WEBVTT - A List Of Games You Would Surely Lose to a Computer 0:00:01.120 --> 0:00:04.560 Welcome to stuff you should know from how Stuff Works 0:00:04.600 --> 0:00:13.920 dot com. Hey, and welcome to the podcast. I'm Josh Clark. 0:00:14.200 --> 0:00:18.720 There's Charles w Chuck Bryant, there's Jerry over there. I'm 0:00:18.720 --> 0:00:20.840 just gonna come out and tell everybody making fun of 0:00:20.880 --> 0:00:26.120 me for some weird reason, vaguely weird ways. But I'm 0:00:26.120 --> 0:00:28.400 all right. So, Chuck, I have a story for you. 0:00:29.560 --> 0:00:33.479 I'm gonna take us back to the seventeen seventies and 0:00:33.560 --> 0:00:39.519 the swinging town of Vienna, not Virginia, not Vanna Georgia, 0:00:40.120 --> 0:00:42.839 but you know that's how they pronounce it, right, Vanna 0:00:43.280 --> 0:00:49.440 Yanna sausages, right, Vienna, Austria. You ever been there? Vienna, Austria. No, 0:00:49.880 --> 0:00:53.640 I've been to Brussels. That was pretty close. Vienna's lovely, 0:00:53.840 --> 0:00:57.800 I'm sure. I think it's a lot like Brussels. Very clean, 0:00:57.920 --> 0:01:01.480 lovely town. I just remember to be very clean, Yeah, 0:01:01.600 --> 0:01:06.280 very clean, gorgeous architecture, weird little angled side streets. They're 0:01:06.400 --> 0:01:10.840 very narrow, very pretty tone. So we're in Vienna and 0:01:10.920 --> 0:01:15.360 there is a dude skulking about going to the royal 0:01:15.520 --> 0:01:21.720 Palace in Vienna. His name is Wolfgang von Kimberlin, and 0:01:21.800 --> 0:01:25.479 he's an inventor. He's an engineer. He's a pretty sharp dude, 0:01:25.480 --> 0:01:27.680 and he's got with him what would come to be 0:01:27.800 --> 0:01:31.360 known as the Turk, but he called it the mechanical 0:01:31.440 --> 0:01:36.240 Turk or the automaton chess player, and that's what it was. 0:01:36.880 --> 0:01:40.800 It was a It was a wooden figure that moved mechanically, 0:01:41.240 --> 0:01:44.560 seated at a cabinet and on top of the cabinet 0:01:44.920 --> 0:01:47.840 was a chessboard. And when he brought it out to 0:01:47.920 --> 0:01:52.960 show to the Royal Court, he would um, it was cool, 0:01:53.240 --> 0:01:57.640 kind of but nothing they hadn't seen before, because automata 0:01:57.920 --> 0:02:01.320 was kind of a hip thing by that. Yeah, people 0:02:01.400 --> 0:02:08.360 loved building these engineering these automatum machines to do various things. 0:02:08.720 --> 0:02:11.000 And people are just knocked out by the fact that, 0:02:11.880 --> 0:02:15.079 you know, you hide these gears and levers behind wood 0:02:15.160 --> 0:02:17.720 or a cloth and it looks as though there's a 0:02:18.320 --> 0:02:19.799 real well I'm not real, but you know what I 0:02:19.840 --> 0:02:23.200 mean is that it's like a real machine. Yeah, but not. 0:02:23.360 --> 0:02:25.640 They weren't fooled anything like is that a real man? 0:02:26.280 --> 0:02:29.080 It was, but it was for their time. It was 0:02:29.160 --> 0:02:33.200 so advanced looking that it's like us seeing ex machina 0:02:33.840 --> 0:02:37.160 in the movie theater that makes sense. Yeah, no, it 0:02:37.200 --> 0:02:39.520 does make sense. But imagine seeing like x Makina and 0:02:39.600 --> 0:02:44.639 being like I've seen this before. This isn't anything special, okay. Yeah, 0:02:44.720 --> 0:02:47.600 and this thing, to be clear, look like a is 0:02:47.639 --> 0:02:53.720 it Zoltar or Zoltan from Big Zoltar Sultan? I don't know, 0:02:53.760 --> 0:02:55.600 it's one of those two, one of those two. Like this, 0:02:55.720 --> 0:02:58.480 this guy's wearing a turban and it's in a glass 0:02:58.520 --> 0:03:03.359 case like bust like you know, like a chest up thing. Yeah, 0:03:03.400 --> 0:03:06.040 he's seated at this this cabinet, so there's no need 0:03:06.120 --> 0:03:08.520 for legs or anything like that. Yeah. But the thing, 0:03:08.600 --> 0:03:12.600 this is what was amazing about the Turk. He could 0:03:12.680 --> 0:03:16.919 play chess, and he could play chess really well. So yeah, 0:03:16.960 --> 0:03:19.760 he was like an automaton and he moved all herky 