WEBVTT - The Journey To AI 0:00:04.160 --> 0:00:07.160 Get in touch with technology with tech Stuff from how 0:00:07.240 --> 0:00:14.000 stuff works dot com. Hey there, welcome to tech Stuff. 0:00:14.040 --> 0:00:17.600 I'm your host, Jonathan Strickland. I'm an executive producer at 0:00:17.600 --> 0:00:21.160 how Stuff Works and I love all things tech. And yeah, 0:00:21.440 --> 0:00:25.000 I am at the IBM Think two thousand eighteen conference, 0:00:25.000 --> 0:00:27.680 which is why this sounds a little different than normal. 0:00:28.760 --> 0:00:32.400 I am in a hotel room over at the Excalibur Casino, 0:00:33.000 --> 0:00:35.080 and I wanted to talk a little bit about what 0:00:35.200 --> 0:00:39.040 I saw and some of the talks that I went to, 0:00:39.840 --> 0:00:43.319 and I learned a lot of interesting things. Now, one 0:00:43.360 --> 0:00:46.559 thing to say is that the THINK conference it's all 0:00:46.600 --> 0:00:51.320 about IBM and ibm S partners and customers. And unlike 0:00:51.400 --> 0:00:53.280 a lot of companies that we deal with on a 0:00:53.360 --> 0:00:59.160 day to day basis, IBM doesn't really have consumer facing businesses. 0:00:59.200 --> 0:01:01.480 In other words, it's not like you go to the 0:01:01.560 --> 0:01:05.440 store and you go buy IBM stuff. IBM mostly makes 0:01:05.480 --> 0:01:11.080 things for other companies and as such, we don't necessarily 0:01:11.160 --> 0:01:14.200 have to uh, we don't necessarily encounter it directly. We 0:01:14.319 --> 0:01:17.960 encounter IBMS products because they are inside other things that 0:01:18.120 --> 0:01:22.440 we are using. So uh, it's interesting to go to 0:01:22.520 --> 0:01:25.679 these events and to hear these talks, because a lot 0:01:25.720 --> 0:01:29.960 of it is stuff that is very much relevant for 0:01:30.400 --> 0:01:35.800 business leaders or for I T. Professionals, or for UH 0:01:36.200 --> 0:01:39.840 infrastructure engineers that kind of thing, but less so for 0:01:39.880 --> 0:01:43.880 the general public unless you step back a little bit. 0:01:44.160 --> 0:01:47.199 Even so, there were some really interesting talks that talked 0:01:47.240 --> 0:01:51.560 about UH where the future is headed as far as 0:01:53.040 --> 0:01:56.800 very big, broad technologies, and I thought that that would 0:01:56.800 --> 0:01:59.000 be the best way to kind of tackle this, to 0:01:59.080 --> 0:02:02.800 talk about these sort of trends that have been identified 0:02:03.200 --> 0:02:06.320 and these predictions that have been made about these kinds 0:02:06.360 --> 0:02:08.080 of tech, because those are the sort of things that 0:02:08.120 --> 0:02:10.960 are going to affect us moving forward, us being, you know, 0:02:11.000 --> 0:02:14.000 the average person as opposed to people who are running 0:02:14.200 --> 0:02:17.000 a tech company. One of the things that they talked 0:02:17.040 --> 0:02:21.680 about UH both at the the keynote speech that was 0:02:21.919 --> 0:02:24.640 technically the very first big keynote speech that was a 0:02:24.800 --> 0:02:29.080 Jenny Romti. Jenny Rometti is the CEO of IBM. She 0:02:29.680 --> 0:02:34.000 got up and spoke very directly to IBMS partners and customers. 0:02:34.600 --> 0:02:39.160 She talked about how there are different laws that we 0:02:39.240 --> 0:02:43.400 have created more like observations really that UM that have 0:02:44.120 --> 0:02:47.519 described the way technology has developed over the years. Now. 0:02:47.680 --> 0:02:50.520 The most famous one is one I've talked about numerous 0:02:50.560 --> 0:02:55.119 times on this show. That would be Moore's law, Moore's law, 0:02:55.240 --> 0:02:57.800 which was proposed by Gordon Moore. Of course he didn't 0:02:57.880 --> 0:03:00.240 call it Moore's law. He just made an observe. Sian 0:03:01.200 --> 0:03:06.160 was about how every eighteen months or so year and 0:03:06.160 --> 0:03:09.840 a half to two years, the number of discrete components 0:03:10.440 --> 0:03:15.000 meaning transistors at that time on a microchip we're doubling. 