WEBVTT - Jonathan Attends IBM Think 2018 0:00:04.160 --> 0:00:07.160 Get in tech with technology with tech Stuff from how 0:00:07.240 --> 0:00:14.200 stuff works dot com. Hey there, and welcome to tech Stuff. 0:00:14.240 --> 0:00:18.159 I'm your host, Jonathan Strickland. I'm an executive producer with 0:00:18.239 --> 0:00:22.320 how Stuff Works and I love all things tech. This week, 0:00:22.720 --> 0:00:26.400 I've got a series of special shows in addition to 0:00:26.440 --> 0:00:29.720 our normal episodes that I am happy to bring with you. 0:00:29.720 --> 0:00:32.640 You may notice that things sound a little different than 0:00:32.680 --> 0:00:35.400 they usually do. That is because I am not currently 0:00:35.600 --> 0:00:40.880 in the massive, wonderful, underground, super secret studio at how 0:00:40.960 --> 0:00:45.440 stuff Works. Instead, I am on the road. I'm actually 0:00:45.479 --> 0:00:50.120 in Las Vegas, Nevada, where I am attending the inaugural 0:00:50.479 --> 0:00:55.280 IBM Think Conference. It's taking place between March nineteenth and 0:00:55.320 --> 0:01:01.360 March ten and it's kind of a big industry conference 0:01:01.440 --> 0:01:04.880 for people who are in the text space, who work 0:01:05.000 --> 0:01:09.199 with computers and software and hardware. Kind of a place 0:01:09.200 --> 0:01:11.480 for them to meet, to network and to learn a 0:01:11.480 --> 0:01:16.880 lot more about some bleeding edge technologies. They have lots 0:01:16.920 --> 0:01:19.880 of different activities going on all week, and I am 0:01:19.959 --> 0:01:22.720 going to be here attending some of the events, talking 0:01:22.760 --> 0:01:27.080 to people, getting more information, and recording special episodes for you, 0:01:27.440 --> 0:01:32.440 my beloved listeners. So I hope you enjoy this special 0:01:32.480 --> 0:01:35.920 series and I can't wait to really dive in and 0:01:36.040 --> 0:01:40.080 learn more about these topics. It looks really interesting. Now. 0:01:40.240 --> 0:01:42.680 One thing I will say before I jump in and 0:01:42.680 --> 0:01:44.640 give you kind of an overview of what I have 0:01:44.840 --> 0:01:49.960 to expect this week is that it's crazy busy, y'all. 0:01:50.000 --> 0:01:54.040 I mean, the conference has got so many people attending. 0:01:54.080 --> 0:01:57.080 I haven't seen any numbers about how many people are 0:01:57.120 --> 0:02:00.520 actually at this thing. But it's taking place at Mandalay 0:02:00.600 --> 0:02:04.160 Bay as one of the resorts and casinos here at 0:02:04.240 --> 0:02:07.480 Las Vegas, and it has a very large convention center. 0:02:07.800 --> 0:02:10.440 During ce S, you may have heard me do several 0:02:10.440 --> 0:02:13.400 episodes about ce S over the years. This is the 0:02:13.440 --> 0:02:17.680 place where the press events happened before the show floor opens. 0:02:17.680 --> 0:02:20.519 At the Las Vegas Convention Center at Mandalaid Bay, you 0:02:20.560 --> 0:02:23.680 would have all the different rooms booked up for the 0:02:23.760 --> 0:02:29.079 various press conferences that companies would use to unveil their 0:02:29.200 --> 0:02:32.160 latest and greatest products they're coming out over the next 0:02:32.400 --> 0:02:36.600 year or two. Well, at IBM think that's what is 0:02:36.639 --> 0:02:39.800 being used as the meeting space and the keynote space, 0:02:40.120 --> 0:02:45.560 presentation space for all sorts of different I T computer 0:02:46.840 --> 0:02:50.680 topics and activities. It's kind of overwhelming, to be honest. 