0:03:19.880 --> 0:03:23.480 jerky or whatever. But he could play you in chess, 0:03:23.760 --> 0:03:29.080 which was a huge, huge advance at the time. Like, 0:03:29.160 --> 0:03:33.160 this is something that wouldn't come up again until the nineties, 0:03:33.280 --> 0:03:36.800 more than two hundred years later. This thing, this automaton 0:03:37.120 --> 0:03:40.480 could play a human being in chess and beat them. Well, yeah, 0:03:40.480 --> 0:03:42.840 and it looked like when the game started, it would 0:03:42.880 --> 0:03:45.480 look down at the chessboard and like cock his head 0:03:45.560 --> 0:03:48.360 like m what should my first move be? And if 0:03:48.400 --> 0:03:50.960 people I love this part, if people tried to cheat. 0:03:51.600 --> 0:03:54.720 Apparently Napoleon tried to cheat this thing because this guy 0:03:55.240 --> 0:03:58.840 he debuted at the Being's court, but then it, you know, 0:03:58.840 --> 0:04:01.320 it went on a world tour. Yeah, and he was 0:04:01.360 --> 0:04:03.920 even it was taken over by his successor to the 0:04:03.960 --> 0:04:06.440 guy who toured with it. Even further, people went nuts 0:04:06.440 --> 0:04:08.240 for this stuff. They did. They loved it because they 0:04:08.240 --> 0:04:10.760 were like, this is crazy. I can't believe what I'm seeing. 0:04:10.960 --> 0:04:13.080 Most people, though, were not taken in by it. There 0:04:13.120 --> 0:04:16.320 like there's some trick here. But von kempel In and 0:04:16.400 --> 0:04:19.200 the guy who came after him, I don't remember his name. Um, 0:04:19.240 --> 0:04:22.360 they would they would demonstrate. You could open this cabinet 0:04:22.360 --> 0:04:25.640 and you could see all the workings of the mechanical 0:04:25.680 --> 0:04:29.240 turk inside. Right. So what I was saying is, if 0:04:29.240 --> 0:04:34.440 this thing's since a cheater like Napoleon supposedly did it, 0:04:34.520 --> 0:04:37.000 would you know Napoleon would move a piece out of 0:04:37.000 --> 0:04:40.599 turner illegally or something. This dude, the turk Turk one 0:04:40.920 --> 0:04:43.119 two would pick up the chess piece move it back 0:04:43.200 --> 0:04:45.719 as if to say like, no, no, Napoleon, I see 0:04:45.720 --> 0:04:47.920 what you're doing. And then if the person attempted to 0:04:47.960 --> 0:04:49.840 move it again, I don't know how many times on 0:04:49.960 --> 0:04:53.040 you two or three times, eventually it would just go 0:04:53.240 --> 0:04:56.599 ah and wipe his hand across the board and knock 0:04:56.600 --> 0:04:59.480 off all the pieces. Game over, which is pretty great, 0:05:00.040 --> 0:05:03.600 A nice little feature, yeah it is. But it even 0:05:03.640 --> 0:05:06.960 showed even more that this thing was thinking for itself. 0:05:07.640 --> 0:05:10.760 That's the key here, right. Chess had been for a 0:05:10.839 --> 0:05:15.520 very long time viewed as only something that a human 0:05:16.000 --> 0:05:18.800 would be capable of because it took a human intellect. 0:05:19.160 --> 0:05:22.600 And there was actually a guy in English um engineer, 0:05:23.120 --> 0:05:26.640 I think he was a mechanical engineer. His name was 0:05:27.400 --> 0:05:30.800 Robert Willis. He said that um chess was in quote 0:05:30.800 --> 0:05:34.360 the province of intellect alone. So the idea that there 0:05:34.400 --> 0:05:38.800 was this automaton playing chess blew people away. But again 0:05:38.880 --> 0:05:41.400 people figured out, like, okay, there's something going on here. 0:05:41.680 --> 0:05:44.840 We think that um von Kempelin is controlling this thing 0:05:44.920 --> 0:05:48.920 remotely somehow, maybe using magnets or whatever. Other people hit 0:05:48.960 --> 0:05:52.240 upon the idea that there was a small person inside 0:05:52.240 --> 0:05:55.080 the cabinet who would hide when the cab when the 0:05:55.120 --> 0:05:57.840 workings were shown, when the cabinet was open to show 0:05:57.839 --> 0:06:00.640 the workings, and then when the cabin was closed again 0:06:00.680 --> 0:06:03.360 and the mechanical turk started playing, the person had crawled 0:06:03.360 --> 0:06:05.960 back out and