0:03:15.720 --> 0:03:22.120 And this observation wasn't about necessarily our technological capabilities, like 0:03:22.160 --> 0:03:25.239 the ability to make things that small. It was more 0:03:25.280 --> 0:03:30.639 about the fact that economics demanded that this was the case, 0:03:30.720 --> 0:03:33.840 that there was enough of a demand two in to 0:03:34.000 --> 0:03:39.520 give an incentive to manufacturing facilities that made these microchips 0:03:39.600 --> 0:03:43.280 to try and make ever smaller components to make more 0:03:43.320 --> 0:03:48.040 powerful processors. So, in other words, it wasn't so much 0:03:48.080 --> 0:03:52.400 that we had this these egghead scientists locked in the 0:03:52.480 --> 0:03:55.720 laboratory coming up with new ways to make transistors smaller. 0:03:56.000 --> 0:04:00.320 It was more like we had money in wheelbarrows out side, 0:04:00.800 --> 0:04:03.520 and we can only get that money if we made 0:04:03.560 --> 0:04:08.240 smaller transistors. And so it was really an economic driven law. 0:04:08.400 --> 0:04:11.680 But the effect that we have on us, it doesn't 0:04:11.680 --> 0:04:13.800 really matter. The economic part we can kind of ignore. 0:04:13.880 --> 0:04:16.880 What we look at is the fact that our processing 0:04:16.960 --> 0:04:21.360 power effectively doubles every eighteen months or so. So every 0:04:21.440 --> 0:04:24.640 year and a half to two years, the machines were 0:04:24.760 --> 0:04:28.440 using are twice as powerful as the ones that were 0:04:28.760 --> 0:04:31.960 two years ago. Uh, And that's kind of cool. It 0:04:32.000 --> 0:04:38.880 means that we keep getting these incredibly sophisticated machines on 0:04:38.920 --> 0:04:43.200 a regular basis, and a lot of the technology sectors 0:04:43.240 --> 0:04:48.440 businesses depend upon the continuation of Moore's law. Later on, 0:04:48.920 --> 0:04:52.839 I was at a talk with Dr Michio Kaku, who 0:04:53.000 --> 0:04:56.760 is a famous physicist and futurist. He talked a little 0:04:56.800 --> 0:04:59.680 bit about the end of the era of Moore's law. 0:04:59.720 --> 0:05:02.480 He did not give a specific prediction as to when 0:05:02.520 --> 0:05:05.320 it would end, but he did say that based just 0:05:05.440 --> 0:05:08.840 purely on physics alone, it will end. What he meant 0:05:08.880 --> 0:05:12.760 by that is, More's law depends on us shrinking these 0:05:12.800 --> 0:05:16.600 components down more and more and more. Once you get 0:05:16.600 --> 0:05:19.440 to the point where the quantum world comes into play, 0:05:19.760 --> 0:05:22.120 this gets really tricky and I've talked about this before, 0:05:22.120 --> 0:05:25.040 to the fact that if you were to create logic 0:05:25.080 --> 0:05:28.960 gates that are so thin that an electron could potentially 0:05:29.040 --> 0:05:32.240 exist on the other side of a logic gate, then 0:05:32.320 --> 0:05:35.719 sometimes an electron is going to be on the other 0:05:35.760 --> 0:05:38.520 side of the uh the electron gates sort of like 0:05:38.839 --> 0:05:42.880 it had tunneled through, except it had not physically tunneled 0:05:43.000 --> 0:05:46.560 through the wall. It's just that it had the probability 0:05:46.760 --> 0:05:49.360 of potentially being on the other side of that wall. 0:05:49.400 --> 0:05:51.280 And as long as there's a probability, it means that 0:05:51.400 --> 0:05:54.599 sometimes that does happen. Even though that you know, in 0:05:54.640 --> 0:05:57.240 the classical world we would say, well, there's a barrier there. 0:05:57.279 --> 0:05:59.880 You can't just go through. A barrier didn't go through. 0:05:59.880 --> 0:06:02.479 It just appeared on the other side because there was 0:06:02.520 --> 0:06:05.280 a chance it could. And if there's a chance, then 0:06:05.320 --> 0:06:08.800 sometimes that does happen. Well, even beyond that, even if 0:06:08.839 --> 0:06:12.839 you say, well, we'll keep figuring out ways to counteract 0:06:13.080 --> 0:06:18.440 this quantum effect so that we can keep having microprocessors 0:06:18.440 --> 0:06:22.640 that are accurate even with quantum tunneling being an issue, 0:06:22.880 --> 0:06:25.479 you eventually get down to the point where you're at 0:06:25.520 --> 0:06:30.120 the atomic scale, meaning the components