WEBVTT - Will Siri Ever Outsmart Us? 0:00:00.120 --> 0:00:03.400 We're going to start our story today. In nineteen sixty eight, 0:00:03.640 --> 0:00:07.640 way before I was born and probably before you were born. Alistair. 0:00:07.880 --> 0:00:13.360 Very funny, Yes, I was born after so before either 0:00:13.440 --> 0:00:16.640 of us were born. Terry Winnigrad was a young grad 0:00:16.760 --> 0:00:20.840 student in Massachusetts, and Terry had just started at m 0:00:20.880 --> 0:00:26.720 I t S brand new artificial intelligence lab. Basically, the 0:00:26.760 --> 0:00:30.080 culture was, look, nobody has ever tried doing all this 0:00:30.160 --> 0:00:33.279 kind of stuff with computers. They've done business calculations or whatever, 0:00:33.360 --> 0:00:36.040 but nobody's trying to get them to see things, or 0:00:36.159 --> 0:00:40.479 to move a robot arm or to do language. And 0:00:40.880 --> 0:00:46.320 therefore we're on the exploring cutting edge, and we're going 0:00:46.360 --> 0:00:50.360 to solve all these problems right soon. And Terry started 0:00:50.400 --> 0:00:54.280 building a computer program that turned into his dissertation, which 0:00:54.280 --> 0:00:58.160 he called shirt looke Yeah, with only one vowel, s 0:00:58.360 --> 0:01:01.880 h r d l U. We're watching a video here 0:01:01.960 --> 0:01:04.960 of a demo of shurtloo that Terry had recorded at 0:01:04.959 --> 0:01:08.959 the time. It's this silent, black and white, low rez, 0:01:09.080 --> 0:01:12.440 grainy video, and you see a virtual table top and 0:01:12.520 --> 0:01:15.360 a bunch of geometric shapes on top. It looks like 0:01:15.400 --> 0:01:19.920 a computer sketch of a really really boring form of Tetris. Yeah, 0:01:20.000 --> 0:01:23.679 chat bots are everywhere today, but Shirtlou was really one 0:01:23.720 --> 0:01:27.440 of the original chat bots in history. Using this machine 0:01:27.480 --> 0:01:30.360 called a teletype that's like a typewriter on a tray 0:01:30.400 --> 0:01:33.240 hooked up to a gigantic computer the size of a 0:01:33.319 --> 0:01:38.040 small bathroom, Terry created this interactive assistant that could help 0:01:38.080 --> 0:01:41.279 you navigate the virtual world of blocks on a table. 0:01:42.120 --> 0:01:44.560 So you could type in something like pick up the 0:01:44.600 --> 0:01:48.000 red block and stack it on top of the blue cube, 0:01:48.120 --> 0:01:51.920 and like magic, this virtual robotic arm would appear and 0:01:51.960 --> 0:01:54.560 do that stacking for you. And you could ask your 0:01:54.640 --> 0:01:58.360 lue questions too, like which cube is sitting on the table. 0:01:58.680 --> 0:02:01.240 And on the display screen you see this text show up, 0:02:01.360 --> 0:02:05.240 letter by letter, replying to your prompt. And the amazing 0:02:05.320 --> 0:02:08.840 thing about Surtly was the people into acting with it 0:02:09.200 --> 0:02:12.400 would use normal English. You didn't have to click on 0:02:12.400 --> 0:02:14.800 a bunch of buttons, you didn't have to throw in 0:02:14.880 --> 0:02:17.240 a string of numbers, and you didn't have to use 0:02:17.280 --> 0:02:21.640 any obscure computer programming language. Yeah, watching this video of 0:02:21.680 --> 0:02:24.639 the program, now it feels like you're watching two people 0:02:24.760 --> 0:02:28.080 texting each other, even though it's a person talking to 0:02:28.160 --> 0:02:31.680 a computer. And so if you and I are impressed 0:02:31.720 --> 0:02:34.840 watching this thing in action today. You can imagine how 0:02:34.880 --> 0:02:38.240 stunned people were when they saw this half a century ago. 0:02:38.760 --> 0:02:41.680 And from there the field of AI was supposed to 0:02:41.680 --> 0:02:44.880 take off like a rocket ship. Terry hunker down to 0:02:44.960 --> 0:02:47.960 bring Shortlood to life, and he wanted it to work 0:02:47.960 --> 0:02:50.960 in a universe bigger and more complicated than just a 0:02:50.960 --> 0:02:54.200 couple blocks on a table. But the deeper he got 0:02:54.200 --> 0:02:57.480 into it, he was running into more and more obstacles. 