What you must know before AGI arrives | Carnegie Mellon University Po-Shen Loh (YouTube Video Transcript)

(00:00:00) Your YouTube transcript will appear here (00:00:00) For many many years, [music] humans were (00:00:03) the top species, the most capable things (00:00:06) on this planet. Soon, it will not be (00:00:08) that case. The creativity [music] (00:00:10) in the AI can probably surpass what we (00:00:14) can do, too. The AI is advancing so fast (00:00:17) with my own kids. I have three kids. I'm (00:00:20) a math professor at Carnegie Melon (00:00:21) University. I'm not even sure what's (00:00:23) going to happen to university. In fact, (00:00:25) actually, I'm going to say as a parent, (00:00:27) I don't really even care too [music] (00:00:29) much if all of my children go to (00:00:31) university because I think that at this (00:00:33) point, by the time they go to (00:00:34) university, boy, the world will be so (00:00:37) different that the most important skill (00:00:39) that you could have is that ability to (00:00:41) synthesize your own idea. I interview (00:00:44) lots and lots of high school students (00:00:46) who want to work with me. And during the (00:00:49) interview, the way that I interview is I (00:00:51) ask them questions until it's very clear (00:00:53) from their body language that they have (00:00:55) never seen this question before. I wait (00:00:57) until it's really clear that they have (00:00:59) never seen this before. And then I want (00:01:01) to see how you think. The expectation is (00:01:04) that you won't solve it. And so then I (00:01:06) start to give hints. And then I want to (00:01:08) see how quickly can you synthesize them (00:01:11) into a solution for a problem we have (00:01:13) never seen before. Actually, (00:01:16) that's also creativity. We really need (00:01:19) this skill. Now, I'm Po Shanlow. I'm a (00:01:23) mathematician who likes to solve real (00:01:25) world problems. In real life, I'm a math (00:01:27) professor at Carnegie Melon University, (00:01:29) but I'm also a social entrepreneur where (00:01:31) I've been running my own educational (00:01:34) solution, which tries to make the world (00:01:36) a more thoughtful place. (00:01:42) My biggest surprise was last year the (00:01:45) International Math Olympiad problems, (00:01:47) four of them were solved by Google's (00:01:50) artificial intelligence. The (00:01:51) International Math Olympiad has six (00:01:53) questions and all six of the questions (00:01:55) are very very original. They are so (00:01:57) original that when the national coaches (00:01:59) meet, they all look at the problems and (00:02:02) they all try to make sure nothing too (00:02:04) similar to those problems has ever (00:02:05) appeared in any contest or anywhere in (00:02:07) the world before. The questions are (00:02:09) supposed to be really original. But (00:02:11) nevertheless, the artificial (00:02:12) intelligence was able to come up with (00:02:14) solutions to four out of six, which is (00:02:15) more than I can do. [music] The only (00:02:17) unique thing about human intelligence is (00:02:19) that we hopefully care that humans still (00:02:21) [music] exist. The creativity in the AI (00:02:25) can probably surpass what we can do, (00:02:28) too. In schools, one of the biggest (00:02:31) places where students are using AI to (00:02:34) cheat on their homework is for their (00:02:36) writing. This unfortunately could make a (00:02:39) huge problem for human civilization (00:02:41) because you just have to think what is (00:02:43) that AI anyway? It's a large language (00:02:45) model. How is that AI so good? [music] (00:02:48) It's because it's good at language. It's (00:02:50) good at looking at the patterns of words (00:02:52) that often appear. If many [music] kids (00:02:55) lose this ability, we'll get many kids (00:02:58) who grow up and aren't able [music] to (00:03:00) think logically. All they're able to do (00:03:02) is just take whatever anyone else gives (00:03:04) them. They'll just be dependent. If (00:03:06) you're already grown up and you already (00:03:08) have that skill and you're using the AI (00:03:10) to achieve that task because that's for (00:03:12) your job, great. [music] Okay, you're (00:03:14) using it to do a job. But if you're in (00:03:16) school, why are you doing that writing? (00:03:19) It's not because the writing you make is (00:03:21) going to make money directly. No, no, (00:03:23) that writing is actually part of your (00:03:25) own learning. Using AI to do your (00:03:27) writing homework in school [music] is (00:03:29) like saying, "I'm not going to run a (00:03:31) mile