you're creating are made 0:06:30.120 --> 0:06:33.120 out of atoms themselves. At this stage, you really it 0:06:33.360 --> 0:06:37.320 would be really difficult to counteract those quantum effects and 0:06:37.480 --> 0:06:43.279 you would have to abandon this particular approach to computer 0:06:43.360 --> 0:06:47.960 science and computer architecture, or else it would just collapse 0:06:48.000 --> 0:06:52.520 in on itself. So Moore's law, while it was incredibly important, 0:06:52.560 --> 0:06:56.240 and it continues to be incredibly important right now. Um Ever, 0:06:56.320 --> 0:07:01.400 since you know the transistor was invented, it it it 0:07:01.440 --> 0:07:04.240 only represents the first kind of wave of laws. The 0:07:04.279 --> 0:07:07.239 next law that they talked about was one they called 0:07:07.440 --> 0:07:13.520 Metcalf's law. Uh. Metcalf's law is actually pretty commonly referred 0:07:13.560 --> 0:07:17.720 to law, just not necessarily among you know, regular folks 0:07:17.760 --> 0:07:21.080 like me and you. But Metcalf's law is about the 0:07:21.240 --> 0:07:27.040 value of a network. So how do you measure how 0:07:27.160 --> 0:07:30.160 valuable a network is? Like if you look at a 0:07:30.240 --> 0:07:33.080 network of devices, and then you look at a different 0:07:33.120 --> 0:07:35.440 network of devices, how could you say which one is 0:07:35.440 --> 0:07:39.080 is quote unquote worth more. Metca Cat's law gives you 0:07:39.160 --> 0:07:43.400 that that measurement. It states that the value of a network, 0:07:43.920 --> 0:07:48.400 of a telecommunications network is proportional to the square of 0:07:48.440 --> 0:07:52.720 the number of connected nodes in the system. So, however 0:07:52.800 --> 0:07:55.480 many nodes are there, and the node can be any 0:07:55.520 --> 0:07:58.520 connected device. They could be a computer, it could be 0:07:58.520 --> 0:08:01.280 a smartphone, could be a table, it could be a 0:08:01.360 --> 0:08:09.040 game console. Those nodes collectively end up determining the value 0:08:09.200 --> 0:08:12.480 of the telecommunications network. When you square the number of 0:08:12.480 --> 0:08:16.280 those nodes. It's those interconnections that make the network of 0:08:16.320 --> 0:08:20.640 a valuable. This is incredibly important again in the world 0:08:20.640 --> 0:08:23.840 of business, less so probably for for me and you. 0:08:24.480 --> 0:08:27.640 The third one, the third law that they were proposing, 0:08:28.400 --> 0:08:32.560 would be what they were cheekily referring to as Watson's law. Watson, 0:08:32.720 --> 0:08:38.520 of course, is not just and artificially intelligent platform for 0:08:38.720 --> 0:08:43.680 IBM and for IBMS customers and partners. Watson also refers 0:08:43.720 --> 0:08:47.120 to the founder of IBM what was his name, his 0:08:47.200 --> 0:08:53.200 last name, But Watson's law would be about how the 0:08:53.240 --> 0:08:56.720 amount of data in a system can be leveraged to 0:08:56.880 --> 0:09:00.920 get the amount of knowledge out of that data. This 0:09:01.040 --> 0:09:04.840 sort of as as data grows exponentially, our ability to 0:09:04.960 --> 0:09:08.640 leverage knowledge from that data grows exponentially. So what the 0:09:08.640 --> 0:09:13.080 heck does that mean? Well, think of data as just 0:09:13.200 --> 0:09:17.040 points of information that are not necessarily connected to one another. 0:09:17.160 --> 0:09:22.280 They're not structured necessarily. This would be as if I 0:09:22.360 --> 0:09:25.400 recorded a podcast and I just started to say random 0:09:25.800 --> 0:09:29.720 words into the microphone, and I did that for forty 0:09:29.800 --> 0:09:33.120 five minutes to an hour. And Okay, smart Alex, you 0:09:33.200 --> 0:09:36.000 might think that's how I do it now, but you're 0:09:36.120 --> 0:09:38.960 you're just mean, you're meaning heads. That's not how I 0:09:39.000 --> 0:09:42.360 do it. I actually think this stuff out and I 0:09:43.000 --> 0:09:47.520 structure my data so that I create a foundation and 0:09:47.559 --> 0:09:50.320 then I build upon it. That's a very easy way 0:09:50.360 --> 0:09:53.560 to get knowledge, right. You