WEBVTT 1 00:00:00.320 --> 00:00:01.080 Dot Org. 2 00:00:06.040 --> 00:00:10.960 KCAA is your CNBC News affiliate. We're the station that 3 00:00:11.400 --> 00:00:13.080 gets down to business. 4 00:00:16.120 --> 00:00:18.800 The views and opinions expressed by individuals on the following 5 00:00:18.839 --> 00:00:22.199 program do not necessarily reflect those of the network, Guys 6 00:00:22.199 --> 00:00:26.800 Guy Radio and its platforms. 7 00:00:28.399 --> 00:00:35.359 It's Guys Guy Radio. Here's your host, Robert Manny. 8 00:00:37.280 --> 00:00:40.520 Welcome to Guys Guys Radio. This is your host, Robert Manny, 9 00:00:40.600 --> 00:00:43.600 welcoming you to the show where men and women can 10 00:00:43.600 --> 00:00:46.840 be at their best and everyone wins. Guys Guys Radio. 11 00:00:47.240 --> 00:00:51.359 We're here to inform you, inspire you, empower you, and 12 00:00:51.560 --> 00:00:54.840 get you to think, feel, and who knows, maybe even 13 00:00:54.880 --> 00:00:59.000 act by virtue of the journeys, stories, experiences and insights 14 00:00:59.159 --> 00:01:01.479 of the guests. I you each and every week to 15 00:01:01.520 --> 00:01:04.319 the show, and once again this week is no exception. 16 00:01:04.400 --> 00:01:08.480 Today we're going to talk about AI artificial intelligence with 17 00:01:08.680 --> 00:01:14.959 author Agg Liu, who's got a book written with Ashwari 18 00:01:15.319 --> 00:01:19.319 How are PhD. The book is called The Handy Artificial 19 00:01:19.519 --> 00:01:22.840 Intelligence answer Book. It is terrific and I can't wait 20 00:01:22.879 --> 00:01:26.640 to get into our conversations with Agg because I think 21 00:01:26.959 --> 00:01:28.439 I'm going to learn a lot and you're going to 22 00:01:28.519 --> 00:01:29.760 learn a lot, and we're going to have a lot 23 00:01:29.799 --> 00:01:31.719 of fun here on Guys Guys Radio. So I ask 24 00:01:31.760 --> 00:01:35.640 yourself what does artificial intelligence mean for work, for society, 25 00:01:35.719 --> 00:01:38.280 for our lives? What is it capable of right now? 26 00:01:38.280 --> 00:01:40.879 And what might it be capable of in the future. 27 00:01:42.480 --> 00:01:46.439 It's popping up everywhere Google Search is consumer customer service calls. 28 00:01:46.480 --> 00:01:48.799 I just had one the other day where AI handled 29 00:01:48.840 --> 00:01:54.159 everything perfectly, Netflix, YouTube, TikTok, algorithms, real time traffic updates, 30 00:01:54.280 --> 00:01:57.239 and routing Siri at Alexa, to name just a few. 31 00:01:57.280 --> 00:02:00.519 It seems that AI is pervading more, more and more 32 00:02:00.560 --> 00:02:02.840 of our society and we've got to get a handler 33 00:02:02.879 --> 00:02:05.159 on it. A lot of people ask, is AI going 34 00:02:05.239 --> 00:02:08.960 to take my job? Well, I think the key is 35 00:02:08.960 --> 00:02:11.520 that you've got to learn how to manage AI. It's 36 00:02:11.520 --> 00:02:14.759 a tool. It's going to take some jobs, replace some jobs, 37 00:02:14.759 --> 00:02:17.000 and it's going to create other jobs. So you have 38 00:02:17.080 --> 00:02:19.240 to be on the creative edge when it comes to 39 00:02:19.479 --> 00:02:24.159 doing things that a computer or algorithms can't do. The creativity, 40 00:02:24.439 --> 00:02:27.879 the emotional connection with other people, the problem solving, and 41 00:02:28.080 --> 00:02:29.960 that's how I think people are going to be able 42 00:02:29.960 --> 00:02:32.560 to continue to grow and get ahead. Our side is 43 00:02:32.599 --> 00:02:34.280 going to change. It's the same thing when you know, 44 00:02:34.319 --> 00:02:37.039 the Internet came up, and email came up, and fax 45 00:02:37.120 --> 00:02:39.280 machines came up, and the telephone came up, and the 46 00:02:39.280 --> 00:02:43.159 TV came up. When all of these things were unleashed 47 00:02:43.199 --> 00:02:46.840 on our society, we managed to deal with it and 48 00:02:46.879 --> 00:02:49.479 we managed to continue growing. And I think that's what's 49 00:02:49.520 --> 00:02:51.599 going to happen in AI. And today we're going to 50 00:02:51.599 --> 00:02:54.520 talk about the earliest forms of AI, how it works, 51 00:02:54.680 --> 00:02:56.960 what some of the ethical questions are, what are some 52 00:02:57.039 --> 00:02:59.800 of the fears associated with the idea of thinking machines, 53 00:03:00.039 --> 00:03:03.919 and more with special guests agg Liu right here on 54 00:03:04.000 --> 00:03:06.199 today's Guys Guys Radio. 