for exercise. I'm going to drive my (00:03:33) car one mile for exercise." how much (00:03:36) exercise you get, you get none. You're (00:03:37) going to grow up and you're not going to (00:03:39) be able to be as physically fit. Similar (00:03:41) thing here with mentally fit. And just (00:03:43) this observation that the power of the (00:03:46) large language model is the L the (00:03:48) language. That's why we need to really (00:03:50) make sure that all of our kids and if (00:03:52) you're watching this and you're (00:03:53) students, this is why you need to be (00:03:55) really really good with language for the (00:03:57) next generation. All of these skills (00:03:59) like reading and writing, communication, (00:04:01) logic, these are all going to be very (00:04:03) important because these are how you (00:04:05) develop a a good way to think. (00:04:08) People used to go to school to learn how (00:04:10) to do the homework and do the exams. (00:04:12) Today, everyone needs to learn how to (00:04:15) grade the homework. This is the huge (00:04:17) difference. I've done lots of different (00:04:19) kinds of teaching. I teach people all (00:04:21) the way from the International Math (00:04:23) Olympiad team. I will also go to schools (00:04:25) and teach sixth grade in schools where (00:04:28) unfortunately there might not even be (00:04:30) any math teacher for the whole seventh (00:04:31) grade. So I go and cover the entire (00:04:33) range of education. I find this to be (00:04:36) very interesting because that helps me (00:04:37) to learn what the challenges are. The (00:04:39) whole point of a school math test is to (00:04:41) see whether or not you listened and you (00:04:43) practiced. In fact, all the math (00:04:45) competition problems in the US and also (00:04:47) in many parts of the world are of this (00:04:49) type. Which is why today the way that I (00:04:51) approach education and training [music] (00:04:53) is to try to help as many people as (00:04:56) possible learn how to do those questions (00:04:59) which they won't have seen before. But I (00:05:01) want to emphasize the reason I I've been (00:05:03) doing a lot of work on this [music] (00:05:04) nowadays is because when I was doing (00:05:07) math competitions in the 1980s the way (00:05:09) you got good at it was [music] by (00:05:11) thinking. Every problem which was new (00:05:14) was a chance to practice mental (00:05:16) flexibility. Today, unfortunately, (00:05:19) there's a huge industry around test (00:05:22) preparation and cramming where people (00:05:25) try to help students get high scores on (00:05:28) these strange math questions by showing (00:05:30) you all of the strange math questions (00:05:32) that you might possibly see. And that (00:05:34) involves studying for many, many, many, (00:05:37) many hours. So that the hope of the (00:05:39) parents is that when the students (00:05:41) [music] see the test questions, they are (00:05:43) never surprised that they have done (00:05:45) everything many many many time. This (00:05:47) causes students to have to go to school (00:05:49) school and after school and so many (00:05:49) and after school [music] and so many (00:05:51) hours actually very bad for the student. (00:05:51) hours actually very bad for the student. (00:05:53) But even worse, it takes away the (00:05:53) But even worse, it takes away the (00:05:55) students chance to [music] invent. So (00:05:55) students chance to invent. So that's why (00:05:58) that's why what the world needs now is a (00:05:58) what the world needs now is a largecale (00:06:00) largecale way for everyone to learn how (00:06:01) way for everyone to learn how to grade (00:06:04) to grade homework, for everyone to learn (00:06:04) homework, for everyone to learn how to (00:06:06) how to come up with their own way of (00:06:06) come up with their own way of thinking, (00:06:07) thinking, not just how to do the (00:06:08) not just how to do the problems. (00:06:10) problems. (00:06:15) This causes students to have to go to (00:06:15) A long time ago when I started with (00:06:16) education, I was actually just thinking (00:06:18) about how to help people do math (00:06:20) problems. Today when I think back to (00:06:22) that time, I think I was probably a (00:06:25) solution looking for a problem in the (00:06:27) sense that uh somehow I thought it would (00:06:28) be very important for people to be good (00:06:30) at math. But then things that happened (00:06:32) later in my life as I became the (00:06:33) national coach of the US Olympic math (00:06:35) team, I saw situations where there were (00:06:38) so many so clever, so