have the structured format, you 0:09:53.600 --> 0:09:56.559 can digest it, you can synthesize it. You can then 0:09:56.760 --> 0:10:00.280 use that yourself. But if the data is unstru shirt, 0:10:00.640 --> 0:10:05.959 and the data is about a lot of seemingly unconnected things, 0:10:06.320 --> 0:10:10.719 and it's spread across multiple types of files, Let's say 0:10:10.800 --> 0:10:15.400 that you've got an enormous folder, uh, and that folder 0:10:15.480 --> 0:10:20.320 contains files that are video files, they are documents, their presentations, 0:10:20.679 --> 0:10:24.120 their spreadsheets, they're all these different things that that on 0:10:24.200 --> 0:10:28.280 casual glance don't have any connectivity to them. How can 0:10:28.320 --> 0:10:31.800 you make that useful so that you can actually leverage 0:10:31.840 --> 0:10:35.439 that data and do stuff with it? And that's kind 0:10:35.480 --> 0:10:38.240 of what IBM was focusing on. And that's really where 0:10:38.240 --> 0:10:41.120 they were talking about Watson quite a lot. It wasn't 0:10:41.480 --> 0:10:45.080 a lot of people think of Watson as this, uh, 0:10:45.120 --> 0:10:48.520 this the supercomputer that played on Jeopardy, which is not accurate. 0:10:48.840 --> 0:10:53.600 Watson is not a supercomputer. The machine that ran Watson 0:10:54.320 --> 0:10:59.720 was just a machine. It was not the entity itself. Uh. 0:11:00.000 --> 0:11:01.959 If you want to do to get a little metaphysical 0:11:02.000 --> 0:11:04.080 with this, you could actually think about a human being 0:11:04.120 --> 0:11:06.840 and you ask, well, what is the human Is the 0:11:06.920 --> 0:11:10.800 human being the body, the physical form, or is it 0:11:10.840 --> 0:11:15.079 the mind? The person, the personality, the emotions, the memories, 0:11:15.520 --> 0:11:19.560 the things that are that inhabit the body and also 0:11:19.640 --> 0:11:23.280 that control the body. Is that the person? And you 0:11:23.360 --> 0:11:25.920 might argue, well, it's actually the collective. It's the body 0:11:25.960 --> 0:11:28.280 and the mind, and I think that's a valid argument. 0:11:28.280 --> 0:11:32.800 You could also argue that Watson ultimately is a platform 0:11:33.200 --> 0:11:37.360 and the physical machine that runs that platform. I probably 0:11:37.400 --> 0:11:39.600 wouldn't argue with you too much there either, except I'd 0:11:39.640 --> 0:11:42.600 say that the platform is more important than anything else 0:11:42.640 --> 0:11:45.440 in this in this particular case, And by platform I 0:11:45.480 --> 0:11:48.480 really just means set of rules, set of algorithms that 0:11:48.520 --> 0:11:52.000 Watson uses in order to process information, to look for meaning, 0:11:52.040 --> 0:11:58.960 to look for results. So let's take that Jeopardy example, Uh, Jeopardy. 0:11:59.040 --> 0:12:01.760 In Jeopardy, Wat's and played against two former champions, one 0:12:01.760 --> 0:12:04.040 of whom now Records podcast for How Stuff Works. So 0:12:04.080 --> 0:12:08.880 that's kind of awesome. And Watson was playing by looking 0:12:08.920 --> 0:12:11.280 at a clue. We're looking quote unquote. It was the 0:12:11.320 --> 0:12:15.400 clues being fed to Watson and then going through its 0:12:15.559 --> 0:12:18.520 massive amount of data and trying to use that to 0:12:18.559 --> 0:12:21.200 figure out what the answer is. And it wasn't just 0:12:21.280 --> 0:12:24.679 looking at a list of trivia or facts. It's not 0:12:24.760 --> 0:12:29.200 like it's looking at an enormous table and every cell 0:12:29.800 --> 0:12:33.600 in that table is filled with a different fact, like 0:12:33.840 --> 0:12:37.520 George Washington was the first President of the United States. Instead, 0:12:37.520 --> 0:12:42.240 it's looking at a massive library of information and pulling 0:12:42.880 --> 0:12:46.760 bits and pieces of information together to formulate an idea 0:12:46.760 --> 0:12:50.280 of what the answer is. And if that formulation reaches 0:12:50.320 --> 0:12:54.120 a certain threshold of confidence, Watson would then ring in 0:12:54.600 --> 0:12:59.000 and present that answer. So it's it's not that it's 0:12:59.360 --> 0:13:03.520 looking at, uh you know, a very long trivia book. 0:13:03.920 --> 0:13:07.880 It's looking at all this information and drawing conclusions from 0:13:07.880 --> 0:13:10.640 it the way similar to how a human being would 0:13:10.760 --> 0:13:17.680 not not completely analogous, but similar and uh so, using Watson, 0:13:18.120 --> 0:13:24.119 you could leverage your unstructured data. You put Watson into 0:13:24.240 --> 0:13:26.800 work at this, and Watson would start to look for 0:13:27.000 --> 0:13:32.920 meaningful connections between data points and pulling relevant information about 0:13:32.960 --> 0:13:36.839 any given query. So then Watson becomes an agent that 0:13:36.960 --> 0:13:40.320 you could interact with. And this agent's job is kind 0:13:40.360 --> 0:13:43.440 of like a reference librarian. It's to go to the 0:13:43.480 --> 0:13:47.000 massive amount of information that's at its disposal and return 0:13:47.080 --> 0:13:50.040 to you the relevant points of information. This is not 0:13:50.160 --> 0:13:54.080 that different from the way people were thinking about web 0:13:54.160 --> 0:13:57.679 three point oh when that was a big discussion. H 0:13:57.840 --> 0:14:00.720 you may remember that like people to talk about how 0:14:01.559 --> 0:14:03.959 right now? If you use a search engine, typically the 0:14:03.960 --> 0:14:05.959 way it works as you type something in the search 0:14:06.000 --> 0:14:09.040 engine and it pulls up a list of websites that 0:14:09.120 --> 0:14:12.079 may or may not have what you're looking for on 0:14:12.160 --> 0:14:15.240 those websites. So if you might you might be looking 0:14:15.280 --> 0:14:20.200 for a let's say it's a um A history of 0:14:20.720 --> 0:14:25.480 the Crusades, and you type that into the search engine 0:14:25.520 --> 0:14:28.080 and it pulls for you a bunch of different sites 0:14:28.160 --> 0:14:30.480 written by different people. Some of them might be very 0:14:30.520 --> 0:14:33.640 easy to read and understand. Some of them might be 0:14:33.760 --> 0:14:36.280 less easy to read, but they might be more accurate 0:14:36.400 --> 0:14:41.160 and more uh unbiased. With the information you don't necessarily 0:14:41.200 --> 0:14:43.320 know at the top of it. You have to go 0:14:43.400 --> 0:14:45.960 through and read all that yourself. But the web three 0:14:45.960 --> 0:14:48.760 point of search engines. This was something that Will from 0:14:48.760 --> 0:14:53.840 Alpha was trying to be would pull the relevant information 0:14:54.160 --> 0:14:58.640 not websites, but the relevant information from those websites and 0:14:58.720 --> 0:15:02.120 present it to you. And that way you could look 0:15:02.160 --> 0:15:05.320 over the important bits of information, you skip over everything 0:15:05.320 --> 0:15:09.120 else you're given the correct context. In theory, you could 0:15:09.120 --> 0:15:12.440 even have an agent like this that could learn about 0:15:12.680 --> 0:15:16.240 you and your learning styles and thus present the information 0:15:16.280 --> 0:15:19.240 to you in a way that is most helpful to you. 0:15:19.840 --> 0:15:22.080 So it's a very big difference between the way we 0:15:22.120 --> 0:15:25.360 do searches now and the way that this proposed method 0:15:25.440 --> 0:15:27.520 would work. And that's kind of what Watson is doing. 0:15:28.000 --> 0:15:32.200 So you've got this this user facing aspect of Watson. 0:15:32.280 --> 0:15:35.400 It's kind of like a chat bot, and you can 0:15:35.640 --> 0:15:39.080 send that chat bot requests and then the chat bot 0:15:39.120 --> 0:15:42.280 will try and pull the information for you, or you 0:15:42.280 --> 0:15:44.920 can use it to generate reports. Let's say that you 0:15:45.000 --> 0:15:48.680 are a business owner and you want to look at 0:15:48.800 --> 0:15:54.440 some information that's gonna pull things from presentations, predictions, results. 0:15:54.480 --> 0:15:57.360 Maybe you've got like a end of the quarter report. 