55 00:03:06.840 --> 00:03:08.439 It's Guys Guy Radio. 56 00:03:10.120 --> 00:03:13.159 Okay, Guys Guys Radio. It is the interview portion of 57 00:03:13.199 --> 00:03:16.319 our show, and today we're going to talk about AI. 58 00:03:16.599 --> 00:03:20.680 My special guest is agg Liu, author of the Handy 59 00:03:21.639 --> 00:03:25.520 Artificial Intelligence answer Book, a practical guide designed to help 60 00:03:25.680 --> 00:03:30.159 everyday people, professionals, entrepreneurs, and curious minds understand one of 61 00:03:30.199 --> 00:03:35.120 the most transformative technologies of our time, Artificial intelligence. You know, 62 00:03:35.199 --> 00:03:38.439 AI is no longer something reserved for scientists and tech giants. 63 00:03:38.400 --> 00:03:41.639 It's already influencing how we work, how we communicate, how 64 00:03:41.680 --> 00:03:44.599 we create, how we learn, how we shop, invest and 65 00:03:44.639 --> 00:03:48.439 make decisions. Yet many people remain uncertain about what AI 66 00:03:48.639 --> 00:03:51.840 actually is, how it works, and how it will affect 67 00:03:51.840 --> 00:03:53.919 their lives. And believe me, it will. So today we'll 68 00:03:53.960 --> 00:03:58.080 separate fact from fiction, explore the opportunities and risks of AI, 69 00:03:58.199 --> 00:04:02.039 and discuss how listeners can use the powerful technology wisely 70 00:04:02.120 --> 00:04:06.800 and effectively. Welcome to Guys Guys Radio A G. G Liu. 71 00:04:07.439 --> 00:04:09.560 Hello, Hello, good to be here. 72 00:04:09.840 --> 00:04:11.800 So I guess we should start at the beginning. You 73 00:04:11.879 --> 00:04:14.080 wrote this book with your writing. 74 00:04:13.800 --> 00:04:16.279 Partner, doctor Ashwari Powar. 75 00:04:16.439 --> 00:04:19.120 Yeah, here's here's the book. I'm in the middle of it, 76 00:04:19.160 --> 00:04:20.639 so I have a piece of favor there. I'll take 77 00:04:20.680 --> 00:04:23.560 it out. But it's fantastic. And the first thing I 78 00:04:23.600 --> 00:04:25.680 wondered when I was reading the book was I wonder 79 00:04:25.680 --> 00:04:28.920 if AI wrote this book for you. 80 00:04:28.959 --> 00:04:33.319 No, we did this all all manually, all manually. We 81 00:04:33.000 --> 00:04:36.160 we we really wanted the goal here to be something 82 00:04:36.199 --> 00:04:39.279 that would be accessible to people, right. And I think 83 00:04:39.800 --> 00:04:41.480 in this age of AI, where it's so easy to 84 00:04:41.480 --> 00:04:43.680 get AI to write something for you, the thing that's 85 00:04:43.759 --> 00:04:46.120 going to be more important is knowing exactly what to 86 00:04:46.160 --> 00:04:48.639 write knowing exactly what you can do that the I can't. 87 00:04:49.399 --> 00:04:53.920 And so when we created this, you know, the publisher 88 00:04:55.040 --> 00:04:57.839 with the publisher Visible in Press, they came to me 89 00:04:58.360 --> 00:05:00.240 and they said, hey, we have this great idea, want 90 00:05:00.279 --> 00:05:02.959 to write this book in this series about AI. And 91 00:05:03.399 --> 00:05:06.319 I created a draft, worked with doctor Powart to sort 92 00:05:06.319 --> 00:05:09.560 of turn it into the current format. And yeah, it's 93 00:05:09.600 --> 00:05:11.600 really meant to be for a general audience. It's meant 94 00:05:11.639 --> 00:05:14.639 to be for people, you know, like your listeners. 