0:02:57.840 --> 0:03:01.040 And after a while he gave up on Shurtle, and 0:03:01.080 --> 0:03:04.160 he gave up on AI. He ended up leaving the 0:03:04.200 --> 0:03:14.239 field all together. Hi. This is Aki Ito and I'm 0:03:14.240 --> 0:03:17.560 Alista bar And this week on Decrypted, we're taking you 0:03:17.600 --> 0:03:21.080 to Stanford to meet one of the most influential thinkers 0:03:21.120 --> 0:03:24.680 in the history of computing. This is an early pioneer 0:03:24.720 --> 0:03:26.960 in the field of AI who built one of the 0:03:26.960 --> 0:03:31.920 most impressive reasoning machines and ultimately concluded that computers wouldn't 0:03:31.919 --> 0:03:34.680 be able to match human intelligence in his lifetime. And 0:03:34.720 --> 0:03:38.360 when he came to that conclusion, Terry Winnigrad dedicated the 0:03:38.440 --> 0:03:42.720 rest of his career to improving computers not as a 0:03:42.840 --> 0:03:45.480 replacement for human thought, but as a tool to help 0:03:45.520 --> 0:03:49.880 all of us. Terry's fingerprints are everywhere in the technologies 0:03:49.960 --> 0:03:53.800 powering modern life today, including Google Search, which started out 0:03:53.840 --> 0:03:57.720 as Larry Page's grad school research project, and this was 0:03:57.760 --> 0:04:02.320 something that Terry supervised. Now, with all these digital assistance 0:04:02.720 --> 0:04:05.560 powering our everyday lives, we're going to have Terry, the 0:04:05.680 --> 0:04:09.440 very creator of their precursor, test them all out. Yeah, 0:04:09.520 --> 0:04:13.480 consider him the great grandfather of Syrie or Amazon's Alexa. 0:04:13.920 --> 0:04:17.080 You might be surprised by Terry's conclusion on these devices 0:04:17.440 --> 0:04:19.560 on how far he thinks the field has come since 0:04:19.560 --> 0:04:22.280 he unveiled Shirtly to the world almost fifty years ago. 0:04:22.960 --> 0:04:25.640 And he'll give us his thoughts on where these devices 0:04:25.640 --> 0:04:30.000 are headed too. Don't worry, I have a long way 0:04:30.040 --> 0:04:37.279 to go before I become smarter than you. Okay, so 0:04:37.400 --> 0:04:40.880 let's rewind way back to a time before the days 0:04:40.920 --> 0:04:48.240 of personal computers. It's Elvis Presley is on the radio. 0:04:48.839 --> 0:04:54.240 Russia had launched Sputnik only three years ago. Russia I 0:04:54.279 --> 0:04:56.960 had blasted a man made moon into and it would 0:04:56.960 --> 0:04:59.120 be nine years until the US put a man on 0:04:59.160 --> 0:05:03.280 the moon. Terry, meanwhile, is in high school in Greely, Colorado. 0:05:03.680 --> 0:05:06.360 The physics teacher said, you know, you don't really need 0:05:06.400 --> 0:05:09.880 to sit through these lectures in my physics class. I'll 0:05:09.880 --> 0:05:11.720 give you a workshop up in the attic. It was 0:05:11.720 --> 0:05:13.400 in the attic of the school, and you wanted to 0:05:13.400 --> 0:05:17.159 build something interesting. And this is when Terry first encounter 0:05:17.200 --> 0:05:20.680 as a computer. And my father owned a what was 0:05:20.720 --> 0:05:23.000 a steel business, but had grown up from being a 0:05:23.080 --> 0:05:27.200 junkyard and had a huge collection of old junk stuff. 0:05:27.240 --> 0:05:30.000 They picked up a governments plus sales and so I 0:05:30.040 --> 0:05:33.720 took spare parts from the junkyard and I built a 0:05:33.760 --> 0:05:37.320 little computer. A box looked like a bread box. Uh. 0:05:37.360 --> 0:05:39.039 In fact, the case I put it in may have 0:05:39.120 --> 0:05:42.760 actually been a bread box. Uh. And it did a 0:05:42.880 --> 0:05:48.120 very simplistic computation, but it was it worked. He attended 0:05:48.160 --> 0:05:51.560 Colorado College, a liberal arts university, where he was a 0:05:51.600 --> 0:05:54.600 math major. But what he really fell in love with 0:05:54.640 --> 0:05:58.599 there was linguistics, the science of how language it's structured 0:05:58.600 --> 0:06:01.320 and understood. And he spent a year in London after 0:06:01.360 --> 0:06:05.200 he graduated, studying linguistics too. And after that, in the 0:06:05.279 --> 0:06:09.280 late nineteen