0:02:50.800 --> 0:02:55.120 In fact, I am now kind of unwinding in my 0:02:55.200 --> 0:02:58.359 hotel room before I have to go back out and 0:02:58.840 --> 0:03:01.040 jump right back into I'm going to go to a 0:03:01.120 --> 0:03:05.600 session where they're going to talk about five big topics 0:03:05.600 --> 0:03:08.400 in science and technology. So I'm really looking forward to it. 0:03:08.440 --> 0:03:12.160 There's some very interesting people who are here presenting, and 0:03:12.200 --> 0:03:16.080 I can't wait to learn more. Some of the areas 0:03:16.120 --> 0:03:20.120 that we're going to talk about this week will include blockchain. Now, 0:03:20.160 --> 0:03:22.480 if you've been listening to text stuff over the past 0:03:22.520 --> 0:03:25.919 few months, you've probably heard a recent episode I did 0:03:26.040 --> 0:03:33.520 about blockchain. Blockchain is the technology that underlies cryptocurrencies like bitcoin, 0:03:33.600 --> 0:03:35.920 and in fact, I would argue bitcoin is probably the 0:03:35.960 --> 0:03:41.440 most famous use of blockchain technology, but blockchain is not 0:03:42.080 --> 0:03:48.960 only good for bitcoins and cryptocurrencies. That's one use for it, 0:03:49.000 --> 0:03:50.880 but not the only one. In fact, a lot of 0:03:50.920 --> 0:03:55.200 people argue that blockchain is going to be the next 0:03:56.680 --> 0:04:01.320 um next evolution of the web. So you might remember 0:04:01.680 --> 0:04:04.880 people were talking about web two point oh more than 0:04:04.920 --> 0:04:08.840 a decade ago, and blockchain would be like the next step. 0:04:08.920 --> 0:04:12.040 So what does that even mean? Well, what one point 0:04:12.040 --> 0:04:16.279 oh was the sort of websites you first saw, primarily 0:04:16.480 --> 0:04:20.760 when the web first launched, that would be websites that 0:04:20.760 --> 0:04:24.000 were pretty much static. They did not change, They were 0:04:24.040 --> 0:04:28.800 not really interactive. They did not have any capacity to 0:04:28.960 --> 0:04:33.839 have users add to or change things in any meaningful way. 0:04:34.080 --> 0:04:38.159 And so it was kind of like looking at a 0:04:38.240 --> 0:04:41.919 newspaper or a magazine, something that's in a fixed format 0:04:42.480 --> 0:04:47.839 that doesn't have any real um, any real, any real 0:04:47.960 --> 0:04:52.039 special thing about it. Right, there's nothing that gives it 0:04:52.120 --> 0:04:56.640 any sort of uh sense that you your presence matters, 0:04:56.720 --> 0:04:59.960 or that that's even registering at all, unless there was 0:05:00.040 --> 0:05:03.760 one of those little helpful counters, as was often the 0:05:03.760 --> 0:05:05.840 case in early websites, where it would tell you how 0:05:05.839 --> 0:05:09.120 many visitors had been to that website above. From that, 0:05:09.160 --> 0:05:12.960 there really wasn't anything that indicated that you mattered at all. 0:05:13.160 --> 0:05:15.160 So that was web one point. Oh and a lot 0:05:15.200 --> 0:05:21.120 of those websites ended up kind of just being uh momentary. 0:05:21.200 --> 0:05:24.240 They didn't stick around because there was no point in 0:05:24.240 --> 0:05:26.560 going back and visiting after you've seen them once. They 0:05:26.680 --> 0:05:29.839 weren't ever going to change. If they did change, you 0:05:29.880 --> 0:05:32.680 didn't necessarily know about it, and so you would just 0:05:32.720 --> 0:05:34.839 have to go back and visit and see if anything 0:05:34.880 --> 0:05:36.440 had changed since the last time you were there. It 0:05:36.480 --> 0:05:41.320 was kind of inconvenient. Web two point oh arguably was 0:05:41.800 --> 0:05:47.200 when websites started to incorporate interactive features where users could 0:05:47.240 --> 0:05:50.960 come into the website and things would change, they'd be dynamic. 0:05:51.320 --> 0:05:55.440 Sometimes this meant that there were animated aspects to the 0:05:55.440 --> 0:05:57.800 web page. Sometimes it just meant that there was a 0:05:57.800 --> 0:06:01.200 way for users to leave feedback. For example, some people 0:06:01.240 --> 0:06:05.320 point to Amazon as being an early web dew point 0:06:05.360 --> 0:06:10.039 oh style web page because you could leave user reviews 0:06:10.680 --> 0:06:14.080 on the site, so you could actually impact what happened 0:06:14.279 --> 0:06:17.559 with products on Amazon dot Com. And then the course 0:06:17.640 --> 0:06:22.520 later on, Amazon was also incorporating things like recommendation engines 0:06:22.600 --> 0:06:28.000 that would try to engage users and and encourage them 0:06:28.040 --> 0:06:31.599 to spend more money and to buy more things, and 0:06:31.720 --> 0:06:35.600 that was also and indication that was web two point oh. 0:06:35.600 --> 0:06:38.640 It was something that was more than just visit website, 0:06:38.800 --> 0:06:42.080 read some information and leave. Blockchain is supposedly going to 0:06:42.200 --> 0:06:46.200 be the next step, and really it comes down to 0:06:47.440 --> 0:06:50.400 the very nature of blockchain itself, which is a peer 0:06:50.400 --> 0:06:54.200 to peer technology within a network. So you're not necessarily 0:06:54.240 --> 0:06:57.520 talking about Internet wide. It's a network within the Internet. 