95 00:05:14.920 --> 00:05:17.480 You know, it's amazing. I've gone through the book. It's 96 00:05:17.519 --> 00:05:18.879 not one of these books that you have to read 97 00:05:18.920 --> 00:05:21.160 from page one to the end, but it's just a 98 00:05:21.160 --> 00:05:24.720 list of questions that are really good and has very clear, 99 00:05:24.839 --> 00:05:27.879 easy to digest answers. And I think anybody who has 100 00:05:27.920 --> 00:05:29.879 an interest in AI, and if you don't have an 101 00:05:29.879 --> 00:05:32.879 interest in AI, you basically don't have an interest in 102 00:05:32.920 --> 00:05:35.480 being part of the culture going forward. Because this is 103 00:05:35.600 --> 00:05:38.279 the it's here to stay. And the key is, as 104 00:05:38.319 --> 00:05:40.600 I tell my thirteen year old son, you know, don't 105 00:05:40.600 --> 00:05:42.680 worry about it taking your job. Learn how to use 106 00:05:42.720 --> 00:05:44.600 it because it's going to create so much more in 107 00:05:44.720 --> 00:05:47.199 terms of jobs. So nothing to be afraid of. But 108 00:05:47.360 --> 00:05:50.560 a couple of guardrails, of course, but concerns but not fear, 109 00:05:50.680 --> 00:05:53.000 and let's let's work with it. So for the benefit 110 00:05:53.040 --> 00:05:56.279 of everybody, because we hear, you know, about computing and Google, 111 00:05:56.360 --> 00:05:58.199 and then we hear about AI, and I think a 112 00:05:58.199 --> 00:06:00.600 lot of people are not sure how to make the 113 00:06:00.639 --> 00:06:04.360 differentiation between what's the difference between like googling something and 114 00:06:04.800 --> 00:06:08.600 asking a service like CHATGBT, what's the difference between AI 115 00:06:08.680 --> 00:06:12.079 and data processing computer typical computer stuff. 116 00:06:12.160 --> 00:06:14.920 Yeah, so AI came out of that data processing idea, 117 00:06:15.160 --> 00:06:17.959 but it works a little differently. So the AIS that 118 00:06:18.000 --> 00:06:20.800 we talk about today usually are something called a large 119 00:06:20.879 --> 00:06:24.000 language model. Large because it's got tons and tons of 120 00:06:24.000 --> 00:06:26.600 these numbers inside of it that it uses to make 121 00:06:26.639 --> 00:06:29.160 its processing. And it's a language model because it works 122 00:06:29.160 --> 00:06:31.759 with human language. So the difference between something like a 123 00:06:31.800 --> 00:06:34.240 standard program and a large language model is that when 124 00:06:34.279 --> 00:06:37.040 we write a computer program, even something as complicated as 125 00:06:37.120 --> 00:06:42.000 say Google Search was before it became AI related, every 126 00:06:42.360 --> 00:06:45.680 line of code in there was created by someone with 127 00:06:45.720 --> 00:06:49.079 a specific purpose. So you'd have a whole giant team 128 00:06:49.120 --> 00:06:52.199 of coders. They would work together to create this gigantic 129 00:06:52.240 --> 00:06:55.959 algorithm and everything, and they would know how each part works. 130 00:06:56.680 --> 00:06:59.439 With AI, it's a little bit different. With AI. What 131 00:06:59.480 --> 00:07:02.920 you do is create a machine that can learn, hence 132 00:07:02.959 --> 00:07:05.639 the term machine learning, which is often thrown around in 133 00:07:05.680 --> 00:07:09.319 the same sentences as AI, And basically the way it 134 00:07:09.360 --> 00:07:11.040 learns it's not quite the same as a human but 135 00:07:11.120 --> 00:07:13.600 it has all these sort of tweakable numbers in it, 136 00:07:14.360 --> 00:07:17.759 and you give it tons and tons and tons of examples, 137 00:07:17.800 --> 00:07:20.519 you know, most of the Internet at this point, and 138 00:07:22.079 --> 00:07:25.279 it learns how to write human language by looking at 139 00:07:25.279 --> 00:07:27.319 those examples and trying to predict what's the next word 140 00:07:27.360 --> 00:07:29.439 in a sentence or a paragraph or something like that. 141 00:07:30.120 --> 00:07:34.040 And so it doesn't work in a way where we 142 00:07:34.160 --> 00:07:36.279 know how each line of it works. In fact, there's 143 00:07:36.319 --> 00:07:38.720 a lot of effort right now to try and understand 144 00:07:38.720 --> 00:07:42.199 what goes on inside these inside their little programming circuits here. 145 00:07:43.000 --> 00:07:45.920 So when you ask something like Chad GBT, it's not 146 00:07:46.279 --> 00:07:48.959 going to be necessarily quite as reliable as something like 147 00:07:49.000 --> 00:07:53.519 a Google search or quite as organize the same way. 148 00:07:53.560 --> 00:07:55.160 But it's going to have different strengths. It can do 149 00:07:55.199 --> 00:07:57.600 some things that are hard to program explicitly, things that 150 00:07:57.600 --> 00:08:00.879 are hard to explain about, you know, creative making, poetry, 151 00:08:00.920 --> 00:08:01.439 things like that. 152 00:08:02.879 --> 00:08:06.639 Fantastic. So what's the biggest mixed misconception and that people 153 00:08:06.720 --> 00:08:09.800 have about AI right now? The general population? 