0:15:57.960 --> 0:16:00.840 Maybe you want to take a look at formation from 0:16:01.040 --> 0:16:04.840 reports from your supply chain. All this kind of complicated 0:16:04.920 --> 0:16:10.200 stuff and Watson could go out, curate and present this 0:16:10.280 --> 0:16:13.560 information in a way that has meaning to you, that 0:16:13.560 --> 0:16:16.520 where you can understand what's going on and you can 0:16:16.600 --> 0:16:21.720 draw conclusions. Uh. This actually was a pretty interesting concept 0:16:21.840 --> 0:16:25.440 to me. I mean, I've seen some implementations of Watson 0:16:26.000 --> 0:16:30.000 that do this, and they do it in such a simple, 0:16:30.640 --> 0:16:35.320 seemingly simple way that's deceptive. You start to forget that 0:16:35.360 --> 0:16:39.800 there is a very powerful computer algorithm that is controlling 0:16:39.800 --> 0:16:44.240 all of this because the implementation itself might be pretty straightforward. 0:16:44.760 --> 0:16:47.760 So for an example, I went to the Weather Company 0:16:48.200 --> 0:16:51.720 last year in TV and while I was there, I 0:16:51.760 --> 0:16:54.280 had a chance to talk to a team that was 0:16:54.360 --> 0:16:58.840 using Watson in a lot of different implementations, and uh, 0:16:58.880 --> 0:17:00.880 you know, they were using it as is the basis 0:17:01.000 --> 0:17:04.959 of a customer service platform or to respond to requests. 0:17:05.680 --> 0:17:08.919 And when you first look at that, it looks deceptively simple. 0:17:08.960 --> 0:17:11.120 You're asking, well, what's the weather going to be like? 0:17:11.240 --> 0:17:14.639 And you get results, Uh that that doesn't seem like 0:17:14.680 --> 0:17:16.600 it's that hard. You would figure that, oh, well, they're 0:17:16.640 --> 0:17:19.399 just gonna pull whatever the record is for my location 0:17:20.000 --> 0:17:23.520 for tomorrow and present it to me. But a lot 0:17:23.560 --> 0:17:25.879 more could be going on behind the scenes, and I 0:17:25.920 --> 0:17:28.399 think that's part of the problem that IBM has been 0:17:28.440 --> 0:17:31.920 dealing with and kind of one of the reasons why 0:17:31.960 --> 0:17:34.520 they've made such a big deal of it at this conference. 0:17:35.119 --> 0:17:40.320 It's because the perception of what Watson is maybe a 0:17:40.359 --> 0:17:46.440 little too narrow, a little too uh uh focused on 0:17:47.680 --> 0:17:51.399 little aspects of what Watson does and ignores the big picture. 0:17:51.800 --> 0:17:56.440 So they've they've definitely doubled down on that. I went 0:17:56.520 --> 0:17:59.920 to a talk called Journey to AI that was really 0:18:00.160 --> 0:18:04.159 all about this, and they talked all about the the 0:18:04.240 --> 0:18:09.280 different variations of artificial intelligence, and uh one of the 0:18:09.359 --> 0:18:13.080 things they mentioned was the very different views of what 0:18:13.240 --> 0:18:18.000 AI is. For example, you've got simple AI. Simple AI 0:18:18.040 --> 0:18:19.800 would include some of the stuff I talked about in 0:18:19.800 --> 0:18:24.199 a previous episode about the little aspects of intelligence that 0:18:24.280 --> 0:18:29.280 are very very narrow, just to slice the pie of intelligence, 0:18:29.359 --> 0:18:32.560 but they do represent what intelligence is in in just 0:18:32.600 --> 0:18:36.320 a very specific application. So image recognition is an example 0:18:36.359 --> 0:18:40.600 of that, or voice recognition or natural language processing even 0:18:40.760 --> 0:18:44.399 as part of that. These are all aspects of intelligence. 0:18:44.920 --> 0:18:47.720 You would not call a machine that lacks one of 0:18:47.760 --> 0:18:51.399 these things truly intelligent, but you also wouldn't call a 0:18:51.440 --> 0:18:54.720 machine that only has one of these things truly intelligent. 