0:06:58.040 --> 0:07:01.080 Now it might be a network that spans multip pole networks, 0:07:01.240 --> 0:07:04.280 because it's all peers that are connecting to one another 0:07:04.680 --> 0:07:10.400 and using a method that creates blocks of data. Thus 0:07:10.480 --> 0:07:13.080 the block in the block chain, and each block is 0:07:13.400 --> 0:07:16.760 sequentially added to a chain of other blocks, and it 0:07:16.800 --> 0:07:20.000 all dates back to the very first block in the chain, 0:07:20.160 --> 0:07:23.520 and there's information within each block that can be traced 0:07:23.560 --> 0:07:26.760 all the way back to that first one. Uh. The 0:07:26.800 --> 0:07:31.240 transactions that happen within a block are recorded within that block, 0:07:32.000 --> 0:07:36.360 and as other computers inside the peer to peer network 0:07:37.240 --> 0:07:42.560 verify and validate those transactions, whatever they may be, they 0:07:42.600 --> 0:07:45.680 then are able to create the next block in the blockchain. 0:07:45.760 --> 0:07:49.280 With cryptocurrency, you get a reward for this. That's what's 0:07:49.280 --> 0:07:54.080 called mining. It's when in bitcoin your computer participates in 0:07:54.280 --> 0:07:59.480 solving a particularly tough math problem essentially, and if you 0:07:59.560 --> 0:08:05.119 do end that ends up being the validation for previous transactions. 0:08:05.600 --> 0:08:09.680 The transactions all are codified as a block added to 0:08:09.720 --> 0:08:13.640 the end of the chain, and everyone in the peer 0:08:13.640 --> 0:08:16.760 to peer network has a has access to a ledger 0:08:17.080 --> 0:08:20.600 that is updated across the entire network. So the ledger 0:08:20.920 --> 0:08:23.480 is the full record of all transactions dating back to 0:08:23.520 --> 0:08:26.720 the very first one. They're going to be very deep 0:08:26.760 --> 0:08:30.440 dives on blockchain technology here this week, as people talk 0:08:30.480 --> 0:08:34.240 about the different ways to use it beyond cryptocurrencies, and 0:08:34.480 --> 0:08:41.080 how it could serve as a backbone to future web interactions. Honestly, 0:08:41.120 --> 0:08:44.160 I'm very eager to hear more about that, because while 0:08:44.200 --> 0:08:47.400 I've just described kind of a very high level of 0:08:47.400 --> 0:08:52.400 what blockchain is, I I am really curious about how 0:08:52.679 --> 0:08:57.400 this is going to be used to actually, uh form 0:08:57.840 --> 0:09:01.840 the the spine of the web in the future. Was 0:09:01.880 --> 0:09:04.400 it going to be besides a way of keeping track 0:09:04.440 --> 0:09:09.760 of transactions, like whether it's currencies or property or anything 0:09:09.760 --> 0:09:12.920 along those lines, what can it do beyond that? And 0:09:12.960 --> 0:09:15.760 I honestly don't fully understand that, so I look forward 0:09:15.760 --> 0:09:18.560 to learning more while I'm here. That's just one of 0:09:18.600 --> 0:09:22.520 the aspects that will be covered during this conference. Another one, 0:09:22.640 --> 0:09:26.360 a big one, is artificial intelligence. Now I've talked about 0:09:26.440 --> 0:09:30.040 that a lot on tech stuff as well. Artificial intelligence 0:09:30.040 --> 0:09:32.280 A lot of people when they hear that, I think 0:09:32.360 --> 0:09:36.240 they immediately go to the idea of a machine that 0:09:36.360 --> 0:09:41.400 can quote unquote think like a person. It is mimicking 0:09:41.440 --> 0:09:46.880 the way brains think, or it somehow possesses human like intelligence, 0:09:46.960 --> 0:09:50.600 or perhaps even superhuman intelligence, where it's able to think 0:09:50.600 --> 0:09:54.160 