154 00:08:10.560 --> 00:08:14.920 Ooh, the general population. I would say a lot of 155 00:08:14.959 --> 00:08:17.199 people don't know what AI is capable of right now. 156 00:08:18.399 --> 00:08:21.560 I think a lot of people if they've tried AI, 157 00:08:21.759 --> 00:08:24.800 they you know, used it when chat GPT first came out, 158 00:08:24.800 --> 00:08:27.040 and they think, Okay, that's you know, from twenty twenty two. 159 00:08:27.240 --> 00:08:30.920 That's exactly how it is now. But it's gotten a 160 00:08:30.959 --> 00:08:33.600 lot better and different in certain ways. 161 00:08:34.320 --> 00:08:34.519 You know. 162 00:08:34.559 --> 00:08:37.440 Now AI it can work with text, it can actually 163 00:08:37.559 --> 00:08:40.759 write things out for you, it can work with images, 164 00:08:40.840 --> 00:08:42.720 it can work with so many different sorts of things. 165 00:08:44.039 --> 00:08:46.879 And one of the big problems with AI that's been 166 00:08:46.919 --> 00:08:48.519 true for a very long time and it's still true, 167 00:08:48.600 --> 00:08:50.559 is that it's not one hundred percent reliable. You know, 168 00:08:50.600 --> 00:08:53.840 sometimes it'll make things up. So the companies call this 169 00:08:53.919 --> 00:08:57.159 the hallucination. I prefer the term confabulation myself. It's just 170 00:08:57.519 --> 00:09:01.000 basically shamelessly guessing an answer. But what you can do 171 00:09:01.080 --> 00:09:03.879 is you can ask the AI where it got its 172 00:09:03.879 --> 00:09:07.159 information from, and it'll be able to send you links 173 00:09:07.159 --> 00:09:10.480 to other places on the internet. And that means you 174 00:09:10.559 --> 00:09:13.720 as a person can back up what this AI is saying. 175 00:09:13.919 --> 00:09:17.000 So again it's about, like what you're saying, how you 176 00:09:17.120 --> 00:09:19.639 use AI learning, how to use it effectively to help 177 00:09:19.679 --> 00:09:20.960 you instead of to mislead you. 178 00:09:21.519 --> 00:09:25.879 It's two thoughts here, agg One is why we question 179 00:09:26.120 --> 00:09:29.080 why AI seem to have shown up all at once. 180 00:09:29.960 --> 00:09:32.879 It's been being used, but I think it's been kind 181 00:09:32.879 --> 00:09:38.320 of unveiled to the general public recently. But oh is 182 00:09:38.360 --> 00:09:39.000 that correct? 183 00:09:39.720 --> 00:09:42.759 Yeah, so really it hit the public scene all at once, 184 00:09:43.159 --> 00:09:45.320 sort of in twenty twenty two with the release of 185 00:09:45.399 --> 00:09:51.519 chat GBT. The original version basically the methods of computer science, 186 00:09:51.600 --> 00:09:55.759 the logic behind how we build these systems has existed 187 00:09:56.120 --> 00:09:59.799 in some form for decades. It sort of became a 188 00:10:00.200 --> 00:10:03.240 to in research field in the twenty tens or so, 189 00:10:05.519 --> 00:10:09.679 but it wasn't hooked up to enough computing power to 190 00:10:09.720 --> 00:10:12.200 actually make something that would be useful for the general 191 00:10:12.240 --> 00:10:14.639 public or interesting to the general public. It was just, 192 00:10:14.679 --> 00:10:16.320 you know, a research tool. It could do something like 193 00:10:16.360 --> 00:10:18.720 play the game of Go, for instance. It was an 194 00:10:18.720 --> 00:10:22.159 ancient Chinese game that was famously beaten in twenty sixteen. 195 00:10:23.279 --> 00:10:26.600 They beat the best human experts. So it's it's right 196 00:10:26.679 --> 00:10:30.720 when that chat GPT moment was the first time that 197 00:10:30.759 --> 00:10:34.200 these older algorithms were connected to enough computing power to matter. 198 00:10:34.559 --> 00:10:39.080 And once companies saw how big of a prize that was, 199 00:10:39.799 --> 00:10:42.240 that's when they started throwing tons and tons and tons 200 00:10:42.240 --> 00:10:44.799 of their computer power and their money to build new 201 00:10:44.799 --> 00:10:47.480 computers and everything like that in order to make this 202 00:10:48.039 --> 00:10:49.279 the behemoth that is today. 