capable people who (00:06:41) were still so depressed. And (00:06:43) furthermore, after they graduated from (00:06:44) high school, they even didn't really (00:06:46) know what to do next because they (00:06:47) thought that the point of life was to (00:06:49) find ways to prove you're better than (00:06:50) other people. That's when I realized we (00:06:54) actually will do much better if we think (00:06:56) about the philosophy to start with, (00:06:57) right? The philosophy in life should not (00:06:59) be how do I outdo everyone else? If you (00:07:02) do [music] that, you will you'll (00:07:03) probably never be satisfied. But if your (00:07:06) philosophy in life is, hey, it is (00:07:08) actually addictive to make a bunch of (00:07:10) other people happy. Oh, now I can do it (00:07:13) for five people. Oh, now I can do it for (00:07:15) 500 people. Oh, wow. Now I can get (00:07:18) thousand people to come to this thing. (00:07:19) The more that you do, the more you want (00:07:21) to do. And the fun part is that (00:07:22) correlates also with traditional (00:07:24) success. Then I realized, ah, I should (00:07:27) be trying to push this worldwide. And if (00:07:30) I don't do it, who will? with the things (00:07:33) I've done in my life, I now have an (00:07:36) opportunity to go and say, you know, (00:07:39) I've seen what happens if you go all the (00:07:41) way in pure competition. I've seen what (00:07:43) happens if you go all the way and just (00:07:44) practice problems to do the best on (00:07:46) tests. (00:07:48) Actually, that's not the right target. (00:07:50) [music] (00:07:52) And I realized that because of my (00:07:53) background, I would be able to shift (00:07:56) mindsets. Then I said, okay, this is (00:07:58) what I have to do. Money doesn't buy you (00:08:00) happiness, but money is important for (00:08:03) impact and influence. [music] (00:08:04) So, in fact, it's very important that (00:08:06) the things that we build are capable of (00:08:08) generating enough money to create the (00:08:10) impact. This just happens to be what (00:08:11) drives me. 10 years ago, I had this (00:08:14) crazy idea that maybe if we made a (00:08:17) website that would collect people's ways (00:08:19) of explaining math and science topics, (00:08:22) then maybe people would explain the math (00:08:24) and science topics and it would be free (00:08:25) and everyone would be able to learn math (00:08:27) and science. And I remember thinking, (00:08:28) "Oh, that can't be very hard. We'll be (00:08:30) done with that in a few months. I'm glad (00:08:33) I thought that because I'm still (00:08:34) working." So, I had this whole thing (00:08:36) called XP. We were making a website with (00:08:38) free explanations, but that didn't (00:08:39) actually have business model of its own. (00:08:41) It wasn't generating money. So, I had to (00:08:44) find some way to support all of that. In (00:08:45) 2019 in April, we started creating our (00:08:48) own version of that in the United States (00:08:50) of America where we took charge of (00:08:52) filming me teaching and then we had a a (00:08:55) product which was consisted of me (00:08:57) teaching math that people could watch (00:08:59) recordings of and they would pay for it (00:09:01) and this this made some amount of money. (00:09:03) But then this one we still found there (00:09:04) were pain points. And finally about two (00:09:06) years ago I realized you know what what (00:09:09) people really want is to have a live (00:09:13) human experience with somebody else who (00:09:16) is an expert. The only problem is that's (00:09:17) quite rare and hard to find. Uh and (00:09:19) there's also another challenge which is (00:09:20) that ideally that person you're talking (00:09:22) to is friendly. Uh if the person knows a (00:09:24) lot but is not friendly that's actually (00:09:26) not useful either. Right? This is the (00:09:28) hardest thing to deliver in education (00:09:30) because it's the least scalable. Of (00:09:31) course, in entrepreneurship world, we (00:09:33) always think about scalability and yes, (00:09:35) you can find one brilliant coach who (00:09:37) teaches 10 students or maybe even 20 or (00:09:40) maybe even even 100. That's a small (00:09:42) scale uh compared to the size of the (00:09:43) world. And then that's when I suddenly (00:09:45) realized I can make a giant win-win-win (00:09:47) situation. So the main thing that I do (00:09:49) now is an ecosystem. It's actually an (00:09:52) ecosystem that I invented which unites (00:09:55) many different types of people to all (00:09:57) contribute [music] in ways where (00:09:59) everyone