even better than humans are. But artificial intelligence is a 0:09:54.280 --> 0:09:57.400 much more broad category than that. That would be a 0:09:57.520 --> 0:10:03.400 very specific niche definition of what artificial intelligence is. Artificial 0:10:03.440 --> 0:10:10.000 intelligence encapsulates multiple disciplines, and it is also uh something 0:10:10.040 --> 0:10:14.880 that involves lots and lots of different parts of intelligence, 0:10:14.920 --> 0:10:18.200 not just this cognition that we think of, but think 0:10:18.240 --> 0:10:22.880 of things like image recognition with computers. Teaching a computer 0:10:23.000 --> 0:10:28.640 to identify a specific kind of thing to extrapolate from knowledge, 0:10:29.120 --> 0:10:31.920 it's tricky. You might be able to teach a computer, 0:10:32.000 --> 0:10:36.200 for example, that a mug is a mug like a 0:10:36.240 --> 0:10:40.320 coffee mug. You've got a red coffee mug has got 0:10:40.520 --> 0:10:44.840 sort of a curved cut mug. It's not, you know, 0:10:44.960 --> 0:10:47.199 just straight up and down. It's got kind of kind 0:10:47.200 --> 0:10:50.720 of bowls out a little bit. And you take images 0:10:50.760 --> 0:10:53.280 of it from a certain angle with certain lighting, and 0:10:53.400 --> 0:10:56.600 you feed it to a computer and you let them 0:10:56.760 --> 0:10:59.760 You have your computer breakdown the image so it can 0:10:59.800 --> 0:11:02.760 I identify where the all the borders are, what are 0:11:02.760 --> 0:11:05.440 what is the mug versus, what is the background versus 0:11:05.520 --> 0:11:09.520 what is the platform it's on? And you Essentially, you 0:11:09.600 --> 0:11:12.840 teach the computer, hey, this is a mug. Computers don't 0:11:12.880 --> 0:11:18.400 magically then understand that all other containers that have that 0:11:18.559 --> 0:11:22.760 same basic sort of shape um are mugs. If you 0:11:23.000 --> 0:11:26.840 showed another picture of that same mug with different lighting, 0:11:27.360 --> 0:11:30.120 the computer might not be able to tell you that 0:11:30.120 --> 0:11:32.760 that's a mug. Or it's at a different angle. Maybe 0:11:32.920 --> 0:11:35.080 you turn it so that the handle is facing a 0:11:35.080 --> 0:11:37.920 different way, and then you showed the computer the image, 0:11:37.960 --> 0:11:39.800 it may not be able to figure out that that 0:11:39.960 --> 0:11:42.920 is also a mug. If it's a different color, if 0:11:42.960 --> 0:11:46.840 it's a different shape. All of these things are tricky 0:11:46.880 --> 0:11:50.080 for machines. Like as human beings, we can see an 0:11:50.120 --> 0:11:55.040 example or two of something and then we're pretty good. 0:11:55.440 --> 0:11:57.959 We then can say, all right, I now I get 0:11:58.000 --> 0:12:01.320 the general idea of the things that that are the 0:12:01.400 --> 0:12:05.680 constitute a mug. And when I encounter something else, even 0:12:05.720 --> 0:12:08.320 if it is of a different shape, a different color, 0:12:08.520 --> 0:12:12.120 different lighting, on a different surface, maybe it's even got 0:12:12.280 --> 0:12:15.400 a different type of liquid inside of it, I still 0:12:15.440 --> 0:12:18.559 can figure out that that thing is a mug. Because 0:12:18.800 --> 0:12:23.640 I can extrapolate. I can take what I've learned and 0:12:23.720 --> 0:12:28.360 extend it beyond just the few instances that I have encountered. 0:12:28.880 --> 0:12:31.880 And then if I walk into a mug store, I 0:12:31.960 --> 0:12:34.199 don't look at just one and say that's a mug, 0:12:34.240 --> 0:12:35.880 but I have no idea what the rest of these 0:12:35.880 --> 0:12:38.880 things are. I can actually say, oh, these are all 0:12:38.960 --> 0:12:42.120 different types of mugs and different sizes and shapes and colors. 