it was actually controlling it. This seems 0:06:06.000 --> 0:06:08.720 to be the case that there was a person controlling it, 0:06:09.160 --> 0:06:13.960 but the idea that it was it was a machine 0:06:14.120 --> 0:06:17.800 that could think and beat humans in chess, had like 0:06:17.920 --> 0:06:24.279 kind of unsettling implications. Yeah, this author Philip Thickness, great name, 0:06:25.360 --> 0:06:30.520 British author for sure, Philip Thickness. He uh, he said, 0:06:30.720 --> 0:06:32.920 and you know, people like you said all those more 0:06:32.960 --> 0:06:37.000 complicated explanations in this article you sent. Astuteley points out 0:06:37.440 --> 0:06:40.520 that he followed Occam's razor and basically said, he's got 0:06:40.560 --> 0:06:42.640 a little kid in there. He's got a little a 0:06:42.680 --> 0:06:45.839 little Bobby Fisher in there that's really good at chess, 0:06:46.480 --> 0:06:50.160 and that's what's going on. And other people speculated that other, 0:06:50.360 --> 0:06:52.800 you know, little people might be in there, Um, just 0:06:52.880 --> 0:06:55.159 adults who would fit in there. But then you know, 0:06:55.200 --> 0:06:57.480 there's the explanation that he would open it up and 0:06:57.880 --> 0:06:59.880 shine a candle around and say, you know nothing, just 0:07:00.000 --> 0:07:07.520 see here everyone. Um, so what should we reveal the 0:07:07.560 --> 0:07:11.320 real deal? Sure? I think I did already. Well, I 0:07:11.320 --> 0:07:14.120 don't think you spelled it out is we'll spell it out. 0:07:14.320 --> 0:07:16.920 There was a little person in there, Yeah, not just 0:07:16.960 --> 0:07:20.000 one little person, but they would travel around and recruit people. 0:07:20.040 --> 0:07:22.000 I guess people would get tired of being in there, 0:07:22.280 --> 0:07:24.440 or they forget about them and they'd starve and have 0:07:24.560 --> 0:07:27.120 to replace them. But it really was a trick. There 0:07:27.160 --> 0:07:29.640 was a little person in there. They did the same 0:07:29.680 --> 0:07:31.680 thing as like the magic acts. You know, when they 0:07:31.920 --> 0:07:34.880 saw a person in half. It's that the lady just 0:07:35.320 --> 0:07:37.960 gets into a tiny little ball in one section of 0:07:37.960 --> 0:07:40.360 that box. But my thing is this, like, this is 0:07:40.400 --> 0:07:43.560 not a satisfying explanation to me, Chuck, I think it's great. 0:07:43.800 --> 0:07:48.040 How did the person keep up with the board above? Well? 0:07:48.080 --> 0:07:50.640 I mean some I don't know if they ever proved 0:07:50.680 --> 0:07:53.560 exactly how it was going. That's what I'm saying, Okay, 0:07:53.560 --> 0:07:57.800 whether or not I think they that the zoltar or 0:07:57.800 --> 0:08:01.760 I'm sorry, the turk was just hollowed out and you 0:08:01.760 --> 0:08:05.640 you would just put your arms through the arm so 0:08:05.680 --> 0:08:07.760 you would call up into the church. Yeah, you would 0:08:07.880 --> 0:08:10.800 become the churk. When the churk would fuse, That's what 0:08:10.920 --> 0:08:13.400 some people thought. I think that's what Edgar Allen Poe 0:08:13.480 --> 0:08:18.240 thought too. He would Other people thought that there were 0:08:18.240 --> 0:08:22.160 the the the little person was underneath in the cabinet 0:08:22.480 --> 0:08:24.880 operating the turk with levers and stuff like that. Whether 0:08:24.880 --> 0:08:26.680 it could have been a mirror or something, you know, 0:08:26.880 --> 0:08:29.880 I guess a little telescopic mirror. That's what's getting me 0:08:29.960 --> 0:08:32.640 is how would they keep up with the game. You 0:08:32.640 --> 0:08:34.520 could keep track of the game, but how could you 0:08:34.559 --> 0:08:36.400 see where the other person moved? You would know where 0:08:36.400 --> 0:08:38.120 you moved, but you wouldn't be able to see where 0:08:38.120 --> 0:08:40.360 the other person. That's what I don't get. Just mirrors 0:08:40.440 --> 0:08:43.320 smoking mirrors maybe so. But the point is is it 0:08:43.360 --> 0:08:45.400 was a fake. It was a fraud, but it raised 0:08:45.480 --> 