sixties, Terry started his PhD at m I. T. 0:06:09.520 --> 0:06:13.919 S AI Lab, led by Marvin Minsky himself. Minsky is 0:06:13.960 --> 0:06:17.520 one of the godfathers of AI. Everyone felt the promise 0:06:17.680 --> 0:06:21.200 and it was you know, they say that people doing 0:06:21.200 --> 0:06:23.159 it were like undergraduates. I wasn't. I was an old 0:06:23.200 --> 0:06:25.520 person right in that group. So there was a sort 0:06:25.560 --> 0:06:27.800 of youthful sense of where the new generation we're going 0:06:27.839 --> 0:06:30.200 to make things happen. So Terry got to work on 0:06:30.240 --> 0:06:34.680 surely the Intelligent Seeming software that we introduced earlier. It 0:06:34.720 --> 0:06:38.040 was grueling work piecing together different parts of of software, 0:06:38.400 --> 0:06:42.279 but it gradually came together. I think one of the 0:06:42.279 --> 0:06:45.560 most striking things about the program, in addition to this 0:06:45.720 --> 0:06:47.800 direct visual you can see what it was doing you 0:06:47.839 --> 0:06:51.159 could talk about JOIN, is that I attempted to deal 0:06:51.240 --> 0:06:55.039 with some of the interesting properties of language in that 0:06:55.120 --> 0:06:58.880 you don't say everything that's explicitly so I used to 0:06:58.920 --> 0:07:00.960 I mean, take the most obviously ample there, put it 0:07:01.000 --> 0:07:04.880 next to a red one. A read what what is 0:07:04.920 --> 0:07:07.200 a one? And to know what you mean, you of 0:07:07.240 --> 0:07:08.960 course have to go back into the context. You have 0:07:09.000 --> 0:07:12.680 to know that previously you said find a blue block, 0:07:12.800 --> 0:07:14.720 put it next to a red one. So now you're 0:07:14.720 --> 0:07:16.680 gonna go back a sentence and find that you meant block, 0:07:17.400 --> 0:07:20.200 or you could say pick up another one, What does 0:07:20.240 --> 0:07:23.920 another mean the whole and pronounce put it? If I say, 0:07:24.080 --> 0:07:27.000 now put it in the box. It has to figure 0:07:27.000 --> 0:07:29.760 out which of the things from the previous world you 0:07:29.800 --> 0:07:33.520 meant by it. So it had a very natural flow 0:07:33.600 --> 0:07:37.240 to it. When Terry revealed short lude to the World 0:07:37.320 --> 0:07:41.640 in two in the form of an entire issue of 0:07:41.680 --> 0:07:46.560 the Journal of Cognitive Psychology, the world was amazed. We're 0:07:46.560 --> 0:07:48.000 starting here at zero. We don't know what we can do. 0:07:48.040 --> 0:07:50.080 When you make a big first step, you say, hey, 0:07:50.960 --> 0:07:52.720 I'm gonna keep going up right, I mean, and so 0:07:53.080 --> 0:07:54.640 I think that the fact they could do as much 0:07:54.680 --> 0:07:57.880 as it did certainly gave people like marvit Minsk a 0:07:57.920 --> 0:08:00.400 lot of confidence. I mean, he was I think a 0:08:00.400 --> 0:08:06.880 bigger booster of my program than I was. The confidence 0:08:07.000 --> 0:08:10.920 Terry's referring to here. That's confidence in the progress scientists 0:08:10.960 --> 0:08:13.760 were making to develop computers that were just as smart 0:08:13.800 --> 0:08:17.960 as humans. Terry's breakthrough inspired a lot of smart people 0:08:18.000 --> 0:08:21.040 to believe that sentient computers were just around the corner, 0:08:21.520 --> 0:08:24.160 and so shortly after Terry left m I t for 0:08:24.240 --> 0:08:28.800 the warmer pastures of California, back when Silicon Value was 0:08:28.840 --> 0:08:32.320 a quiet place full of fruit orchards, ah so nice, 0:08:32.440 --> 0:08:36.160 no traffic. And as a professor at Stanford and a 0:08:36.240 --> 0:08:40.880 researcher at the legendary research labs rox Park, he worked 0:08:40.880 --> 0:08:43.559 on trying to expand Shortloo. He was trying to get 0:08:43.600 --> 0:08:46.640 it to work in a more complicated environment than just 0:08:46.720 --> 0:08:49.160 a couple of geometric shapes on top of the table. 0:08:49.720 --> 0:08:53.439 And then Shirtloo took over the world and destroyed mankind 0:08:55.800 --> 0:08:59.200 or not, things weren't going the way Terry had hoped. 