0:12:42.360 --> 0:12:45.200 Computers are not good at that generally speaking. This is 0:12:45.440 --> 0:12:47.760 one of the reasons why you may have heard about 0:12:47.760 --> 0:12:51.959 that story of a computer being fed thousands upon thousands 0:12:51.960 --> 0:12:56.640 of images of cat pictures so that the computer could 0:12:56.720 --> 0:13:00.640 learn what a cat is. Because without all of that, 0:13:01.040 --> 0:13:06.600 without the this huge body of examples, the computer simply 0:13:06.640 --> 0:13:10.840 can't extrapolate and figure out what is is not a cat. 0:13:11.120 --> 0:13:14.600 And even after all of those images were fed to 0:13:14.640 --> 0:13:17.320 the computer, it's still had some issues. It wasn't like 0:13:17.400 --> 0:13:20.960 it had magically understood what a cat was. It did 0:13:20.960 --> 0:13:24.319 not have an innate grasp of cat nous, and by 0:13:24.360 --> 0:13:26.800 that I mean the qualities that make a cat not 0:13:27.320 --> 0:13:32.080 a character from Hunger games. So artificial intelligence, again is 0:13:32.280 --> 0:13:36.800 very much a multidisciplinary thing, image recognition being just one 0:13:36.960 --> 0:13:41.080 tiny part of it. There are numerous other aspects to 0:13:41.200 --> 0:13:48.040 artificial intelligence, including things like voice recognition and natural language processing. 0:13:48.280 --> 0:13:52.480 Natural language processing, of course, is where a computer starts 0:13:52.520 --> 0:13:55.640 to learn what we mean when we say things a 0:13:55.679 --> 0:14:00.120 certain way. Now, again, just like with our ability to 0:14:00.200 --> 0:14:04.840 extrapolate based upon the limited examples we have seen of 0:14:04.880 --> 0:14:08.719 any one object, with image recognition, natural language, we can 0:14:08.720 --> 0:14:12.480 have different ways of saying the same thing and still 0:14:12.600 --> 0:14:17.480 mean the same meaning. So I might say it's raining, 0:14:17.840 --> 0:14:21.080 or it's pouring outside, or it's coming down like cats 0:14:21.080 --> 0:14:23.680 and dogs. Those are all different ways of me saying 0:14:23.680 --> 0:14:25.760 that water is falling from the sky in the form 0:14:25.800 --> 0:14:29.080 of precipitation, and you will get wet if you walk outside. 0:14:29.680 --> 0:14:32.920 And people who have had just a limited amount of 0:14:33.480 --> 0:14:36.520 exposure to these sort of ideas can pick up on that, 0:14:36.640 --> 0:14:40.200 but computers, again do not unless you expressly tell them. So. 0:14:41.080 --> 0:14:44.400 The way we say things, the word order we choose, 0:14:44.480 --> 0:14:48.840 the syntax, the the accent we may speak with, the 0:14:48.880 --> 0:14:52.440 dialect if you prefer that we might speak with, the 0:14:52.840 --> 0:14:57.000 emphasis we would place on different words, the speed at 0:14:57.040 --> 0:14:59.880 which we speak. All of these different factors make it 0:15:00.080 --> 0:15:03.760 very challenging for computers to understand us. That being said, 0:15:04.240 --> 0:15:07.120 voice recognition and natural language processing has come a far 0:15:07.200 --> 0:15:10.000 away in a short amount of time. Over the last 0:15:10.040 --> 0:15:12.840 couple of decades, it has really advanced quite a bit, 0:15:12.880 --> 0:15:15.600 which is why we see all of these digital personal 0:15:15.600 --> 0:15:19.240 assistance that we can talk to like Siri and Alexa 0:15:19.440 --> 0:15:24.040 and Google Assistant, and they have become pretty good and 0:15:24.200 --> 0:15:27.280 understanding us. They're not perfect, but they're pretty good at 0:15:27.280 --> 0:15:30.000 figuring out what we want, even if we say things 0:15:30.080 --> 0:15:33.480 in different ways. But that's another aspect of artificial intelligence. 0:15:33.800 --> 0:15:35.600 All of these sort of things are the kind of 0:15:35.600 --> 0:15:38.560 topics that will be talked about during the IBM Think Conference, 0:15:38.920 --> 0:15:41.600 And again they're gonna go into much greater detail and 0:15:41.640 --> 0:15:46.920 talk about real world applications for AI, not just theory, 0:15:47.400 --> 0:15:50.640 not just how far we've come, but how we can 0:15:50.680 --> 0:15:53.880 put it to use. There's more to come about the 0:15:53.920 --> 0:15:58.240 IBM Think Conference, but before I get much further into it, 0:15:58.320 --> 0:16:09.320 let's take a quick break to and our sponsor IBM Watson. 