0:08:49.960 some really big questions about the idea of a machine 0:08:50.000 --> 0:08:53.320 beating a person at something like chess. Yeah, and it 0:08:53.760 --> 0:08:58.480 really peaked the mind of one Charles Babbage. He was 0:08:59.160 --> 0:09:00.840 he was a kid or or young at least at 0:09:00.880 --> 0:09:04.000 the time when he saw the turk in person, and 0:09:04.040 --> 0:09:06.920 a few years afterward he began work on something called 0:09:06.920 --> 0:09:11.120 the Difference Engine, which was a machine that he designed 0:09:11.120 --> 0:09:16.480 too to calculate mathematics automatically. So at some point to 0:09:16.559 --> 0:09:20.320 this is kind of maybe the beginnings of humans trying 0:09:20.360 --> 0:09:24.840 to create a I well, yeah, with Babbage's differential machine 0:09:25.240 --> 0:09:28.199 or difference machine, yeah, difference engine. But at the very least, 0:09:28.400 --> 0:09:31.440 what this is is the first that I know of 0:09:31.520 --> 0:09:35.840 example of man versus machine, even though it was really 0:09:35.920 --> 0:09:39.000 man versus man because it was a man in the machine. 0:09:39.120 --> 0:09:41.680 It was a fraud, yeah, but it was it. It 0:09:41.720 --> 0:09:45.679 sparked that idea, It definitely did. And that's something that UM, 0:09:45.720 --> 0:09:48.760 like chess in particular, has always been like this idea 0:09:49.800 --> 0:09:52.040 of like, if you can teach a machine to play chess, 0:09:52.080 --> 0:09:55.720 you really achieved a milestone. And there's been you know, 0:09:55.880 --> 0:10:00.559 plenty of programs, was most notably Deep Blue, which we'll 0:10:00.600 --> 0:10:04.679 talk about, but there's there's been this idea that like 0:10:04.800 --> 0:10:08.920 part of AI is chess, teaching it to play chess. 0:10:09.400 --> 0:10:12.960 But they, the people who develop a I never set 0:10:13.000 --> 0:10:16.360 out to make a chess playing AI, just to make 0:10:16.400 --> 0:10:18.959 a machine that can play chess. That's not the point. 0:10:19.520 --> 0:10:23.199 Chess has always been this way to demonstrate the progress 0:10:23.200 --> 0:10:27.160 of artificial intelligence because it's a complex game that you 0:10:27.280 --> 0:10:32.240 can't just program it, like it almost has to learn. Uh, well, 0:10:32.320 --> 0:10:34.560 it depends on how you come at it at first, right, 0:10:34.679 --> 0:10:38.920 So initially they did try to program it Okay, there's 0:10:39.040 --> 0:10:44.400 this for for from basically nineteen fifty two the about 0:10:44.480 --> 0:10:49.760 the mid like about say sixty years right. That is 0:10:49.760 --> 0:10:53.719 how they approached AI and chess is you figured out 0:10:53.800 --> 0:10:58.880 how to break chess down and explain it to a computer. Now, 0:10:59.040 --> 0:11:02.160 what if you could, ideally you would have this this 0:11:02.320 --> 0:11:07.120 computer or this AI, this artificial intelligence UM be able 0:11:07.160 --> 0:11:12.240 to think about the outcome of every possible every possible 0:11:12.240 --> 0:11:15.840 outcome of a move before making it. That's just not possible. 0:11:15.920 --> 0:11:18.559 It's still today we don't have computers that can do that, right, 0:11:19.000 --> 0:11:20.559 So what you have to do is figure out how 0:11:20.559 --> 0:11:24.080 to create shortcuts for the machine, give it best practices, 0:11:24.520 --> 0:11:27.199 that kind of thing. And that was actually laid out 0:11:27.360 --> 0:11:30.319 in by a guy named Claude Shannon, who is the 0:11:30.360 --> 0:11:34.120 father of information theory. And he wrote a paper with 0:11:34.200 --> 0:11:37.440 a pretty on the nose title called programming a Computer 0:11:37.520 --> 0:11:40.000 for Playing Chess. And you have to say it like 0:11:40.040 --> 0:11:42.199 that when you say the name, it's got a question 0:11:42.240 --> 0:11:44.320 market in the end, right. But he laid out two 0:11:44.360 --> 0:11:50.560 big things, um. One is creating a function of the 0:11:50.720 --> 0:11:55.040 different moves, and then another one is called a mini max. 0:11:56.000 --> 0:11:58.360 And if those