do it when I have growth. Democracies need growth. They don't work without it. Autocracies are fine without growth because they're not voting, right? And obviously, we want to be democratic. You have to, if I've learned anything in all these years working on this, democracy is important, and you have to fight for it. You have to fight for freedom.
So, um, let's go to questions and, and, and next, Eric, but before we do, just what do you say to some of the AI skeptics who talk about the pace of change being slower than, uh, you're anticipating, the incredible demand for energy, the incredible demand for computing power, frankly, the limits of human ingenuity hitting some sort of a wall at some point? Is there any chance that this is going to roll out a lot slower than?
Well, at the moment, they're wrong, and it looks like it's underhyped. Um, so the machine that I'm a part of is very good at announcing success before it delivers. Take a look at crypto, right? Crypto was going to take over the entire world. Um, speaking as a computer scientist, crypto is a great way of moving money around. It is not going to, with layer two chains, going to change the way we write programs. The latency doesn't work. All of us understood that. But the people promoting it still got valuations. But the AI stuff is real, right? It really does diffuse across every industry, every platform, and so forth. And it's a fundamental accelerant now, um, for each of the crazy, crazy initiatives. So, for example, Elon is busy raising another $10 billion for what he's doing. Doesn't have a lot of users right now. Most people would bet for Elon. He has a track record of delivering on these things. That's a, that's a something we could debate. But what you can't debate is that not only is the usage there, I'll give another example. There are 320 million users of ChatGPT, and their revenue is $10 billion-ish. By the way, this product is two years old. Two years old. Yeah. By the way, you know, we should have invented it. In fact, it's annoying that the people were at Google and they left to go to OpenAI and they invented it there. So, Google has responded with what they claim is a better, and that's how competition works. But all of this has happened in the last year, and the next year, the arrival of these agents, broadly speaking, are going to transform everything. And so, for example, you'll have to have an agent orchestrator to determine which agents you use. You're going to have to have an agent store. You're going to have to have agent security frameworks. You're going to have to have all the things that you have for the equivalent of an app store. Plus, they have to be able to talk to each other. And one of the things I will tell you is that the agents will initially talk in English to each other, which is fine because we can see what they're doing. When the agents on their own decide to come up with their own computer language to talk to each other, unplug them. You won't know what they're doing.
Eric, Miami is a real estate town. Real quick, tell your story about an AI agent building a new house.
No, it's, it's that this one's easy. Um, I happen to like Florida, and I happen to like Miami, and I want to build a new house, let's say up over there. Um, so, um, so here's, here's what I would do, right? I would say to a human, "Study the laws, look for some land, negotiate the prices." Now, each of these is steps. Um, and by the way, when you get the price negotiated, buy it, and then hire an architect and produce some plans, and then hire another contractor. And they're always overruns, so you have to watch them, and all this kind of stuff. And eventually pay them, and then take 20% away from them because they were late. You know, it's sort of how it works, right? And then there's a big fight. Every one of those things is an agent. So, the first agent is, "Identify where, where we can build a nice house." The second one is, "Analyze the all of the overlapping laws, right?" Nobody can understand. It takes a specialist to explain to you what the rules are about how high and all of these things. Um, the next one is to generate a design. These computers can generate really beautiful house designs, right? Um, get ready. They can. Um, the cost accounting and all of the planning, that's pretty easy. Selecting a contractor. Well, maybe the agent could actually go and call people on its own and get references and listen to them and listen and reason, because remember, these things are beginning to reason. Um, build the house, um, take pictures of the house. The, if somebody takes pictures and gives it, the agent can decide if progress is being made or it's being lied to. Right? Am I scaring you enough?
I was in India and I met with two founders. These are young, young men who are obviously super smart. And I said, "What are you doing?" And he said, "Well, there's a lot of small businesses in India." I said, "Yes, there's like about 100 billion of them. There is like immense, and none of them have IT skills." And I said, "Yes, that's obvious." You know, it's, you know, Joe and Bob in Indian language are, you know, live there, and it's their family business, and it's small. So, they're building an application which takes, you take pictures of the inventory, and that's all. So, what does the system do? It takes pictures. It figures out what the business is. It can do that. It's easy enough. It can then do a calculation of inventory. It can do that. And then it can figure out which products are selling by watching the changes of inventory. Kind of obvious. So, it can build an inventory system. Now, in India, using as an example, they tend to use WhatsApp Business to talk to everybody. So, it can generate promotions and so forth on a per-customer basis. Now, everything I just told you comes from a command study, my business, and make me more money. That's the point. You're dealing with a human-like level of intelligence to solve an important problem that you care about. You can be very sure that every one of those companies would love to make 10% more just from taking pictures of their inventory. Mhm.