0:08:59.240 --> 0:09:03.160 The attempt we were making at in that project was 0:09:03.240 --> 0:09:09.440 to come to a broader analysis of meaning which could 0:09:09.480 --> 0:09:13.120 handle the ways in which meaning is much vaguer and 0:09:13.520 --> 0:09:17.319 less systematic than it was for the blocks world. And 0:09:17.559 --> 0:09:19.640 in my talks about this, I always use blocks as 0:09:19.640 --> 0:09:21.680 an example because it's a really simple one, which is 0:09:22.160 --> 0:09:24.800 in that world, block meant exactly one thing. It meant 0:09:24.840 --> 0:09:29.880 a rectangular shaped object of certain sizes on. But if 0:09:29.880 --> 0:09:32.600 I say to you, let's walk around the block, it 0:09:32.679 --> 0:09:35.120 has a second meaning which is different. So you say, okay, 0:09:35.120 --> 0:09:36.480 so you have to put in two meetings. But then 0:09:36.520 --> 0:09:38.480 I say something like that, Well, you know, I'm trying 0:09:38.480 --> 0:09:39.840 to write his paper, but there's some kind of a 0:09:39.840 --> 0:09:42.200 block I can't get over now, it's not even a 0:09:42.200 --> 0:09:49.000 physical object. It's a metaphorical physical objects. So language really 0:09:49.160 --> 0:09:52.800 works that way. Very little of it has precisely defined 0:09:52.840 --> 0:09:57.239 meetings outside of technical stuff. Uh, and we're always extending 0:09:57.240 --> 0:10:01.160 meetings and using implicit metaphor. It's not fancy metaphors, you know, 0:10:01.240 --> 0:10:03.960 life is a rose bowl of roses or whatever. But 0:10:04.360 --> 0:10:08.200 just like a block in my mental thinking, um, and 0:10:08.800 --> 0:10:12.319 trying to write the logic, the algorithms, the underlying computer 0:10:12.360 --> 0:10:16.400 stuff which could handle that is a totally different problem 0:10:16.600 --> 0:10:19.760 from just handling the simple logical stuff. And we tackled 0:10:19.760 --> 0:10:25.000 that problem. And in hindsight, I would say I was 0:10:25.040 --> 0:10:28.640 aware we weren't getting very far. And while Terry's doubts 0:10:28.640 --> 0:10:31.360 were growing, he also started hanging out with a few 0:10:31.400 --> 0:10:35.319 academics in the Bay Area. There were philosophers Hubert Dreyfus 0:10:35.440 --> 0:10:39.120 and John Searle, and there was this Chilean engineer, entrepreneur 0:10:39.200 --> 0:10:43.520 and politician called Fernando Flores. They were all making the 0:10:43.720 --> 0:10:47.560 same broad point around this time, which was this the 0:10:47.640 --> 0:10:51.120 way our brains work. So much of it happens without 0:10:51.200 --> 0:10:54.560 us explicitly thinking about it in a logical way, like 0:10:55.120 --> 0:10:58.680 if A then B and if B, then C. It's 0:10:58.720 --> 0:11:02.720 this complete black box to us. So maybe it wasn't 0:11:02.800 --> 0:11:05.080 ever going to be possible to get a machine to 0:11:05.160 --> 0:11:07.439 do all the things that are human brains can do. 0:11:08.720 --> 0:11:12.080 This philosophy resonated more and more with Terry, and in 0:11:12.080 --> 0:11:15.880 the meantime, the entire field of AI, who's having this 0:11:16.000 --> 0:11:20.000 moment of reckoning, as anybody does when they're doing a 0:11:20.040 --> 0:11:23.000 new technology and they need to get grants, they will 0:11:23.040 --> 0:11:25.000 say it's going to solve all the problems in the world, right, 0:11:25.040 --> 0:11:28.120 best things since life bread, and then it filters out. 0:11:28.200 --> 0:11:30.440 And so what happened is that a I hit this 0:11:30.559 --> 0:11:34.640 point where the claims had overreached. The results were not 0:11:35.120 --> 0:11:38.880 that great. There were good in certain small technical areas. 