0:18:55.040 --> 0:18:58.199 So if I have a smartphone and the smartphone is 0:18:58.240 --> 0:19:02.320 able to recognize uh images, so i'm I'm I point 0:19:02.359 --> 0:19:05.280 my smartphone at something and it even labels what that 0:19:05.359 --> 0:19:07.960 thing is. Maybe it says, oh, well, that's a specific 0:19:08.080 --> 0:19:12.320 model and make of car, or maybe it says that 0:19:12.440 --> 0:19:18.040 building is a historic landmark, or this park is going 0:19:18.080 --> 0:19:22.359 to have a concert uh at at it the next day, 0:19:22.480 --> 0:19:25.680 or something along those lines. That's cool. That image recognition 0:19:25.760 --> 0:19:29.840 is really cool, but I wouldn't call my smartphone intelligent. Similarly, 0:19:29.880 --> 0:19:32.359 if my smartphone happens to have one of those digital 0:19:32.400 --> 0:19:34.680 assistance on it, and it does, I've got an Android phone, 0:19:34.720 --> 0:19:37.439 so I've got the Google Assistant on there. Um, I 0:19:37.480 --> 0:19:41.159 can talk to that and it can retrieve information for me. 0:19:41.480 --> 0:19:43.879 It can do tasks for me. I can use it 0:19:43.920 --> 0:19:47.440 to make calls, I can use it to send text messages, 0:19:47.560 --> 0:19:49.600 or I can use it to search for information on 0:19:49.640 --> 0:19:52.280 my phone or on the internet. I still wouldn't call 0:19:52.320 --> 0:19:56.320 my phone intelligent. It has an aspect of intelligence. Similarly, 0:19:56.320 --> 0:20:01.000 if I had a supercomputer that could listen to voice commands, 0:20:01.080 --> 0:20:04.320 respond in natural language, and do these other things, but 0:20:04.359 --> 0:20:06.639 it couldn't do any image recognition. I would feel I 0:20:06.640 --> 0:20:09.159 would I would notice that lack, and I wouldn't call 0:20:09.240 --> 0:20:12.480 that intelligent. On the other side of the scale, you 0:20:12.560 --> 0:20:17.040 have general AI, where you know, the classic image of 0:20:17.040 --> 0:20:20.480 this is you've got a big machine. They can do 0:20:21.760 --> 0:20:25.440 uh that can do general thinking, like thinking that's analogous 0:20:25.440 --> 0:20:28.840 to human thinking. It can process information, it can draw conclusions, 0:20:28.880 --> 0:20:33.080 that can synthesize data. It can um innovate. It may 0:20:33.119 --> 0:20:36.760 even be self aware, although the weather or not self 0:20:36.800 --> 0:20:39.919 awareness is directly tied to intelligence is a matter of 0:20:39.960 --> 0:20:45.160 philosophical debate. Talking about general AI, I mean, that's that's 0:20:45.160 --> 0:20:49.600 a hard, hard goal to hit. We honestly don't know 0:20:49.880 --> 0:20:52.720 what it will take to get there. It may be 0:20:53.000 --> 0:20:56.879 that we are thirty years away from having a true 0:20:57.040 --> 0:21:00.359 general AI, It may be much longer than that, it 0:21:00.400 --> 0:21:03.320 maybe a century away, or it may even be impossible 0:21:03.320 --> 0:21:08.000 for us to do based upon our technological abilities. Right now, 0:21:08.119 --> 0:21:12.960 most technologists think that it is attainable, but they don't 0:21:13.000 --> 0:21:15.440 know exactly what it's going to take to get there. 0:21:15.600 --> 0:21:19.480 So there's some argument about the timeline, But there are 0:21:19.480 --> 0:21:22.880 a lot of interesting things that can happen between those 0:21:22.960 --> 0:21:28.760 simple versions of AI, and that that crazy general AI 0:21:28.800 --> 0:21:31.679 that that you know, science fiction writers write about and 0:21:32.119 --> 0:21:36.320 warn us about. And that's where this this ability to 0:21:36.600 --> 0:21:42.760 deal with unstructured data comes in and h designing AI 0:21:42.880 --> 0:21:45.439 is part of that problem. But as they mentioned in 0:21:45.600 --> 0:21:49.240 multiple presentations here at IBM, it's not just building the 0:21:49.320 --> 0:21:54.240 artificial intelligence to do this that's a challenge. It's also 0:21:54.359 --> 0:22:00.200 incorporating that artificial intelligence into existing work practices because, as 0:22:00.240 --> 0:22:04.400 most businesses have existed for a while now, it's not 0:22:04.520 --> 0:22:07.480 like you can just slot AI n necessarily. It's not 0:22:07.560 --> 0:22:10.200 like a module you plug in and everything works properly. 