is winning. one pain point (00:10:01) which was for the people learning math. (00:10:02) Then the second pain point was from the (00:10:04) people who are very very strong at math (00:10:06) already from which building the EQ will (00:10:08) be even better. Although I do want to (00:10:09) emphasize this is helping them finish up (00:10:12) to become extraordinary. And the thing (00:10:14) that made me realize the key that made (00:10:15) me realize I could put everything (00:10:16) together was an experience that I had (00:10:18) about five six years ago which is that I (00:10:21) also took improvisational comedy classes (00:10:23) myself. improvisational comedy classes (00:10:25) are acting classes. [music] And I was (00:10:28) doing that because I was trying to learn (00:10:29) how to communicate better to get more (00:10:30) people interested in math. But I (00:10:32) realized that even a math nerd like me (00:10:34) can take those classes and then become (00:10:36) able to talk to a few more people. So (00:10:38) then I realized, let me add [music] (00:10:40) that. And then I walked over to our (00:10:42) drama department and I found out that (00:10:44) actually there are lots of people who (00:10:46) have extraordinary drama skills who are (00:10:48) actually indeed very interested in paid (00:10:50) part-time jobs to help to coach the high (00:10:53) school students. So that's the third (00:10:54) painoint. The third pain point is there (00:10:56) are people who absolutely love what (00:10:57) they're they're passionate about what (00:10:58) they're doing in the acting and drama (00:11:00) [music] world, but there's a practical (00:11:01) need which is well how to find a stable (00:11:04) part-time job flexible hours that they (00:11:07) can use to support their passions. So (00:11:09) suddenly win-win-win we have all three (00:11:11) lined up and that's why this thing (00:11:13) scales actually the everyone winning is (00:11:15) very important because I work with high (00:11:16) school students and so in our in our (00:11:18) company anytime anyone wants to ask high (00:11:21) school students to do anything my answer (00:11:23) to my employees is always that thing you (00:11:25) want to ask that high schooler to do can (00:11:27) we explain to their parent why for a (00:11:30) very busy high school student that thing (00:11:32) is the best thing they can do with their (00:11:33) time if I cannot explain that they're (00:11:35) not doing it so this this is the (00:11:37) discipline that we run it to this is (00:11:38) this how strongly we make it a win-win (00:11:40) situation. We will never have a high (00:11:42) school student doing something unless I (00:11:44) could explain myself to their parent. We (00:11:46) suggested for your daughter to do this (00:11:48) because it's really good for her to do (00:11:49) this. [music] The thing is beforehand it (00:11:52) was hard to imagine there would be a way (00:11:54) to do that until the answer became oh (00:11:56) yeah because while doing this they will (00:11:57) get to learn from a Broadway or (00:11:59) Hollywood quality actor or actress (00:12:01) that's going to help them become (00:12:02) extraordinarily successful. now win. But (00:12:05) you see, this took eight years to come (00:12:06) up with, two more years to scale. The (00:12:08) speed at which we grow is purely just (00:12:10) based on how long it takes for people (00:12:12) who have middle school children to (00:12:14) realize that, oh, there are these (00:12:16) classes here where the class looks as (00:12:17) good as a Twitch gaming stream and it's (00:12:19) taught by math geniuses who are smiling. (00:12:21) [music] You know, these are things that (00:12:22) people could not imagine that you'd put (00:12:24) all together at the same time. And as (00:12:26) people discover this, they actually (00:12:27) switch over. They start joining our (00:12:28) classes. Then we can bring more high (00:12:30) school students. And the scaling power (00:12:32) this can go to is we I estimate that (00:12:34) this easily could grow to a 100,000 high (00:12:36) school students in the US. That's 1% of (00:12:38) the US high school [music] students (00:12:40) teaching about a million middle school (00:12:41) students. If you look at our live (00:12:43) program, it looks a little strange (00:12:45) because you'll see that the only (00:12:47) subjects that we teach is a pretty small (00:12:49) number. Algebra, geometry, (00:12:51) combinatorics, number theory. Why do (00:12:54) [music] we cover these? We cover these (00:12:55) because these are a curriculum that (00:12:57) teaches you how to think. The