0:16:09.800 --> 0:16:12.480 You may have heard of it as the computer that 0:16:12.640 --> 0:16:16.560 defeated former Jeopardy champions. Well, it's much more than that. 0:16:17.080 --> 0:16:20.520 Although that definitely was a big publicity stunt. You could 0:16:20.560 --> 0:16:24.440 argue it was the equivalent of when Deep Blue went 0:16:24.560 --> 0:16:30.000 up against Kasparov in the chess competitions. But IBM s 0:16:30.000 --> 0:16:34.720 Watson is sort of in that natural language processing and 0:16:34.960 --> 0:16:37.960 understanding what people are saying and being able to pull 0:16:38.160 --> 0:16:41.520 data based on that. That is sort of an example 0:16:41.560 --> 0:16:44.120 of that. So in Jeopardy it was really put to 0:16:44.160 --> 0:16:48.600 the test because with Jeopardy, you don't just get clues 0:16:48.760 --> 0:16:51.760 and then you have to come up with the right answer. 0:16:52.040 --> 0:16:56.720 Sometimes those clues are puns or rhymes, or they are 0:16:56.920 --> 0:17:01.760 very circumspect, like it's very circuitous the kind of logic 0:17:01.840 --> 0:17:04.159 you have to use to figure out what the actual 0:17:04.240 --> 0:17:06.720 answer is in the form of a question, and Watson 0:17:06.760 --> 0:17:09.359 did really well with it. You may remember that the 0:17:09.359 --> 0:17:12.240 way Watson worked was that it had a massive database. 0:17:12.280 --> 0:17:14.560 It was not connected to the Internet for the purposes 0:17:14.600 --> 0:17:17.639 of Jeopardy. It just had a big database of information 0:17:18.359 --> 0:17:23.760 and it would process whatever the clue was. It would 0:17:23.840 --> 0:17:28.399 reference its database, and it would look for potential answers 0:17:28.480 --> 0:17:32.199 and assign each potential answer a probability, sort of a 0:17:32.240 --> 0:17:37.400 confidence probability. If Watson's confidence probability probability was high enough, 0:17:37.440 --> 0:17:39.280 and as I recall, it was somewhere in the eight 0:17:39.480 --> 0:17:43.400 percentile range, like it had to be at least sure 0:17:43.440 --> 0:17:45.720 that that was the right answer. It would then put 0:17:45.800 --> 0:17:49.800 forward that as its answer in jeopardy. If none of 0:17:49.800 --> 0:17:52.960 the answers that came up with met that level of confidence, 0:17:53.280 --> 0:17:56.800 Watson would not answer. It would not attempt to see 0:17:57.240 --> 0:18:00.119 if perhaps it could take a wild guess and at 0:18:00.160 --> 0:18:02.199 the right answer, and it turned out to work. Watson 0:18:02.280 --> 0:18:05.120 ended up winning that game of Jeopardy. But of course, 0:18:05.119 --> 0:18:09.919 Watson isn't just about showing up former champions. It's about 0:18:10.480 --> 0:18:15.560 all sorts of stuff IBM SUH. Early applications of Watson 0:18:15.600 --> 0:18:20.679 were to include it with medical procedures so that it 0:18:20.720 --> 0:18:25.800 could help doctors diagnose and treat patients. It was not 0:18:25.960 --> 0:18:30.159 replacing doctors. It was not meant to be a robo 0:18:30.320 --> 0:18:34.360 doctor that would treat you based on your symptoms and 0:18:34.400 --> 0:18:37.200 then you just go to the robo pharmacist. But rather 0:18:37.359 --> 0:18:40.920 it was acting as an assistant to help either confirm 0:18:41.040 --> 0:18:44.560 what a doctor was thinking or perhaps narrowed down some 0:18:44.640 --> 0:18:49.680 options of what the cause of any particular symptoms might be. 0:18:50.200 --> 0:18:53.439 And it has seen a lot of use in that field. 0:18:54.000 --> 0:18:57.520 It's gone well beyond that, though some of the other 0:18:57.760 --> 0:19:00.880 applications have been a little silly. You may remember if 0:19:00.880 --> 0:19:04.120 you listen to the Forward Thinking podcast We did an 0:19:04.119 --> 0:19:09.480 episode about Chef Watson that was an application that IBM 0:19:09.520 --> 0:19:15.000 created where Watson would design a meal for you based 0:19:15.040 --> 0:19:18.800 off of ingredients you told it. We're at your disposal. 