203 00:10:50.159 --> 00:10:53.720 And now we've got talking about behemoth. You have companies 204 00:10:53.759 --> 00:10:56.960 like Navidia who make the process and video, yeah, the video, 205 00:10:57.240 --> 00:10:59.480 the processing chips, and then you have all these data 206 00:10:59.480 --> 00:11:02.039 centers being built and people are like, what's going on here? 207 00:11:02.200 --> 00:11:05.600 But using AI uses a lot of energy, I guess, 208 00:11:05.679 --> 00:11:09.240 So could you explain that how that works so people understand. 209 00:11:09.120 --> 00:11:11.039 Yeah, of course, So this is something we talked about 210 00:11:11.039 --> 00:11:14.279 in the book. The deal with AI is that it 211 00:11:14.600 --> 00:11:17.759 uses lots and lots of computing power both to make 212 00:11:18.240 --> 00:11:20.080 the AI do what you want in the first place. 213 00:11:20.080 --> 00:11:22.000 So training the AI, that is going through all those 214 00:11:22.039 --> 00:11:25.399 examples and having it learn the patterns of human language 215 00:11:25.440 --> 00:11:28.919 and how to answer questions and things like that. And 216 00:11:29.759 --> 00:11:33.159 when you're actually running an AI, it's a more complicated 217 00:11:33.159 --> 00:11:36.480 computer operation than something like older Google Search would have been. 218 00:11:37.200 --> 00:11:40.879 So you actually do need a pretty powerful computer somewhere 219 00:11:40.879 --> 00:11:43.240 in the world to be running that. And I don't 220 00:11:43.279 --> 00:11:47.360 want to oversell this. It's still using less energy than say, 221 00:11:47.519 --> 00:11:49.759 running a lawnmower or something like that. It's not going 222 00:11:49.799 --> 00:11:54.080 to be anything that's by itself. Each individual AI search 223 00:11:54.120 --> 00:11:55.919 isn't going to be a big problem. But once you 224 00:11:56.000 --> 00:11:59.919 have millions or billions of people using this thing, then 225 00:12:00.039 --> 00:12:02.440 it multiplies and then all of a sudden it's a problem. 226 00:12:02.480 --> 00:12:04.759 Now the electrical grid has to worry about can it 227 00:12:04.840 --> 00:12:07.960 power all these data centers and everything else that comes 228 00:12:08.000 --> 00:12:08.519 along with that. 229 00:12:09.080 --> 00:12:12.039 Now, as an expert in the area, what do you think, 230 00:12:12.480 --> 00:12:14.600 what's the answer to that? Because people are concerned, as 231 00:12:14.679 --> 00:12:17.960 you've seen all these data centers, it's all about surveillance. 232 00:12:18.159 --> 00:12:20.039 We're going to suck all the power. The grids are 233 00:12:20.039 --> 00:12:22.720 old that we can't protect them, et cetera. What do 234 00:12:22.759 --> 00:12:24.679 you see for SEA as a solution to this. 235 00:12:25.159 --> 00:12:28.039 As a solution to this, I'd say that we've known 236 00:12:28.039 --> 00:12:30.360 for a very long time that the power grid needs upgrading. 237 00:12:30.519 --> 00:12:34.039 It's just something that you know, every kind of person 238 00:12:34.399 --> 00:12:37.279 involved in the actual system says this is something that 239 00:12:37.279 --> 00:12:41.639 we need to work on, and so this is just 240 00:12:41.759 --> 00:12:42.759 adding fuel to that fire. 241 00:12:42.960 --> 00:12:43.159 Right. 242 00:12:43.519 --> 00:12:46.600 The more these data centers they use power, they use 243 00:12:46.639 --> 00:12:49.519 power basically twenty four to seven compared to some of 244 00:12:49.519 --> 00:12:52.320 these other sorts of things. And so I really think 245 00:12:52.360 --> 00:12:54.240 that part that's going to be making the grid smarter, 246 00:12:54.639 --> 00:12:56.960 making it have better energy storage so that you can 247 00:12:57.360 --> 00:12:59.799 take energy that's coming in the daytime or the nighttime 248 00:12:59.840 --> 00:13:03.320 and smooth it out. That also helps because there's a 249 00:13:03.399 --> 00:13:07.159 very specific problem where data centers need power that's organized 250 00:13:07.240 --> 00:13:09.960 very cleanly. So you know that alternating current that's