were the two things that Shannon laid 0:11:58.400 --> 0:12:02.800 out and they established about fifty or sixty years of 0:12:02.840 --> 0:12:06.240 development in teaching an AI to play chess. Yeah, so 0:12:06.320 --> 0:12:10.240 this evaluation function is just sort of the base, the 0:12:10.360 --> 0:12:12.240 very basis of it all kind of where it starts, 0:12:12.280 --> 0:12:14.760 which is you ev kind of give a number to 0:12:14.800 --> 0:12:19.080 create a numerical evaluation based on the state of the 0:12:19.120 --> 0:12:24.040 board at that moment, and assign a real number evaluation 0:12:24.080 --> 0:12:28.040 to it. So, um, the highest number that you would 0:12:28.520 --> 0:12:32.920 shoot for is obviously getting checkmate, getting a king and 0:12:33.000 --> 0:12:36.520 checkmate right, right, So what you've just done now is 0:12:36.720 --> 0:12:40.079 by assigning a number to a state like the pieces 0:12:40.080 --> 0:12:43.240 on a board. Um, what you what you've done is 0:12:43.280 --> 0:12:46.800 to say, like shoot for this number. The higher the number, 0:12:46.880 --> 0:12:48.600 like you're going to give this a either rule. Now, 0:12:48.720 --> 0:12:52.079 the higher the number, the more desirable that this move 0:12:52.200 --> 0:12:55.760 that could lead to that higher number. Function. Evaluation function 0:12:56.200 --> 0:12:58.480 is what you want to do, right, like capture the 0:12:58.600 --> 0:13:00.600 night or capture the queen. Kept of the queen would 0:13:00.600 --> 0:13:04.280 have a higher evaluation number, right exactly. So that's the function. 0:13:04.320 --> 0:13:07.160 And there's another one called the mini max. Yeah, this 0:13:07.200 --> 0:13:10.000 is pretty great where you want to minimize the maximum. 0:13:10.040 --> 0:13:12.520 And this is another shortcut that they taught computers the 0:13:12.520 --> 0:13:15.719 maximum loss, that is right. So what they what they 0:13:15.960 --> 0:13:19.360 talk computers do is so you know, computer can look 0:13:19.400 --> 0:13:24.240 through an entire game every possible outcome. But what you 0:13:24.360 --> 0:13:27.240 there are computers that can look pretty far down the 0:13:27.320 --> 0:13:33.480 line at every possible outcome. And what you can say is, okay, um, 0:13:33.600 --> 0:13:36.880 you want to find the evaluation function that is the 0:13:36.880 --> 0:13:42.120 worst case scenario, the maximum loss, and then find the 0:13:42.200 --> 0:13:46.839 move that will minimize the possibility for that outcome. Yeah. 0:13:46.880 --> 0:13:49.600 By and this is your only limited by your programming power, 0:13:49.679 --> 0:13:51.880 but by looking not only at the state of the 0:13:51.920 --> 0:13:54.440 board right now, but if I make this move and 0:13:54.480 --> 0:13:57.719 I move the the pond to this spot, what are 0:13:57.720 --> 0:14:00.800 the next like three moves possibly that could happen as 0:14:00.800 --> 0:14:03.840 a result of this move. And you're only limited, like 0:14:03.880 --> 0:14:07.679 I said, by programming power. So obviously, the more juice 0:14:07.720 --> 0:14:10.959 you have, the more moves ahead that you can look exactly. 0:14:11.240 --> 0:14:14.199 And then they just shy away from ones with the 0:14:14.280 --> 0:14:17.560 higher function number or lower function number, depending on how 0:14:17.600 --> 0:14:20.960 you've programmed it. But they they they're making these decisions 0:14:21.000 --> 0:14:23.760 based on these rules, um. And then there's other things 0:14:23.840 --> 0:14:27.080 you can do, like little shortcuts to say, if if A, 0:14:27.240 --> 0:14:33.120 if A decision tree leads to a the other players 0:14:33.560 --> 0:14:36.840 King being in checkmate. Don't even think about that move 0:14:36.840 --> 0:14:39.840 any further, don't evaluate any longer, just abandon it because 0:14:39.880 --> 0:14:41.960 we would never want to make that move, right. So 0:14:41.960 --> 0:14:43.960 there's all the shortcuts you can do. And that's what 0:14:44.000 --> 0:14:47.360 they did to teach computers. That's what Deep Blue did 0:14:47.520 --> 0:14:51.720 when it beat Gary Kasparo. It was this huge, massive 0:14:51.760 --> 0:14:56.640 computer that knew a lot of chess, a lot about chess. 