0:22:10.520 --> 0:22:15.720 You might have to reevaluate and redesign work processes in 0:22:15.840 --> 0:22:18.320 order to make this happen. And again, this gets a 0:22:18.320 --> 0:22:21.200 little little dry and technical if you're not really into 0:22:21.240 --> 0:22:23.200 the business side of things. But when you start thinking 0:22:23.200 --> 0:22:26.160 about you realize, yeah, it's not enough to just build 0:22:26.160 --> 0:22:28.400 a tool. You have to figure out how's the best 0:22:28.400 --> 0:22:32.879 way to use that tool with respect to the things 0:22:32.920 --> 0:22:36.920 you're already trying to do that. They started talking about 0:22:36.960 --> 0:22:42.320 impotence match. The engineers were chatting, chatting all about impotence 0:22:42.359 --> 0:22:46.160 match between man and machine to get machines to process 0:22:46.320 --> 0:22:50.000 human language and commands and to return information that would 0:22:50.000 --> 0:22:53.439 be useful to humans, and to eventually get rid of 0:22:53.480 --> 0:22:56.960 that boundary between man and machines so that decisions can 0:22:56.960 --> 0:23:01.000 be made together and implemented together. So this gets into 0:23:01.080 --> 0:23:05.280 that concept of augmented intelligence, not that we are trying 0:23:05.320 --> 0:23:09.760 to create a supercomputer that is incredibly intelligent, and we 0:23:09.800 --> 0:23:13.239 will then reference the supercomputer as if it were an 0:23:13.240 --> 0:23:17.800 oracle or a deity, instead talking about creating machines that 0:23:17.800 --> 0:23:23.440 would work right alongside people, and the machines could help 0:23:23.600 --> 0:23:26.920 fill in the gaps that would be there because of 0:23:26.960 --> 0:23:30.280 the human failings that are in all of us, and 0:23:30.359 --> 0:23:34.679 humans could provide all the bits that machines are not 0:23:34.800 --> 0:23:38.720 good at, and together we could be better. And that 0:23:38.800 --> 0:23:40.159 we have to get to a point where we have 0:23:40.280 --> 0:23:44.280 to trust the machines as a an assistant, and the 0:23:44.280 --> 0:23:47.360 machines have to quote unquote, trust us as teachers. By 0:23:47.359 --> 0:23:49.640 trust us, they don't necessarily mean that the machines are 0:23:49.640 --> 0:23:54.280 going to be harboring doubts, but rather that humans are 0:23:54.280 --> 0:23:57.800 the ones designing these machines, and we have to make 0:23:57.840 --> 0:24:01.560 certain that we do so in a way that is responsible, 0:24:01.640 --> 0:24:06.080 that is ethical, that is inclusive. Otherwise we end up 0:24:06.119 --> 0:24:09.240 with bad machines. And it's not that the machines themselves 0:24:09.240 --> 0:24:14.320 were inherently wicked, but rather they were poorly designed. I've 0:24:14.359 --> 0:24:17.840 got more to say about the Journey to AI presentation 0:24:18.000 --> 0:24:21.200 at IBM THINK, but before I go into it, let's 0:24:21.200 --> 0:24:31.240 take a quick break to thank our sponsor. The folks 0:24:31.240 --> 0:24:35.240 over at IBM are arguing that every single industry across 0:24:35.280 --> 0:24:38.000 the world is going to be affected by this sort 0:24:38.000 --> 0:24:46.560 of transformation of of data and knowledge. They started referencing 0:24:46.600 --> 0:24:52.520 things like retail optimization, or the oil industry, or automotive 0:24:52.520 --> 0:24:56.520 industries shipping. All of these things they said, we're going 0:24:56.560 --> 0:24:59.040 to transform dramatically over the next few years due to 0:24:59.119 --> 0:25:02.520 this kind of technology. Uh. And they talked about how 0:25:03.119 --> 0:25:06.439 the one field you can look at right now that 0:25:06.600 --> 0:25:11.480 is undergoing such a transformation is healthcare. All healthcare is 0:25:11.480 --> 0:25:15.760 is transforming because we are seeing not just advanced tools 0:25:15.840 --> 0:25:20.600 come into hospitals and doctors offices, but also these programs 0:25:20.640 --> 0:25:24.720 like Watson where a doctor can actually turn to Watson 0:25:25.440 --> 0:25:30.439 as a colleague, almost like someone up here who can 0:25:31.560 --> 0:25:35.359 provide more information a second opinion, if you will. In fact, 0:25:35.560 --> 0:25:39.760 IBM brought up some representatives from the American Cancer Society 0:25:40.200 --> 0:25:46.240 and some very prestigious cancer research hospitals to talk about 0:25:46.320 --> 0:25:51.560 this and about how cancer is a really really difficult problem. 0:25:51.640 --> 0:25:56.000