0:19:19.040 --> 0:19:22.120 So you might say, hey, I have chicken, and I've 0:19:22.160 --> 0:19:25.840 got rosemary, and i have some potatoes, and I've got 0:19:25.960 --> 0:19:30.199 some rice, and i've got some green beans. What can 0:19:30.280 --> 0:19:32.879 I make with these? And it would say, all right, well, 0:19:32.960 --> 0:19:34.720 let me come up with a recipe for you, and 0:19:34.760 --> 0:19:37.880 would generate a recipe and it probably includes some ingredients 0:19:37.880 --> 0:19:39.679 that you might not have on hand. It would be 0:19:39.680 --> 0:19:41.679 stuff that you would have to go shop for. But 0:19:41.720 --> 0:19:45.320 the interesting thing was it was dynamically creating these recipes. 0:19:45.359 --> 0:19:49.960 It wasn't accessing a database of recipes and pulling from 0:19:50.000 --> 0:19:54.280 the ones that included the ingredients you mentioned. Instead, it 0:19:54.320 --> 0:19:58.840 would look at how flavors had been combined in the 0:19:58.880 --> 0:20:04.439 past by looking at a huge library of recipes that 0:20:04.520 --> 0:20:08.399 had been fed to Chef Watson. So just imagine a 0:20:08.560 --> 0:20:11.520 library is worth of cookbooks fed to this computer, and 0:20:11.560 --> 0:20:14.080 the computer says, well, based upon what I can see 0:20:14.320 --> 0:20:18.280 from all these recipes, roseberry and chicken do pair well together, 0:20:18.359 --> 0:20:20.760 So we're going to make a recipe that uses those 0:20:20.760 --> 0:20:24.640 two ingredients, and based upon all the different ways I've 0:20:24.640 --> 0:20:28.160 seen to prepare chicken, I think a baked chicken dish 0:20:28.200 --> 0:20:30.760 would be the perfect one to go with, and so 0:20:30.800 --> 0:20:33.760 on and so forth. But it's not pulling a recipe, 0:20:34.119 --> 0:20:37.199 it was creating one based upon its knowledge. It's like 0:20:37.240 --> 0:20:41.919 a chef, not a cook. Uh. This did not always 0:20:41.960 --> 0:20:45.480 necessarily work out great. If you listen to that episode 0:20:45.480 --> 0:20:48.720 of Forward Thinking, you'll hear some of the experiences we 0:20:48.760 --> 0:20:52.600 had as we were testing out Chef Watson. I remember 0:20:52.640 --> 0:20:56.840 a cauliflower fricacy recipe in which the cauliflower was listed 0:20:56.880 --> 0:21:01.199 as optional, which I thought was a bit unusual. Also, Uh, 0:21:01.359 --> 0:21:03.439 the other interesting thing was that you could feed the 0:21:03.480 --> 0:21:07.480 exact same ingredients into Chef Watson multiple times and you'll 0:21:07.520 --> 0:21:10.680 get different results each time. Again because the recipes are 0:21:10.680 --> 0:21:16.160 created dynamically, so it's not again not not looking at 0:21:16.400 --> 0:21:19.440 a recipe it's already written. It's writing a new one 0:21:19.480 --> 0:21:21.720 for you each time you ask it. It's kind of interesting. 0:21:22.520 --> 0:21:25.840 There was also a an example recently where Watson was 0:21:25.920 --> 0:21:30.280 being used as a platform for the Weather Company, where 0:21:30.280 --> 0:21:33.720 if you use the Weather Company's app, Watson was helping 0:21:33.760 --> 0:21:36.480 power that so it's kind of like an an A 0:21:36.600 --> 0:21:41.679 p I. So different developers can use Watson as a 0:21:41.680 --> 0:21:46.119 platform to build different applications, and it all depends on 0:21:46.160 --> 0:21:48.720 what your application needs whether or not Watson is a 0:21:48.720 --> 0:21:51.959 good fit. It doesn't necessarily mean that every single application 0:21:52.560 --> 0:21:55.360 is going to benefit from using Watson. If you want 0:21:55.400 --> 0:21:58.679 to create a game that's like Angry Birds, Watson may 0:21:58.720 --> 0:22:01.480 not be of much use. But if it's anything where 0:22:01.920 --> 0:22:07.320 you have someone asking questions or asking for data, then 0:22:07.560 --> 0:22:10.800 Watson might end up being helpful. And so that Obviously, 0:22:10.840 --> 0:22:15.800 there's a lot of different roundtable discussions and breakout sessions 0:22:15.840 --> 0:22:18.560 all about IBM Watson and how to use it. One 0:22:18.640 --> 0:22:22.199 thing that I think is gonna be particularly interesting in 0:22:22.200 --> 0:22:27.080 a very sensitive subject given extremely recent news, is the 0:22:27.119 --> 0:22:31.840 idea of self driving cars. There