only way (00:12:59) to do that is by giving them questions (00:13:01) that they have never seen in school (00:13:03) before. So, I need to find a source of (00:13:05) problems that you will not see in (00:13:07) school. Actually, that turns out to be (00:13:09) the middle school math competition (00:13:11) curriculum because the people making (00:13:13) those problems were trying to make (00:13:15) problems that you don't see in school. (00:13:16) But the difference between the way we do (00:13:18) it and the way that lots of the training (00:13:20) centers do [music] it is that we are (00:13:22) trying to use those as opportunities to (00:13:24) make you able to practice [music] the (00:13:26) thinking instead of just showing you (00:13:28) doing the question that way enough (00:13:29) times. So the answer is the topics that (00:13:32) we cover specifically chosen because (00:13:34) that will be enough to teach a student (00:13:36) how to generate their own idea. Our (00:13:38) philosophy is if you finish all of that (00:13:40) you will discover that you can learn (00:13:41) anything. Our goal is not to make it so (00:13:44) that we have classes for you for every (00:13:46) year of your life. Our goal is to make (00:13:48) it so that as fast as possible, you (00:13:50) don't need any classes from anyone ever (00:13:52) again. (00:13:54) Right? So that's one particular piece. (00:13:56) But going forward, I think that one of (00:13:57) the skills that people will really need, (00:13:59) it's that aspect of actually wanting to (00:14:04) create value and delight in other (00:14:06) people. Why do I say this? I say this (00:14:08) because for many many years, humans were (00:14:12) the top species, the most capable things (00:14:15) on this planet. Soon, it will not be (00:14:17) that case. (00:14:19) Soon, [music] you're going to have to (00:14:20) work together to survive. The only way (00:14:24) to get other people to want to team up (00:14:26) with you is for you to authentically and (00:14:29) deeply be a person who is motivated by (00:14:32) creating value in the other. If you are (00:14:35) not that way, you are a bad partner and (00:14:37) people will not want to go and team with (00:14:39) you. If they don't team with you, you (00:14:42) will die. [music] (00:14:44) You will lose opportunity because (00:14:47) eventually [music] all of these kinds of (00:14:49) jobs you can use AIS to do. Then why (00:14:51) would anyone want to employ you? Why (00:14:52) would anyone want to have you as as (00:14:54) someone on their team? Presumably only (00:14:56) because they somehow felt like you are (00:14:58) going to create some value and they they (00:15:00) like that vibe. They like you. I think (00:15:02) what we really need is to get more and (00:15:05) more people who are figuring out what (00:15:07) the real problem to solve is. But (00:15:09) unfortunately, sometimes when kids just (00:15:11) think about problems, they don't realize (00:15:13) that the way that you solve a problem is (00:15:16) through empathy and through relating to (00:15:18) other people. Why? [music] Because you (00:15:20) can't solve a problem unless you can (00:15:22) visualize it through their eyes. I do (00:15:25) spend a significant amount of my life (00:15:27) working towards the goal of being better (00:15:30) at simulating the world. I also use AI (00:15:33) for that. I think what I'll give one (00:15:35) exact example. I was just in Nashville, (00:15:37) Tennessee last week and I just saw a (00:15:39) really I thought very very talented (00:15:41) singer in one of the bars there. Wow, (00:15:44) she's good. And I just got curious how (00:15:47) hard is it to be able to get a (00:15:49) performing spot on Broadway in (00:15:51) Nashville. So I asked AI and I was (00:15:53) actually not just interested in AI tell (00:15:55) me about the I want to see the links (00:15:56) [music] you know tell me more what is (00:15:58) the background of this particular place (00:16:00) she's performing at right I I will make (00:16:02) my own conclusions based on knowing oh (00:16:04) oh I see so there's all of these (00:16:06) different people who would want to do (00:16:07) this thing right so if you're one of the (00:16:09) people who got picked to do it at this (00:16:10) prime time [music] oh this makes logical (00:16:12) sense see I'm using the AI to build the (00:16:15) logic inside my brain for understanding (00:16:19) country music performance And why was I (00:16:21) trying to understand? Honestly, it's (00:16:23) because these days I also work with (00:16:24) professional entertainers. So, I'm also (00:16:26) myself always