going 251 00:13:09.960 --> 00:13:14.480 through your wires. If that shape that that current is 252 00:13:14.480 --> 00:13:16.840 making as a little jagged, then the data center can't 253 00:13:16.840 --> 00:13:19.279 really use it very well. So there's a specific thing 254 00:13:19.320 --> 00:13:22.200 they'll do is if that power just loses that quality 255 00:13:22.200 --> 00:13:24.279 to it, they'll switch over to backup power. And then 256 00:13:24.279 --> 00:13:26.000 all of a sudden there's a bunch of extra energy 257 00:13:26.000 --> 00:13:27.799 on the grid, which is almost as big a problem, 258 00:13:28.840 --> 00:13:32.200 and it ends to be dumped somewhere. So basically all 259 00:13:32.240 --> 00:13:33.840 the same kinds of things people have been talking about 260 00:13:33.879 --> 00:13:35.799 for a while, you know, making sure we're having more 261 00:13:35.840 --> 00:13:38.279 efficient power plants and making sure that we have the 262 00:13:38.399 --> 00:13:41.080 energy storage. You know, solar is getting very cheap as 263 00:13:41.159 --> 00:13:44.639 much as it's kind of a meme and so so 264 00:13:44.679 --> 00:13:47.240 I think that's going to be the solution there. But 265 00:13:47.279 --> 00:13:49.360 then also there are some more technical ways to try 266 00:13:49.399 --> 00:13:51.519 and make AI more efficient on the chips that we have, 267 00:13:51.879 --> 00:13:52.720 which I could go into. 268 00:13:52.759 --> 00:13:55.679 Also, where do you see in terms of kind of 269 00:13:55.679 --> 00:13:59.639 the race for AI between the various big powers in 270 00:13:59.679 --> 00:14:02.200 countries on the Earth. Where do you see the US sitting? 271 00:14:02.320 --> 00:14:04.600 Are we in the lead? Are we number two? I mean, 272 00:14:04.679 --> 00:14:08.000 China obviously is highly evolved in the system. How about 273 00:14:08.000 --> 00:14:10.399 the other countries have they have they way behind? Or 274 00:14:10.440 --> 00:14:12.200 is it just the two of us or what do 275 00:14:12.240 --> 00:14:13.480 you see their agg. 276 00:14:13.559 --> 00:14:15.679 So it's pretty much just the US and China in 277 00:14:15.720 --> 00:14:19.000 the race at this point. There's a few European ones. 278 00:14:19.000 --> 00:14:21.919 There's some in the UK is involved in some I 279 00:14:21.919 --> 00:14:25.519 think there's a mysterl that's in I believe France. But 280 00:14:25.720 --> 00:14:28.200 really it's the American companies and the Chinese companies that 281 00:14:28.240 --> 00:14:30.120 are in the lead. But I also want to stress 282 00:14:30.159 --> 00:14:32.639 it's not just all of America are raising against all 283 00:14:32.679 --> 00:14:35.360 of China. It's a few American companies. You've got Google, 284 00:14:35.399 --> 00:14:38.360 You've got open Ai, you've got Anthropic, and they're raising 285 00:14:38.399 --> 00:14:40.559 each other as well as they're raising the Chinese companies. 286 00:14:40.600 --> 00:14:42.600 So you've got your q Win, you've got your Bidy 287 00:14:42.639 --> 00:14:46.120 and Ali Baba, all those guys. So it's not a 288 00:14:46.159 --> 00:14:48.279 monolithic thing in one or the other, I'd say that, 289 00:14:48.320 --> 00:14:51.879 But the US teams are, the US companies are in 290 00:14:51.919 --> 00:14:53.720 the lead right now. You know, the biggest news has 291 00:14:53.759 --> 00:14:56.159 been coming out of companies like Anthropic and Open Ai, 292 00:14:56.320 --> 00:14:59.480 especially with Claude Fable that's been in the news lately. 293 00:15:00.679 --> 00:15:05.000 So there's it's it's not an insurmountable lead. You know, 294 00:15:05.000 --> 00:15:06.840 there's a chance that depending on what we do, China 295 00:15:06.879 --> 00:15:09.000 could catch up. And I also want to stress that 296 00:15:09.399 --> 00:15:15.480 it's it's not a race in the sense that like 297 00:15:15.639 --> 00:15:18.639 if someone makes some AI system that's much better than 298 00:15:18.639 --> 00:15:20.519 everyone else's, we want it to be good, yes, but 299 00:15:20.559 --> 00:15:21.600 we also want it to be safe. 300 00:15:21.840 --> 00:15:24.840 M Okay, when you say safe, what do you mean. 301 00:15:26.200 --> 00:15:28.159 Well, there's a lot of dangers that come with AI. 302 00:15:28.879 --> 00:15:30.480 You know, again, there's a lot of positives, there's also 303 00:15:30.639 --> 00:15:32.399 some dangers that we have to watch out for. So 304 00:15:32.480 --> 00:15:36.919 one thing that's recently been in the news is cybersecurity. 