are some sessions that 0:22:31.880 --> 0:22:34.760 are supposed to be about self driving cars. The reason 0:22:34.840 --> 0:22:38.000 I say sensitive is because the day I'm recording this, 0:22:38.119 --> 0:22:42.040 on March nineteen, two thousand eighteen, UH there was also 0:22:43.000 --> 0:22:46.960 very tragic news out of Arizona, and that news was 0:22:47.000 --> 0:22:53.240 that Uber uh has a fleet of self driving vehicles 0:22:53.240 --> 0:22:56.280 that they've been testing in different markets, including in Arizona, 0:22:56.440 --> 0:22:59.600 and one of those self driving vehicles and SUV ended 0:22:59.680 --> 0:23:05.240 up striking and killing a pedestrian, Elaine Hertzberg. She was 0:23:05.359 --> 0:23:07.920 trying to cross the street. She was walking a bicycle 0:23:08.000 --> 0:23:11.520 across the street in Tempe, Arizona, on the night of 0:23:11.640 --> 0:23:15.040 March eighteen, two thousand eighteen, and she was struck by 0:23:15.040 --> 0:23:18.840 this SUV. And the SUV was an autonomous mode. There 0:23:19.000 --> 0:23:23.760 was a driver in the vehicle because while Uber has 0:23:23.760 --> 0:23:28.639 been testing this autonomous vehicle technology, they have been asked 0:23:28.640 --> 0:23:31.280 to keep an operator behind the wheel of the car 0:23:31.359 --> 0:23:34.679 to take over just in case something goes wrong and 0:23:34.760 --> 0:23:39.720 the vehicle is unable to cope with it. And I'm 0:23:39.760 --> 0:23:42.639 as of the recording of this podcast, I'm not sure 0:23:42.680 --> 0:23:46.480 exactly what happened, whether or not the driver was aware 0:23:46.840 --> 0:23:49.520 of what was about to happen, was unable to respond 0:23:49.560 --> 0:23:53.679 in time. I don't know what the sequence of events was, 0:23:53.880 --> 0:23:59.280 specifically in regards to that operator. But the tragic story 0:23:59.520 --> 0:24:03.560 is that this autonomous car, the self driving car, struck 0:24:03.600 --> 0:24:07.480 and killed a pedestrian. And as far as I can determine, 0:24:07.840 --> 0:24:12.280 it is the first fatality due to an autonomous car. 0:24:12.520 --> 0:24:14.840 Obviously not the first accident. There have been a few 0:24:14.880 --> 0:24:18.399 others and a few that have been attributed specifically to 0:24:18.560 --> 0:24:22.520 the autonomous cars and not to human drivers. But this 0:24:22.720 --> 0:24:27.040 is a terrible, terrible story, and obviously that's going to 0:24:27.400 --> 0:24:31.680 affect the conversations that go on here, So I hope 0:24:31.720 --> 0:24:35.480 to attend some of those and hear what experts in 0:24:35.480 --> 0:24:38.520 the field have to say about this and what is 0:24:38.560 --> 0:24:40.760 the best course of action. Obviously, you want to be 0:24:40.840 --> 0:24:45.800 respectful of the victim and her family, and you want 0:24:45.840 --> 0:24:49.399 to be realistic. You do not want to dismiss this. 0:24:49.520 --> 0:24:53.160 Obviously that would be horrible, it'd be unthinkable. But how 0:24:53.200 --> 0:24:59.240 do you move forward when there's so much momentum technologically 0:24:59.280 --> 0:25:03.360 speaking behind the movement to go to autonomous cars? And 0:25:03.600 --> 0:25:07.720 I'm I want to find out what people's thoughts are 0:25:07.720 --> 0:25:10.879 on this. It may very well be that there aren't 0:25:10.920 --> 0:25:15.199 any prepared sessions to cover this because the news is 0:25:15.280 --> 0:25:18.840 so recent, but I'm sure there will be questions about it. 0:25:19.240 --> 0:25:21.199 There's a bit more I expect to see here at 0:25:21.240 --> 0:25:25.480 the IBM Think Conference, and I'll tell you about it 0:25:25.600 --> 0:25:28.080 in just a minute, but I've got to take a 0:25:28.160 --> 0:25:39.639 quick break to thank our sponsor. Another area that is 0:25:39.680 --> 0:25:41.919 going to get a lot of coverage here at the 0:25:42.040 --> 0:25:46.080 conference is all about cloud and data issues, and oh boy, 0:25:46.240 --> 0:25:50.960 there's a lot to talk about there too, because again, 0:25:51.119 --> 0:25:58.200 recent news has been pretty rough in regards to data mining. 0:25:58.880 --> 0:26:04.000