WEBVTT - AI, Machine Learning and Human Creativity 0:00:03.120 --> 0:00:05.920 Welcome to Stuff to Blow Your Mind from how Stuff 0:00:05.920 --> 0:00:14.600 Works dot com. Hey, welcome to Stuff to Blow your Mind. 0:00:14.640 --> 0:00:18.439 My name is Robert Lamb and my name is Christian Sager. Hey. 0:00:18.640 --> 0:00:21.520 Something that people probably don't know about both of us 0:00:21.640 --> 0:00:24.880 is that outside of doing this show, we're both quote 0:00:24.920 --> 0:00:29.160 unquote creative people, right Like, even in the show. Well, 0:00:29.200 --> 0:00:30.920 in the show, we try to be creative. We can't 0:00:30.960 --> 0:00:33.760 strip the creativity off of cats like that. This is 0:00:33.840 --> 0:00:37.920 very true. But um, you are a fiction writer, which 0:00:37.960 --> 0:00:39.800 I always try to mention on the episodes that you're 0:00:39.800 --> 0:00:42.880 not on so that you're not like somehow embarrassed or 0:00:42.920 --> 0:00:45.960 anything like that. But I try to plug Robert's books. Uh, 0:00:46.000 --> 0:00:48.400 And then I do comics and and in the past 0:00:48.440 --> 0:00:51.800 I've done music as well, and I didn't know that. Yeah, 0:00:51.960 --> 0:00:53.680 and you were you like in a band. You were 0:00:53.680 --> 0:00:56.720 in a band, right, Yeah. And I had a period 0:00:56.760 --> 0:00:58.360 of time I'm gonna talk about in this episode where 0:00:58.400 --> 0:01:02.360 I dabbled with electronic music as well. Um, but yeah, so, 0:01:02.520 --> 0:01:04.639 but I was in a band for a couple of years. 0:01:05.160 --> 0:01:09.600 And Uh, there's this interesting thing going on right now 0:01:09.640 --> 0:01:11.759 that you saw when you went to the World Science 0:01:11.800 --> 0:01:15.720 Festival two weeks ago. That is sort of like trying 0:01:15.760 --> 0:01:19.959 to come to grips with what happens when you teach 0:01:20.400 --> 0:01:23.280 artificial intelligence or really, I guess the correct term is 0:01:23.319 --> 0:01:29.119 machine learning to be creative quote unquote, and like, how 0:01:29.160 --> 0:01:32.039 do people feel about that? And you saw this amazing 0:01:32.040 --> 0:01:34.800 panel where they had people who were dabbling in that, 0:01:34.840 --> 0:01:38.520 both in music and culinary arts and drawing, and then 0:01:38.560 --> 0:01:40.600 they had the psychologist on the panel to who is 0:01:40.640 --> 0:01:43.039 like the skeptic kind of Yeah, yeah, this was a 0:01:43.080 --> 0:01:45.800 really good, good panel. I saw it live at the 0:01:45.800 --> 0:01:48.720 World Science Festival this year, and if you're listening to this, 0:01:48.920 --> 0:01:50.880 you can watch it as well, because this one not 0:01:50.960 --> 0:01:52.520 all the talks are available on video, but this one 0:01:52.560 --> 0:01:54.440 is available on video, and I'll include a link to 0:01:54.600 --> 0:01:56.640 that talk on the landing page for this episode of 0:01:56.680 --> 0:01:58.520 Stuff to Blow Your Mind dot Com. But it was 0:01:58.560 --> 0:02:02.520 titled Computational Creativity. He was moderated by w n y 0:02:02.560 --> 0:02:07.080 C S John Schaefer, and the panelist included artist Song 0:02:07.200 --> 0:02:11.720 Win Chung, computer scientists and musician Jesse Ingle, neuroscientist Peter 0:02:11.800 --> 0:02:16.560 you Rock Say, and engineering theorist A lab of Varshney. Yeah, 0:02:16.600 --> 0:02:18.960 and so the four of them had this interesting discussion 0:02:19.040 --> 0:02:22.079 kind of a back and forth about what what they 0:02:22.080 --> 0:02:25.400 were experimenting with in terms of trying to teach machines 0:02:25.440 --> 0:02:28.640 to be creative, and then more along the lines of 0:02:28.760 --> 0:02:35.240 defining creativity, uh, systematizing creativity for human beings, so that 0:02:35.280 --> 0:02:38.200 you could somehow try to take that and apply it 0:02:38.240 --> 0:02:42.000 to machines and algorithms. Uh. And that the that's where 0:02:42.040 --> 0:02:45.280 Peter Rix, say was sort of like the the neuroscientists, 0:02:45.320 --> 0:02:48.200 not naysayer really, but just like he was doubtful right 0:02:48.320 --> 0:02:51.400 because he was he could not see a way that 0:02:51.520 --> 0:02:55.320 quote unquote creativity as he understood it could be replicated 0:02:55.960 --> 0:02:58.280 uh in machines the way that the other three were 0:02:58.320 --> 0:03:00.960 pretty excited about. Yeah. He I guess if you were 0:03:00.960 --> 0:03:03.560 thinking of this as a like a jam band or something. 