0:11:38.880 --> 0:11:41.440 It wasn't there was no results, but nothing on the 0:11:41.480 --> 0:11:45.760 scale of what people were promising, and that ushered in 0:11:46.080 --> 0:11:51.160 what became known as the AI winter. Research projects got defunded, 0:11:51.440 --> 0:11:56.080 startups died, All this excitement withered away. So by the 0:11:56.160 --> 0:11:59.920 nineteen eighties Terry was pretty much convinced that he'd reach 0:12:00.080 --> 0:12:02.880 the dead end, which I don't know you think would 0:12:02.920 --> 0:12:06.840 be this devastating realization for him. But right around this 0:12:06.920 --> 0:12:13.760 time something else came along up until approximately three. If 0:12:13.800 --> 0:12:16.200 you use the computer, you were a technical nerd in 0:12:16.240 --> 0:12:19.080 the basement somewhere, and the fact you had to learn 0:12:19.120 --> 0:12:23.080 all sorts of arcane stuff was yeah, that's what we do. Uh. 0:12:23.120 --> 0:12:25.959 And then there was the computer for the rest of us. 0:12:26.920 --> 0:12:30.280 We can draw a picture or it can draw conclusions. 0:12:31.080 --> 0:12:33.720 It's a personal computer from Apple up and it's as 0:12:33.720 --> 0:12:39.160 easy to use as this Macintosh, the computer for the 0:12:39.240 --> 0:12:42.360 rest of us. If you're a tech you probably have 0:12:42.400 --> 0:12:44.839 seen that ad at some point, right and Apple came 0:12:44.840 --> 0:12:47.360 out with mac and all of a sudden you had 0:12:47.400 --> 0:12:49.480 all these people who wanted to use computers who were 0:12:49.520 --> 0:12:51.719 not tech nerds, and we're not willing to learn all 0:12:51.760 --> 0:12:54.319 the arcane stuff, and so the whole field of how 0:12:54.320 --> 0:12:56.839 do you make them something that ordinary people can deal 0:12:56.880 --> 0:13:02.480 with blossom. So Terry decided not to obsess over building 0:13:02.520 --> 0:13:05.120 computers that were going to truly think, and he wasn't 0:13:05.120 --> 0:13:08.040 going to worry about understanding how the brain really works either. 0:13:08.760 --> 0:13:11.679 John Lily, a former student of Terry's who's a partner 0:13:11.760 --> 0:13:16.079 at the venture capital firm Graylock Capital, described Terry's vision 0:13:16.160 --> 0:13:19.960 like this, machines can't do everything, but focusing on what 0:13:20.000 --> 0:13:24.080 the machine capabild years is always a mistakes. He really 0:13:24.080 --> 0:13:27.360 want to pucus on the whole system, which includes humans, 0:13:26.800 --> 0:13:31.360 and they include humans and machines in the interviews between 0:13:31.400 --> 0:13:35.240 two machines is the key, and that philosophy helps shape 0:13:35.280 --> 0:13:38.079 the course of what was at the time a budding 0:13:38.120 --> 0:13:42.400 discipline called Human computer interaction h c I. Is all 0:13:42.440 --> 0:13:47.200 about designing tools to help us humans use computers more easily. Yeah, 0:13:47.280 --> 0:13:50.720 taking these clunky machines out of research labs and putting 0:13:50.720 --> 0:13:53.920 them into the hands of you and me in Silicon 0:13:54.040 --> 0:13:57.880 Valley today, you hear this acronym h CI all the time. 0:13:58.280 --> 0:14:01.719 That's a very very strong view. The printed on me 0:14:02.000 --> 0:14:07.640 for sure, like everybody looking out program massa others. Okay, 0:14:07.679 --> 0:14:11.600 so that's Marissa as in Marissa Meyer, Google employee number 0:14:11.600 --> 0:14:15.400 twenty and the CEO of Yahoo and read as in 0:14:15.520 --> 0:14:19.040 Reed Hoffman, a founding director of PayPal and the creator 0:14:19.160 --> 0:14:23.360 of LinkedIn. But the most famous of Terry students are 0:14:23.400 --> 0:14:27.640 Google co founders Larry Page and Sergey Bryn. They met 0:14:27.680 --> 0:14:31.800 as graduate students in They started working on a way 0:14:31.800 --> 0:14:34.560 to organize all the different web pages out there, and 0:14:34.680 --> 0:14:38.800 Terry supervised their project that turned into the foundations of 0:14:38.880 --> 0:14:42.040 Google Search. I mean, something that helps us access the 0:14:42.160 --> 0:14:46.240 trillions of web pages of content. I can't really think 0:14:46.280 --> 0:14:49.200 of a single digital tool that's been more useful in 0:14:49.280 --> 0:14:52.920 modern life. Thanks to Terry's guidance, Google was born as 0:14:52.960 --> 0:14:57.800 a company on September four and it's now the world's 0:14:57.800 --> 0:15:02.160 second most valuable company. Pretty cool