305 00:15:37.440 --> 00:15:41.360 So using the models that are at the cutting edge 306 00:15:41.360 --> 00:15:46.559 of the AI systems to find exploits in the software 307 00:15:46.600 --> 00:15:50.639 that runs banks, or that runs important social media sites, 308 00:15:50.720 --> 00:15:53.720 or anything that has your data in it. And one 309 00:15:53.720 --> 00:15:55.840 of the issues that people are finding is that these 310 00:15:55.879 --> 00:15:58.559 ais are so good at finding bugs in this code 311 00:15:59.159 --> 00:16:01.799 that all of a sudden, the number of bugs that 312 00:16:01.840 --> 00:16:04.840 are known shoots up. So the responsible thing to do, 313 00:16:04.879 --> 00:16:07.559 which is thankfully what has been started to set it 314 00:16:07.559 --> 00:16:11.279 as a trend, is to give these models to the defenders. First, 315 00:16:11.399 --> 00:16:13.039 give it to the software companies, give it to the 316 00:16:13.080 --> 00:16:16.279 banks before you release to the general public. That way 317 00:16:16.320 --> 00:16:18.919 they can patch all the exploits that it defines. So 318 00:16:18.960 --> 00:16:21.720 cybersecurity is one issue. Another issue people talk about is 319 00:16:22.279 --> 00:16:24.639 synthetic biology is a little more speculative. But you know, 320 00:16:24.679 --> 00:16:30.200 maybe AIS could make people's work in biolabs more dangerous 321 00:16:30.200 --> 00:16:32.080 from people with less skill, could make something like a 322 00:16:32.159 --> 00:16:37.320 dangerous kind of germ. But then also in principle, if 323 00:16:37.320 --> 00:16:40.480 an AI gets smart enough, if it really wants something 324 00:16:40.519 --> 00:16:42.919 that's bad for humans, you know, something like turning the 325 00:16:42.960 --> 00:16:46.000 whole world into more computers, then we would have a 326 00:16:46.039 --> 00:16:46.559 problem with. 327 00:16:46.559 --> 00:16:50.480 That, right of course. Okay, yeah, my special guest on 328 00:16:50.759 --> 00:16:56.440 Guys Guys Radio Today, agg Liu and his Agg's partner Ashwari. 329 00:16:56.759 --> 00:17:00.279 How are PhD put together this wonderful book and it's 330 00:17:00.279 --> 00:17:03.919 called The Handy Artificial Intelligence answer Book. You folks did 331 00:17:03.960 --> 00:17:07.039 a great job with it, reily fantastic and I'm reading 332 00:17:07.119 --> 00:17:08.920 it at the same time as my thirteen year old 333 00:17:09.000 --> 00:17:11.000 son and we just you can just open up a 334 00:17:11.079 --> 00:17:13.160 question or ask a question, look it up and the 335 00:17:13.200 --> 00:17:16.039 answer is there. It's just it's really fantastic. And I 336 00:17:16.200 --> 00:17:18.240 urge everybody to get a copy of this because you 337 00:17:18.359 --> 00:17:20.400 really want to learn about this. This is going to 338 00:17:20.440 --> 00:17:23.119 be part of our lives and it's up to us 339 00:17:23.160 --> 00:17:27.839 how we use it. You mentioned AGG finding little holes 340 00:17:27.880 --> 00:17:31.240 and things, and I was reading about medicine how artificial 341 00:17:31.240 --> 00:17:35.480 intelligence can help medicine where maybe they read a scan 342 00:17:35.640 --> 00:17:37.839 or something and they can pick things out ahead of 343 00:17:37.839 --> 00:17:41.200 time that traditionally that the physicians may not be able 344 00:17:41.240 --> 00:17:42.839 to see. Talk to us a little bit about that, 345 00:17:42.880 --> 00:17:45.319 because that's a really good thing, and is there an 346 00:17:45.400 --> 00:17:47.839 issue with that that maybe because sometimes AI is wrong. 