0:14:56.720 --> 0:14:58.880 It had a lot of rules, a lot of incredibly 0:14:59.200 --> 0:15:04.800 intricate programming that was extremely sharp, and it actually wanted 0:15:04.800 --> 0:15:08.920 became the first computer to beat an actual human chess 0:15:09.000 --> 0:15:13.480 grand master in like regulation match play. Yeah. I mean, 0:15:13.640 --> 0:15:16.760 and I don't think caspar Off gets enough credit for 0:15:16.840 --> 0:15:20.000 like willing being willing to do this, because it was 0:15:20.040 --> 0:15:22.960 a big deal for him to lose. It was in 0:15:23.000 --> 0:15:26.400 this community and the AI community. It was sent shock waves, 0:15:27.000 --> 0:15:31.080 and everyone that was alive remembers, even if you didn't 0:15:31.080 --> 0:15:34.320 know anything about either, one remembers Deep Blue being all 0:15:34.320 --> 0:15:36.480 over the news. It was a really big deal. And 0:15:36.560 --> 0:15:39.400 Casperrof put his name on the line and lost. Yeah, 0:15:39.400 --> 0:15:42.240 And I was wondering, Chuck, how how like you would 0:15:42.760 --> 0:15:46.600 get somebody to do that. I'm sure the Mountain of Catch. 0:15:47.040 --> 0:15:48.840 I guess that would probably be part of it. But 0:15:48.880 --> 0:15:52.600 I mean, I don't know. I bet, I bet that's 0:15:52.600 --> 0:15:54.160 out there. We just I just didn't look it up. 0:15:54.280 --> 0:15:58.240 So um, that's possible. It's also possible that they said, look, man, 0:15:58.320 --> 0:16:00.320 like this is chess, we're talking about what ever, But 0:16:00.440 --> 0:16:05.320 really what you're doing is helping advance artificial intelligence because 0:16:05.360 --> 0:16:07.800 we're not at we're not really trying ultimately to win 0:16:07.880 --> 0:16:10.840 chess games. We're trying to cure cancer. Yeah, we're gonna 0:16:10.880 --> 0:16:14.160 take your title because we're gonna beat you, or our 0:16:14.200 --> 0:16:16.880 machine is gonna beat you. But even still, you're gonna 0:16:16.880 --> 0:16:20.320 be helping with cancer. Think of the cancer Casparrow. That's 0:16:20.320 --> 0:16:23.200 probably what they said. Should we take a break? Yeah? Uh, what? 0:16:23.360 --> 0:16:27.280 Well should we tease our special guests? First? Is he okay? 0:16:27.440 --> 0:16:29.360 I can smell him. I don't think we even said 0:16:29.600 --> 0:16:32.840 we're gonna have a special guest. Later in the episode, 0:16:33.360 --> 0:16:36.000 Mr Jonathan Strickling of tech stuff nice. It's just been 0:16:36.000 --> 0:16:38.720 a long time since, like years since we had stricken. 0:16:38.920 --> 0:16:41.000 The last time we had stricken was like two thousand 0:16:41.160 --> 0:16:44.760 nine with the Necronomicon episode what is what where has 0:16:44.800 --> 0:16:47.120 he been besides sitting in between this every day? It's 0:16:47.120 --> 0:16:50.440 been a strickling drought is It's yes, the stricklands coming later. 0:16:50.480 --> 0:16:52.840 But we're gonna come back after this and talk a 0:16:52.840 --> 0:17:18.280 little bit more about man versus machine. Okay, dude, So 0:17:18.560 --> 0:17:22.399 what we just described was how AI was taught to 0:17:22.480 --> 0:17:26.040 play things like chess or to think like you take something, 0:17:26.280 --> 0:17:29.159 you figure out how to break it down into little 0:17:29.240 --> 0:17:33.000 rules and and and things that a computer can think of, right, 0:17:33.240 --> 0:17:35.640 and then follow these kind of rules to to make 0:17:35.680 --> 0:17:38.840 the best decision. That's how it used to be. The 0:17:39.960 --> 0:17:43.359 way that it's done now that everybody's doing now is 0:17:43.359 --> 0:17:48.080 where you are creating a machine that teaches itself. That 0:17:48.240 --> 0:17:51.560 was the breakthrough. You may have noticed back in about 0:17:51.560 --> 0:17:58.240 two all of a sudden, things like Siri and alexa 0:17:58.960 --> 0:18:02.440 um got way better at what they are doing. They 0:18:02.480 --> 0:18:07.000 got way less confused. You're a navigation app got a 0:18:07.040 --> 0:18:10.119 lot better. And the reason why is because these these 0:18:10.160 --> 0:18:12.440 this type, this new type of AI, this new type 0:18:12.440 --> 0:18:17.000 of machine learning that can teach itself and learn on 0:18:17.040 --> 0:18:20.800 its own, just hit the scene and they just started exploding. 