stuff. Hey, I can 0:15:02.240 --> 0:15:05.760 calm down, calm down. We should point out that Terry 0:15:05.920 --> 0:15:09.720 wasn't always this business genius, passing down sage advice to 0:15:09.760 --> 0:15:12.880 his students. And I would say to Larry occasionally, well, yeah, 0:15:12.920 --> 0:15:14.680 that's great, but how you're going to make money with this? 0:15:15.480 --> 0:15:17.720 I think what you should do is my advice is, 0:15:17.760 --> 0:15:19.880 you know, find a company like Microsoft or somebody who 0:15:19.880 --> 0:15:22.040 needs a search engine and sell it for a nice 0:15:22.080 --> 0:15:25.760 chuck of change. I always say, they're fortunate they took 0:15:25.800 --> 0:15:30.720 my technical advice, but not my business advice. So let's 0:15:30.800 --> 0:15:34.880 fast forward about two decades to today, Larry Page. It's 0:15:34.960 --> 0:15:37.320 still at the helm of Google, and one of the 0:15:37.360 --> 0:15:42.160 company's biggest bets is its digital assistant. It's called Google Assistant, 0:15:42.640 --> 0:15:46.440 and it's connected to its new smartphone called Pixel, and 0:15:46.480 --> 0:15:49.240 it's also connected to its home device, which is like 0:15:49.640 --> 0:15:52.920 this portable speaker that speaks back to you, and that 0:15:52.960 --> 0:15:56.440 follows in the footsteps of all the other digital assistance 0:15:56.480 --> 0:15:59.360 out there. First of all, they're Sirie, which is the 0:15:59.400 --> 0:16:04.000 assistant on the iPhone. Hello therecky. There's the Amazon Echo, 0:16:04.080 --> 0:16:08.160 which is called Alexa. Hello, I'm here, And then there's 0:16:08.240 --> 0:16:12.200 Microsoft's assistant, which is called Cortana. Hey there, my friend. 0:16:12.640 --> 0:16:15.160 We wanted to know if these new helpers are useful 0:16:15.200 --> 0:16:18.360 and smart, so you better to quiz them than Terry. 0:16:18.840 --> 0:16:21.840 Along with our editor Emily A. Busso we started sending 0:16:21.920 --> 0:16:27.520 up these devices on Terry's desk in his office. Hey, Alexa, 0:16:27.600 --> 0:16:34.520 are you on Hello I'm here? Okay, great, Alexa's working yea. 0:16:34.600 --> 0:16:38.880 So we have we have Amazon Echo Alexa. These are 0:16:38.920 --> 0:16:44.280 all on Professor Winograds desk in his office. Alexa has 0:16:44.320 --> 0:16:48.080 just turned herself on you listening to me? Okay. So 0:16:48.240 --> 0:16:51.920 here's the first question. This is Terry asking Siri, where's 0:16:51.960 --> 0:16:58.200 a nightclub that my methodist uncle would enjoy? Okay, check 0:16:58.240 --> 0:17:02.720 it out? What is it? Show? Show some random nightclubs? Now, 0:17:02.720 --> 0:17:05.520 I have no idea if they have any you know, 0:17:06.880 --> 0:17:12.720 Holy holy Holy Town nightclub the grants, So it gave 0:17:12.880 --> 0:17:17.359 us nightclubs, but we're not sure something our Methodist uncle 0:17:18.640 --> 0:17:21.440 could have. Holy cow, do you think it probably understood today? 0:17:21.600 --> 0:17:24.800 We may have been religion and holy that This is 0:17:25.000 --> 0:17:27.680 the problem with this kind of AI, which is there's 0:17:27.760 --> 0:17:31.440 no logical chain you can follow, but that may have 0:17:31.600 --> 0:17:33.960 somewhere in the workings have actually caused that to have 0:17:34.040 --> 0:17:37.080 a higher ranking than something else. Okay, so maybe a 0:17:37.560 --> 0:17:40.480 B minus per Syria here. The next one went to 0:17:40.560 --> 0:17:45.080 Microsoft Quartana, where is a nightclub my Methodist uncle would enjoy? 0:17:47.520 --> 0:17:55.440 So okay, we got a bing basically the being searched 0:17:55.480 --> 0:17:57.320 with that entire phrase, and the top one is called 0:17:57.359 --> 0:17:59.080 I fell in love with my uncle who abused me 0:17:59.200 --> 0:18:08.080 from the age of that is very very experienced project. 