scouting. The big heart of (00:16:28) this [music] is I wasn't using the AI to (00:16:31) write the report for me. I was using AI (00:16:33) to make myself better at that particular (00:16:35) goal. Being able to simulate the world (00:16:38) is the superpower that makes someone (00:16:41) able to be a successful entrepreneur. (00:16:43) Simulating the world allows you to (00:16:46) imagine a product or imagine a strategy (00:16:49) and then play it forward in your head. (00:16:51) What would happen if I did this? (00:16:58) The work I do is really focused on (00:17:02) building up autonomous human thinking. (00:17:04) That's why the core word I use for the (00:17:06) philosophy is thoughtful. I've watched (00:17:09) over many years as people are becoming (00:17:12) less and less interested in thinking. I (00:17:15) think that actually happened because (00:17:17) people found out that they can entertain (00:17:19) themselves with iPads and this (00:17:21) unfortunately makes people also not have (00:17:23) as much interest in concentrating and (00:17:26) thinking about something. AI could make (00:17:28) that much worse. The fun part of life (00:17:31) [music] is having your own contribution (00:17:34) to the life that you live. [music] (00:17:35) Actually, I think that's why people (00:17:37) actually like creativity. It's fun. (00:17:39) People like to draw. People like to put (00:17:41) their own flavor. People like to wear (00:17:42) their own fashion. It expresses themsel. (00:17:44) This expression of yourself, it will be (00:17:46) lost if everything you do is efficient (00:17:49) but just reliant on the AI which told (00:17:51) you how to dress today. That's why I (00:17:53) want more and more people to discover (00:17:55) it's fun to think. It's fun to have your (00:17:57) own twist on things, your own your own (00:17:59) idea that you inject inside. The other (00:18:01) dangerous thing that happens if people (00:18:03) lose the ability to to think and reason (00:18:05) is that it makes it far easier to (00:18:08) deceive [music] people. The world is so (00:18:10) complicated that if you look at any (00:18:13) situation in the world, sometimes (00:18:14) depending on how you tell the story of (00:18:17) what happened, you can say statements (00:18:19) [music] that are all true which make you (00:18:22) come up with a different feeling. I (00:18:24) think it's really important for people (00:18:25) to be critical and to for people to be (00:18:28) able to understand what's really going (00:18:29) on because sometimes when someone's (00:18:31) talking to you, they have an agenda. (00:18:32) Like I'll be frank, I have an agenda. (00:18:34) I'm trying to build a more thoughtful (00:18:36) world. And I'm going to be very very (00:18:37) upfront with you on that. Anyone who's (00:18:38) watching this video, I think it's really (00:18:40) important that we have as many people as (00:18:43) possible find out how much fun it is (00:18:45) [music] to delight other people and to (00:18:46) have the ability to think and figure (00:18:49) that out. That's my agenda. [music] But (00:18:51) you see, everyone has an agenda. And if (00:18:54) you can't think for yourself and you (00:18:56) just listen to some authority, what if (00:18:59) that agenda is [music] actually to your (00:19:01) detriment? You'll have no way of (00:19:04) knowing. The technology revolution (00:19:07) really made all of us start to realize (00:19:10) how much of an impact bias has in the (00:19:13) [music] sense that whoever makes some (00:19:16) technology tool has some bias. What does (00:19:19) bias really mean? Well, I guess as a (00:19:21) mathematician, the way I would say it is (00:19:23) 2+ 2 is always four. That's right. (00:19:25) What's the point of life? Oh, I don't (00:19:27) know. Like, there's no clear definition. (00:19:30) What's the point of life? I think the (00:19:31) point of life is is to delight as many (00:19:33) other people as you can, but I know that (00:19:34) you might not necessarily agree, and (00:19:36) it's not a problem. I think it's it's (00:19:38) healthy that we may have different (00:19:40) starting points. The part that becomes (00:19:42) unhealthy is where there is only a very (00:19:45) short menu of options each of which is (00:19:49) followed by a huge number of people. (00:19:51) That's actually where bias comes in (00:19:53) because we just we just mentioned so (00:19:55) [music] far in this in this video a (00:19:56) couple of different uh sources of AI (00:19:59) providers, right? We have we have (00:20:00) Claude, we have OpenAI, there's also (00:20:02) Gemini. If you're in China, there's (00:20:03) DeepSeek. There are all of these, but (00:20:05) that's relatively few [music] if you (00:20:07) think about it. That's that would be (00:20:08) sort of like saying well the world has (00:20:10) lots of different viewpoints it has five (00:20:13) of them really no no the world has 7 and (00:20:15) a half billion [music] (00:20:17) different viewpoints there are 7 and 12 (00:20:18) billion people one of the beautiful (00:20:19) things about I guess humanity is the (00:20:22) fact that there are so many different (00:20:24) ideas all out there and let's be frank (00:20:27) some of the ideas are bad some of the (00:20:29) people are unfortunately in prison (00:20:31) because they decided to kill someone (00:20:33) else hopefully [music] (00:20:34) we all understand that is a bad idea but (00:20:36) the point is there are lots of different (00:20:38) people who are trying different kinds of (00:20:40) ideas, lots of different philosophies (00:20:41) and in this great big marketplace of (00:20:43) ideas that is the [music] world we see (00:20:46) some ideas come out and the the variety (00:20:49) also allows us to have more creativity (00:20:51) perhaps and when I look at the different (00:20:53) AI tools well it's actually well known (00:20:56) that they have certain biases this is (00:20:59) also why for me when I try to get the (00:21:00) news [music] I don't only go to CNN.com (00:21:03) I also go to Fox News I tune my social (00:21:06) media So that my ex is all tracking (00:21:09) Republican [music] (00:21:10) uh right-leaning viewpoints and my (00:21:12) Facebook is all tracking left-leaning (00:21:14) [music] viewpoints. And I look at both (00:21:16) of them every day because I want to see (00:21:18) what's going on. And my expectation is (00:21:20) yes, you're going to be biased. You have (00:21:21) a certain view on the world and you (00:21:23) think you're right. And you're biased (00:21:24) too. You have a way of thinking of the (00:21:25) world. And my [music] job as I simulate (00:21:28) the world is to try to figure out where (00:21:29) do you disagree? Ah, you disagree (00:21:31) perhaps on a few values of how people (00:21:33) should live their life. And then that (00:21:35) causes you to have different ways of (00:21:36) reporting on the story. I [music] think (00:21:38) it's all the more important now that (00:21:40) there's AI out here, which sounds like a (00:21:42) very convincing, reasonable [music] (00:21:43) person. It's even more important that (00:21:45) people look at things and say, "All (00:21:48) right, is that really the story?" (00:21:51) Because I think that the AI is going to (00:21:53) be so good at looking complete that you (00:21:57) may think you have the entire story on a (00:21:59) controversial situation, but you don't. (00:22:02) What I do for fun is I like to meet and (00:22:05) try to understand people whose (00:22:07) backgrounds I don't fully understand (00:22:08) [music] yet. This is actually what I do (00:22:10) for fun. This is also why we're we're in (00:22:13) we're we're talking right now. I happen (00:22:14) to be talking to you in New York City. (00:22:16) The way I got here overnight is I (00:22:18) [music] took the bus, the overnight bus. (00:22:20) You know, some people don't take the (00:22:21) overnight bus because who knows who (00:22:23) you're taking the bus with. [music] But (00:22:25) for me, I'm actually not scared by that. (00:22:26) That's just called real world. You can't (00:22:29) understand the real world unless you (00:22:30) actually start going into various parts (00:22:32) of the real world. I think my message of (00:22:34) how to create value is you cannot create (00:22:36) value if you don't interact with people. (00:22:39) You cannot just theoretically think (00:22:41) about the value. And the more people the (00:22:43) more you can understand people of (00:22:44) different backgrounds. Understand means (00:22:46) have some [music] idea of how they tick. (00:22:48) What are their needs? What are the (00:22:50) limiting factors? What do they want to (00:22:52) do? The better you are at modeling this, (00:22:54) the more effective you will be at coming (00:22:56) up with a solution. I started going to (00:22:58) city after [music] city after city (00:22:59) giving math talks in public parks. I (00:23:02) actually set a schedule. I put a (00:23:03) schedule on my website and I said I'm (00:23:05) going to go to all these cities and (00:23:07) people could just sign up to show up for (00:23:08) the talks. And at the beginning people (00:23:10) were wondering will anyone show up? But (00:23:12) actually there would be like 50 to 100 (00:23:14) people showing up at these talks [music] (00:23:15) in parks. And by the way, that was a fun (00:23:17) journey because in order to do that, I (00:23:18) was traveling around from park to park (00:23:21) with all the AV equipment, [music] (00:23:23) speakers and everything to be able to (00:23:25) have a a stage in in in park shelters (00:23:28) all around the US. But while doing that (00:23:30) in inadvertently, that was customer (00:23:33) discovery because I was able to suddenly (00:23:35) interact with and talk to thousands of (00:23:38) parents and [music] students, which (00:23:40) started to make me realize what kind of (00:23:42) challenges people had. And that's why (00:23:44) just probably about [music] 1 to two (00:23:46) months after that, the big idea came. (00:23:49) The big idea of oh, we can actually have (00:23:52) all these middle school students who (00:23:53) I've met, they can all learn how to (00:23:55) think all at the same time while these (00:23:58) people who are brilliant become (00:24:00) extremely polished so that someday later (00:24:03) in their careers they can be really (00:24:04) successful. So that that that was that (00:24:06) was the idea. It was like somehow you (00:24:08) cannot really find pain points if (00:24:11) [music] you're not seeing people. That's (00:24:13) also why with a lot of the work that I (00:24:14) do, I will go into schools, right? I (00:24:17) love to work on education [music] and I (00:24:19) do it in a way where even just last (00:24:21) night I was writing to somebody who is (00:24:23) involved with a large network of schools (00:24:25) [music] and their schools, I believe, (00:24:27) serve students who are also (00:24:29) disadvantaged. [music] Now, instead of (00:24:31) just putting money or putting resources (00:24:33) from my side, what I said is I'm very (00:24:35) interested. [music] Can we arrange for (00:24:37) me to go into some of your schools and (00:24:38) teach sixth grade? What I'm explaining (00:24:40) is that the way I do anything, if I want (00:24:42) to work in a sector, [music] (00:24:44) I go and myself step in and start doing (00:24:46) the work and see what happens. And this (00:24:48) is actually how I came up with all these (00:24:50) ideas because actually the ideas that (00:24:52) I'm doing are all things I've personally (00:24:53) [music] experienced myself. I've (00:24:55) experienced being a math person taking (00:24:57) acting classes. [music] I've experienced (00:24:59) being a person learning how to think, (00:25:01) taught by somebody who knows a lot of (00:25:03) things and also smiles. So all the (00:25:05) different parts gave me these ideas. (00:25:09) And the other story I'll share is I (00:25:11) don't only go to rural America. I go all (00:25:13) over the world. In fact, as I started (00:25:14) running around rural areas, do I walked (00:25:16) into the elementary school? I walked (00:25:18) into the fourth grade classroom and I (00:25:20) was just going to do my usual thing, Mr. (00:25:21) Poe, the substitute teacher. So, I wrote (00:25:23) on the blackboard, "What is 1 + [music] (00:25:26) 3 + 5 + 7 + 9 =?" [music] As soon as I (00:25:32) wrote equals, behind me, I heard a bunch (00:25:35) of kids yelling 25. I've actually never (00:25:38) experienced before a classroom. The kids (00:25:40) were all suggesting ideas. They were (00:25:42) also very respectful of each other's (00:25:44) ideas. This was one of the best (00:25:46) classrooms I have ever taught. Those are (00:25:49) wonderful kids. And I just told you the (00:25:51) profile of the area was high poverty. I (00:25:54) asked the person who took me around (00:25:56) afterwards, "These kids are amazing. (00:26:01) Do they play games on their phones?" And (00:26:03) the lady told me, "They don't [music] (00:26:05) have phones. It's because of the money. (00:26:07) In fact, they might not even have the (00:26:09) internet access." So, that's an (00:26:10) interesting commentary. [laughter] Uh, (00:26:12) then I said, "What do they do for fun?" (00:26:14) And she said, "Well, they just [music] (00:26:15) figure out how to make their own games." (00:26:18) So, then I realized actually that huge (00:26:20) pool of authentically interested and (00:26:22) [music] curious kids throughout rural (00:26:24) America who are ready, have the wits and (00:26:28) the creativity which they've built from (00:26:29) the way of life that they have, very (00:26:31) [music] smart kids. This could be an (00:26:33) enormous untapped potential for (00:26:36) unleashing science and technology across (00:26:38) not only the US but the the entire (00:26:40) world.