0:18:20.840 --> 0:18:22.920 And one of the things that they were first trained 0:18:22.960 --> 0:18:25.879 on was games. Yeah, and it makes sense. Um. And 0:18:25.920 --> 0:18:29.320 if you thought chess was complicated and difficult, when it 0:18:29.359 --> 0:18:32.600 comes to these new AI s that they're teaching to 0:18:32.600 --> 0:18:36.280 teach themselves game strategy, they said, we might as well 0:18:36.320 --> 0:18:40.919 dive in to the Chinese strategic game Go because it 0:18:40.920 --> 0:18:44.760 has been called the most complex game ever devised by humans. 0:18:45.560 --> 0:18:50.440 And this was actually that was actually quote from Demi Hassabi, 0:18:51.040 --> 0:18:57.840 neuroscientists and the founder of deep Mind, which was deep Mind. 0:18:57.840 --> 0:19:00.320 They were purchased by Google or where they are always 0:19:00.320 --> 0:19:02.280 part of Google. I don't know if they were a 0:19:02.359 --> 0:19:04.960 spun off branch or where they were purchased, but it's 0:19:04.960 --> 0:19:08.159 one of Google's AI outfits. Well, they're they're one of 0:19:08.200 --> 0:19:10.560 the teams, yeah, that they are designing these new programs. 0:19:10.640 --> 0:19:14.440 And to give you an idea of how complex Go is, uh, 0:19:14.480 --> 0:19:17.639 it deals with a board with with different stones and 0:19:17.720 --> 0:19:21.919 there are uh ten how do you even say that 0:19:22.080 --> 0:19:25.080 tend to the hundred and seventy power So that means 0:19:25.119 --> 0:19:28.840 a hundred and seventies zeros uh and take that number 0:19:28.840 --> 0:19:32.679 and that's the number of possible configurations of a Go board. Right, So, 0:19:32.960 --> 0:19:36.800 like you say, chess is very complex and complicated and 0:19:36.800 --> 0:19:40.000 it's very difficult to master Go. And I've never played 0:19:40.040 --> 0:19:44.120 Go of you now, so it's supposedly it's easy to learn, right, 0:19:44.320 --> 0:19:47.760 but very complicated in its simplicity, right right, exactly, it's 0:19:47.880 --> 0:19:50.919 extremely difficult to master. And there was a guy in 0:19:50.960 --> 0:19:55.399 the late nineties, and I'm guessing that that he was 0:19:55.480 --> 0:20:00.119 saying this after Deep Blue beat casper Off. Um. There 0:20:00.200 --> 0:20:03.320 was an astrophysicists from Princeton. He said that it would 0:20:03.320 --> 0:20:05.680 probably be a hundred years before a computer beats the 0:20:05.760 --> 0:20:08.000 human at Go. To give you an idea of just 0:20:08.040 --> 0:20:12.520 how complex Go is, that deeply would just be cast 0:20:12.520 --> 0:20:14.600 prov And this guy is saying it will still be 0:20:14.640 --> 0:20:17.119 a hundred years before any anyone gets beat at Go 0:20:17.240 --> 0:20:19.880 by a computer. And he was someone who knew about 0:20:19.880 --> 0:20:22.400 this stuff, who was an astrophysicist. He wasn't just some 0:20:22.440 --> 0:20:25.800 schmoe at home and drunk in his recliner just making 0:20:26.280 --> 0:20:31.840 asinine predictions. Um. So, And again we've we've said this before, 0:20:31.920 --> 0:20:35.119 but I want to reiterate the people that, uh, I 0:20:35.119 --> 0:20:37.600 think Alpha Go is the name of this program. The 0:20:37.640 --> 0:20:40.760 people that created this a deep mind. They wanted to 0:20:40.800 --> 0:20:45.240 stress that this is a problem solving program. We're just 0:20:45.320 --> 0:20:48.520 teaching it this game at first, to to make it 0:20:48.600 --> 0:20:50.639 learn and to see if it can get good at 0:20:50.640 --> 0:20:53.520 what it does. But they said it is built with 0:20:54.119 --> 0:20:56.480 the idea that any task that has a lot of 0:20:56.560 --> 0:20:59.280 data that is unstructured, and you want to find