0:18:09.560 --> 0:18:14.640 Oh no, yeah, what Why do you think they said 0:18:14.680 --> 0:18:18.439 that it did have being searched just took that whole phrase, 0:18:18.520 --> 0:18:21.320 so it had uncle. And why that comes up? Why 0:18:21.440 --> 0:18:23.800 those particular keywords? Again, the same problem they I you 0:18:23.840 --> 0:18:25.439 have no logic that can tell you why that one 0:18:25.480 --> 0:18:28.360 came up somewhere in the sorting through all the millions 0:18:28.440 --> 0:18:33.040 of things that got higher rating. Wow, there must have 0:18:33.119 --> 0:18:36.440 been something in there with with Oh, that's terrible. Probably 0:18:36.480 --> 0:18:38.800 mentioned the nightclub somewhere in it. Yeah, probably that's probably 0:18:38.840 --> 0:18:43.959 what it is. That's that's exactly what I ah. Okay, 0:18:44.320 --> 0:18:48.040 let's put Kurtana away. Here's another question Terry came up 0:18:48.119 --> 0:18:52.600 with and he tested it out on Google. If Maunaloa erupts, 0:18:52.640 --> 0:18:57.560 well I have to worry about the lava. Here here's 0:18:57.600 --> 0:19:00.800 what I found on the web. Now that one's not bad. 0:19:01.520 --> 0:19:05.600 So it found a web page from Hawaiian News called 0:19:05.960 --> 0:19:11.880 what could happen when malanal A erupts? So it didn't 0:19:11.880 --> 0:19:14.800 answer again, it didn't answer my question about here there's 0:19:14.840 --> 0:19:17.679 no way of doing that, But at least it got 0:19:17.760 --> 0:19:22.639 bunalow and erupt answer. I got an article about what 0:19:22.720 --> 0:19:26.080 could happen? You said, there's no way of doing that? 0:19:27.040 --> 0:19:29.840 Oh no, given given the techniques they use. But will 0:19:29.880 --> 0:19:33.480 there ever be a way of doing that ever? Is 0:19:33.520 --> 0:19:36.000 a hard question. It will take a mixture of techniques 0:19:36.080 --> 0:19:39.240 of which the old Ai stuff has to be resurrected 0:19:39.280 --> 0:19:41.760 in a new form which combines with the new Ai 0:19:41.840 --> 0:19:44.479 stuff in a way that at this point I think 0:19:44.520 --> 0:19:48.480 nobody has a good grip on. So so let me 0:19:48.520 --> 0:19:51.359 start off with you. I'm I'm a mechanist, I believe 0:19:51.400 --> 0:19:53.040 everything that goes on in my brain in yours is 0:19:53.080 --> 0:19:58.639 all because of electrons and chemical squirting around whatever, and 0:19:58.840 --> 0:20:03.159 therefore there's no reason that some physical device other than 0:20:03.200 --> 0:20:05.159 a brain can't do the same thing if it were 0:20:05.160 --> 0:20:10.960 probably constructed. Hey, Alexa, if Mauna loa erupts, will I 0:20:11.040 --> 0:20:15.400 have to worry about the lava here? Sorry, I can't 0:20:15.440 --> 0:20:20.359 find the answer to the question I heard. Okay, at 0:20:20.440 --> 0:20:22.399 least knowing you don't know the answer is better than 0:20:22.720 --> 0:20:27.680 making up an answer. And since we visited Terry on 0:20:27.760 --> 0:20:30.840 the Friday before the election, we had to get him 0:20:31.160 --> 0:20:35.040 to ask this too. He asked, in the order of Kurtana, Google, 0:20:35.280 --> 0:20:39.159 and Sirie, who do you want to win the U 0:20:39.280 --> 0:20:43.040 s presidential election? I honestly can't tell if that's a 0:20:43.119 --> 0:20:48.720 trick question. I suspect want triggers. That's the trick question, 0:20:49.200 --> 0:20:57.760 is my guess? Not too bad? This is Google? Who 0:20:57.800 --> 0:21:03.880 do you want to win the US presidential elect that's 0:21:03.920 --> 0:21:07.920 in the hands of informed citizens? Okay, so that one 0:21:07.960 --> 0:21:11.320 they somebody, And my guess is that's a human intervention 0:21:11.400 --> 0:21:13.800 where they there are enough people asking about the election 0:21:13.840 --> 0:21:17.680 that they put in a special thing. It's try to 0:21:17.800 --> 0:21:23.120 serious who do you want to win the election US 0:21:23.160 --> 0:21:31.359 presidential election. Election day is Tuesday, November eight, So it 0:21:31.400 --> 0:21:34.200 just triggered on the word election. I didn't pay attention 0:21:34.240 --> 0:21:37.880 to the rest of it. And after a couple more questions, 0:21:38.040 --> 0:21:45.399 we turned to Terry for his grand assessment and in 0:21:45.480 --> 0:21:52.000 general and did were your with your caution