347 00:17:48.599 --> 00:17:50.599 So yeah, So I'd say medisine is one of the 348 00:17:50.599 --> 00:17:52.960 areas people are most optimistic about when it comes to AI. 349 00:17:53.119 --> 00:17:55.000 This is one of the areas where we think it's 350 00:17:55.039 --> 00:17:57.160 going to be a great tool. It's already starting to 351 00:17:57.160 --> 00:18:01.160 be a great tool because is an area where your 352 00:18:01.160 --> 00:18:03.920 body is so complicated. There's so many little systems. There's 353 00:18:03.960 --> 00:18:07.400 so much that could be changed by something like a 354 00:18:07.440 --> 00:18:09.240 medicine or something by a treatment like a surgery or 355 00:18:09.319 --> 00:18:13.359 something like that. And having a machine like an AI 356 00:18:13.480 --> 00:18:16.440 that's very good at finding patterns and data that's sort 357 00:18:16.440 --> 00:18:20.720 of its fundamental job, makes it a lot easier in 358 00:18:20.759 --> 00:18:24.759 principle at least to diagnose things. Like they could work 359 00:18:24.920 --> 00:18:28.759 with a radiologist looking at an X ray and see, okay, 360 00:18:29.000 --> 00:18:33.039 maybe I can flag little points on the X ray 361 00:18:33.079 --> 00:18:35.039 image that the radiologists might want to look at further. 362 00:18:35.960 --> 00:18:39.680 Or it could help look at what your specific symptoms are, 363 00:18:39.759 --> 00:18:42.759 what your specific test results are, and find is there 364 00:18:42.759 --> 00:18:45.599 a drug out there in the world already that can 365 00:18:45.720 --> 00:18:48.839 treat your disease maybe better than it would be able 366 00:18:48.839 --> 00:18:50.680 to treat someone else with a vaguely similar but not 367 00:18:50.759 --> 00:18:54.400 quite the same disease. Already in clinical trials, you know, 368 00:18:54.440 --> 00:18:56.839 people are testing medicines that AI has come up with. 369 00:18:57.640 --> 00:19:00.480 The AI is searched through all the possible goals and 370 00:19:01.039 --> 00:19:04.240 to see what might actually be helpful to our bodies 371 00:19:04.279 --> 00:19:05.079 that have hurtful to it. 372 00:19:05.759 --> 00:19:08.920 At that same on the same note, a g G 373 00:19:10.279 --> 00:19:16.359 is a AI being used to find holistic cures, if 374 00:19:16.400 --> 00:19:20.440 you will, for maladies that people have. Because I was 375 00:19:20.480 --> 00:19:23.480 watching the news this morning and one of the guys, 376 00:19:23.480 --> 00:19:26.039 some science or you know, medical expert, came on and 377 00:19:26.160 --> 00:19:29.519 was touting AI developed vaccines. Now, depending on how you 378 00:19:29.519 --> 00:19:32.759 feel about vaccines, some people yeah, no way and uh 379 00:19:32.839 --> 00:19:34.720 and now they're going to use this against us, and 380 00:19:34.799 --> 00:19:37.720 other people like, oh, that's great. What's your take on 381 00:19:38.160 --> 00:19:40.880 AI being I guess a lot of it depends on 382 00:19:41.079 --> 00:19:46.160 who's behind challenging AI to come up with stuff, and 383 00:19:46.240 --> 00:19:48.599 so I guess the question is how's that being done? 384 00:19:48.839 --> 00:19:51.319 And are they looking for natural stuff as well as 385 00:19:51.720 --> 00:19:53.640 you know, farmer based solutions When it. 386 00:19:53.559 --> 00:19:55.880 Comes to of course, of course, you know, I'd say 387 00:19:55.880 --> 00:19:59.200 there's there's nothing saying AI has to do one exact 388 00:19:59.200 --> 00:20:02.359 approach to medicine. You know, there's nothing in the systems 389 00:20:02.400 --> 00:20:04.920 that are saying it has to follow this farming based approach. 390 00:20:04.960 --> 00:20:07.960 The way you're talking about, you know, people are talking 391 00:20:07.960 --> 00:20:09.839 about using it to help people's mental health, to help 392 00:20:09.839 --> 00:20:14.000 people's wellness in their exercise programs or in their general lifestyles, 393 00:20:14.079 --> 00:20:15.839 or you could even look at it, you know, to 394 00:20:15.880 --> 00:20:18.400 find remedies in nature that kind of stuff. You know, 395 00:20:18.559 --> 00:20:21.480 are there plants out there or are their animals or 396 00:20:21.519 --> 00:20:24.680 other sorts of you know, mushrooms I don't know that 397 00:20:24.799 --> 00:20:28.480 have have these these useful properties. 398 00:20:28.000 --> 00:20:28.200 You know. 399 00:20:28.440 --> 00:20:30.839 And so again, AI is a very good pattern matching tool. 400 00:20:31.160 --> 00:20:34.119 It can find patterns and data, it can find information 401 00:20:34.279 --> 00:20:39.079