patterns 0:20:59.280 --> 0:21:02.000 in the data and then decide what to do right, 0:21:02.119 --> 0:21:04.520 And that's kind of like what we're talking about. It's 0:21:04.640 --> 0:21:09.480 it crunches down all these possible options aka data to 0:21:09.640 --> 0:21:11.959 decide what move should I make right, and you can 0:21:12.000 --> 0:21:15.240 apply that. Ideally, they're gonna apply this to Alzheimer's and 0:21:15.280 --> 0:21:18.440 cancer and all sorts of things. Right. That's general general 0:21:18.640 --> 0:21:21.440 purpose thinking, right, yeah, and thinking on the fly to 0:21:22.160 --> 0:21:26.040 faced with novel stuff. So one of the reasons why 0:21:26.160 --> 0:21:28.919 it's good to use games like chess or go or whatever. 0:21:29.520 --> 0:21:33.960 Those are called perfect information games where both players or 0:21:33.960 --> 0:21:37.080 anybody watching has all the information that's available on it, 0:21:37.400 --> 0:21:41.399 there are definite rules their structure. It's a good proving ground. 0:21:41.680 --> 0:21:44.480 But as we'll see, AI makers are getting further and 0:21:44.520 --> 0:21:48.159 further away from those structure games as their AI becomes 0:21:48.160 --> 0:21:53.200 more and more sophisticated, because the structure and the limitations 0:21:53.200 --> 0:21:56.280 aren't necessarily needed anymore. Because these things are starting to 0:21:56.359 --> 0:21:58.680 be able to think on their own in a very 0:21:58.720 --> 0:22:04.680 generalized and even creative way. Yeah, it's really really interesting 0:22:05.560 --> 0:22:07.879 the way that they're like you said um earlier and 0:22:08.040 --> 0:22:10.800 before the break that we we don't have computers that 0:22:10.840 --> 0:22:13.719 can run all the possibilities. So what they teach in 0:22:13.760 --> 0:22:17.679 the case of AlphaGo, this program teaches itself by playing 0:22:17.680 --> 0:22:22.160 itself in these in these games and go specifically, and 0:22:22.240 --> 0:22:24.480 the more it plays itself, the more it learns, and 0:22:24.560 --> 0:22:29.320 the more ability it has during a game to choose 0:22:29.320 --> 0:22:33.200 a move by narrowing down possibilities. So instead of like, well, 0:22:33.200 --> 0:22:37.120 there are twenty million different variations here, by playing itself, 0:22:37.160 --> 0:22:40.080 it's able to say, well, in this scenario, there really 0:22:40.080 --> 0:22:44.120 only fifty different moves that I could or should make. 0:22:44.440 --> 0:22:46.520 Right or that's kind of a simplified way to say it. 0:22:46.560 --> 0:22:49.160 But right, No, but it's true. But that's that's exactly right. 0:22:49.200 --> 0:22:51.560 And what they're doing is basically the same thing that 0:22:51.600 --> 0:22:54.440 a human does. It's and it's going back to its 0:22:54.480 --> 0:22:58.240 memory banks, yeah, exact experience and saying well, I've I've 0:22:58.280 --> 0:23:00.359 been faced with something like this before, and this is 0:23:00.359 --> 0:23:02.879 what I used and it was successful. Forty out of 0:23:02.920 --> 0:23:05.280 fifty times. I'll do this one. This is a pretty 0:23:05.320 --> 0:23:09.080 reasonable move. That is what humans do. Yeah, not only 0:23:09.280 --> 0:23:11.720 I mean, boy, we screwed up the chess episode, but 0:23:11.760 --> 0:23:14.120 I get the idea that when you're a chess master, 0:23:14.320 --> 0:23:16.560 you don't just think what do the numbers say and 0:23:16.560 --> 0:23:19.480 what does the book say? But oh man, I did 0:23:19.520 --> 0:23:23.120 this move that one time and it didn't go as 0:23:23.160 --> 0:23:26.760 the book said. So that's now factored into my thinking, right, 0:23:27.240 --> 0:23:32.120 except imagine being able to learn from scratch and get 0:23:32.160 --> 0:23:35.520 to that point in eight days or eight hours. So 0:23:35.560 --> 0:23:39.080 that Go team, the Alpha Go the first the first 0:23:39.119 --> 0:23:41.760 iteration of Alpha Go. I think they started working on 0:23:41.800 --> 0:23:45.440 it in two thousand fourteen and in two thousands sixteen, 0:23:45.480 --> 0:23:47.800