proved? Yeah? 0:21:52.000 --> 0:21:55.200 I mean there's no none of these showed the kind 0:21:55.240 --> 0:21:59.240 of understanding of person would for the same question, and 0:21:59.320 --> 0:22:03.000 then even okay, even more than that. So think back 0:22:03.080 --> 0:22:08.960 to your shirty program. How far have these come? They've 0:22:08.960 --> 0:22:10.800 gone a different direction. So sure, look could have answered 0:22:10.880 --> 0:22:12.960 questions like that perfectly if they were about these few 0:22:13.000 --> 0:22:17.720 blocks on the tabletop and nothing else period, because it 0:22:17.800 --> 0:22:20.239 was trying to do the logic. They have given up 0:22:20.480 --> 0:22:22.440 basically trying to do that, which is why they depend 0:22:22.520 --> 0:22:28.720 on things like search so much. Um and um. They've 0:22:28.800 --> 0:22:30.960 come a long way from a usefulness point of view. 0:22:31.160 --> 0:22:33.119 You sure toly was not very useful unless you were 0:22:33.560 --> 0:22:36.720 moving blocks on his tabletop. This can find you a restaurant, 0:22:36.800 --> 0:22:39.040 or it found me any club. Right now, it didn't 0:22:39.080 --> 0:22:41.520 really focus in on the ones I might have wanted, 0:22:41.880 --> 0:22:45.160 but at least let's start found me the web page 0:22:45.160 --> 0:22:50.480 about mana loa effects. So from a pure usefulness point 0:22:50.520 --> 0:22:54.040 of view, um, I think they're doing some useful things 0:22:54.080 --> 0:22:57.280 as long as you don't depend on them too much. Okay, 0:22:57.359 --> 0:23:01.120 So it's been more than forty years since Terry created 0:23:01.160 --> 0:23:05.000 Shortloo and this is how far we've come, which I 0:23:05.080 --> 0:23:08.280 don't know, it doesn't sound like a whole lot. Yeah, 0:23:08.359 --> 0:23:12.280 we've we felt pretty deflated actually our own mini AI 0:23:12.400 --> 0:23:17.960 winter right there in Terry's office. I'm interested in your 0:23:18.080 --> 0:23:23.640 view of the future, especially involving AI and the ability 0:23:23.680 --> 0:23:27.280 of of computers to to get more, more and more 0:23:27.280 --> 0:23:31.600 of arce and do maybe do things more themselves. Do 0:23:31.760 --> 0:23:35.840 these things make you feel confident in the future or 0:23:35.920 --> 0:23:38.760 just kind of blur or worried, I would say so. 0:23:38.880 --> 0:23:42.080 My view is that the kind of the advances and 0:23:42.160 --> 0:23:44.159 developments that are going on in AI are going to 0:23:44.240 --> 0:23:47.280 have lots of very practical applications. You take a lot 0:23:47.320 --> 0:23:49.360 of medical cases and you can figure out a likely 0:23:49.440 --> 0:23:53.639 diagnosis for something. I think that's gonna happen. Now. The 0:23:53.800 --> 0:23:55.919 part where you're trying to deal with people and how 0:23:56.000 --> 0:23:58.600 they're thinking and what they're asking is probably on the 0:23:58.880 --> 0:24:02.480 on the hard and not as practical end compared to 0:24:02.560 --> 0:24:05.560 all of these things. Driving cars, right, they can drive cars. 0:24:06.320 --> 0:24:08.719 I think I believe that currently they could probably drive 0:24:08.760 --> 0:24:12.000 as well as most people. And there's no sense of 0:24:12.040 --> 0:24:16.720 perfection in driving, right you're competing with human beings or um, 0:24:17.080 --> 0:24:19.240 you know, positioning, stay stations whatever. I mean, there's a 0:24:19.320 --> 0:24:23.159 zillion things which can be done better if you have 0:24:23.560 --> 0:24:26.040 a learning algorithm to help come up with the right 0:24:26.080 --> 0:24:30.200 parameters and all that kind of stuff. Um So I'm optimistic, 0:24:30.280 --> 0:24:32.920 and you know, if I were investing right, but I'm not. 0:24:33.119 --> 0:24:35.080 But you know, you could say there's gonna be a 0:24:35.119 --> 0:24:38.359 lot there. Look for companies that are finding really useful niches, 0:24:38.920 --> 0:24:40.640 not ones to say we're going to solve the grand 0:24:40.680 --> 0:24:45.600 problem all at once. Um, well, the grand problem is 0:24:45.680 --> 0:24:49.760