0:03:04.320 --> 0:03:07.080 He was the guy with doing the jazzy solos, like 0:03:07.120 --> 0:03:11.600 all the other three had had really particular examples tied 0:03:11.639 --> 0:03:14.320 into their research, and then Say would come in and 0:03:14.360 --> 0:03:16.640 he'd just kind of he'd weigh a little bit. He 0:03:16.720 --> 0:03:19.440 had very interesting stuff to add to Yeah. Yeah, he 0:03:19.480 --> 0:03:21.680 was great, um, and it definitely brought some balance to it. 0:03:21.760 --> 0:03:24.239 So we thought, after you've seen it, you said, you 0:03:24.280 --> 0:03:26.560 gotta check this out. I watched the video, took notes, 0:03:26.800 --> 0:03:29.560 and we thought this is worth us having a conversation 0:03:29.600 --> 0:03:33.440 here on the show about it, especially because we're sort 0:03:33.480 --> 0:03:37.280 of at the intersectionality of science and creativity. Yeah, so, 0:03:38.000 --> 0:03:39.560 what we're not gonna do here today, We're not going 0:03:39.600 --> 0:03:42.720 to just replicate their discussion. We're gonna We're gonna spring 0:03:42.760 --> 0:03:46.120 off their discussion, uh, talk about our own experience, his 0:03:46.120 --> 0:03:48.000 own thoughts, and we're gonna drag in a few additional 0:03:48.080 --> 0:03:51.920 resources as well and in some additional examples, uh, to 0:03:51.920 --> 0:03:54.440 to bring this conversation to you and then hopefully here 0:03:54.440 --> 0:03:58.520 back from all of you creative people about your thoughts 0:03:58.840 --> 0:04:03.800 on machine creativity, machine learning, and the future of human creativity. 0:04:04.040 --> 0:04:06.680 So along those lines, Robert, like, when you hear the 0:04:06.760 --> 0:04:11.040 idea of a machine being creative, how does that make 0:04:11.080 --> 0:04:13.360 you feel as a person who like you know, I 0:04:13.680 --> 0:04:16.719 I imagine you're like me, Like it's work, you know, 0:04:16.839 --> 0:04:20.799 being creative, writing, drawing, playing music. It. You know, there's 0:04:20.880 --> 0:04:26.720 there's something that feels unescapable human about it. Yeah, yeah, 0:04:26.800 --> 0:04:29.160 you know. I I hear, I hear about it, and 0:04:29.160 --> 0:04:32.240 and I read articles or studies where they've taken a 0:04:32.240 --> 0:04:35.680 computer program and it's writing high coups, or it's where 0:04:35.720 --> 0:04:40.000 it's printing paintings, that sort of thing. I see that, 0:04:40.160 --> 0:04:42.719 and then I I think about these stories of say 0:04:42.839 --> 0:04:47.120 journalism jobs like local journalism jobs, jobs being outsourced to 0:04:47.160 --> 0:04:50.919 another country that I hear about, of course, the impact 0:04:50.920 --> 0:04:55.920 on manufacturing, of the impact of automation on manufacturing, and 0:04:55.960 --> 0:04:58.480 that makes me, you know, try and envision the future 0:04:58.880 --> 0:05:02.520 or where suddenly you're going to have programs and machines 0:05:02.600 --> 0:05:06.880 that can replicate what I do. Yeah, we've actually exactly 0:05:06.880 --> 0:05:10.360 to that point. We've actually like encountered that in this business, 0:05:10.760 --> 0:05:13.280 where like people will occasionally come to us and say, like, hey, 0:05:13.560 --> 0:05:16.160 you want to make your job easier podcasting. We've come 0:05:16.240 --> 0:05:19.840 up with this program that will do all the research 0:05:19.960 --> 0:05:22.360 for you, and all you have to do is just 0:05:22.440 --> 0:05:24.719 like you know, printed out and go into the studio 0:05:24.800 --> 0:05:28.719 with it. Uh. And you know that seems really interesting 0:05:28.760 --> 0:05:30.520 to me. That's I don't like I got that email. 0:05:30.640 --> 0:05:33.120 I got the one about where you're basically outsourcing it 0:05:33.120 --> 0:05:36.360 to other researchers. There's that, but then there's also, um, yeah, 0:05:36.400 --> 0:05:38.880 it'll just basically just pluck stuff for you and put 0:05:38.880 --> 0:05:41.119 it into a dock. Uh. This has has been pitched 0:05:41.160 --> 0:05:42.920 to me before at least. And then the other thing 0:05:42.920 --> 0:05:44.800 that I've heard about is like, you know, some of 0:05:44.839 --> 0:05:48.720 our listeners are familiar that we do videos occasionally. Uh, 0:05:48.839 --> 0:05:50.560 the old stuff to blow your mind videos would be 0:05:50.600 --> 0:05:52.919 a good example of this. So essentially you punch in 0:05:53.040 --> 0:05:55.760 parameters of what your video would be about. In this 0:05:56.800 --> 0:06:01.240 software spits out a like video for you that has 0:06:01.400 --> 0:06:05.440 text on the screen and pulls image clips or video 0:06:05.520 --> 0:06:08.960 clips out of a library to just make something rather 0:06:09.000 --> 0:06:12.320 than you're working together with a producer, writing script. All 0:06:12.400 --> 0:06:14.360 the things that we do when we make those videos 0:06:14.400 --> 0:06:16.560 just makes me feel like we need to be marketing. 0:06:16.560 --> 0:06:19.880 Our podcast is a craft podcast, you know, because it's 0:06:19.920 --> 0:06:23.279 like all the research is done by hand, granted, you know, 0:06:23.320 --> 0:06:25.120 on a computer screen. A lot of people don't know this. 0:06:25.200 --> 0:06:27.640 Book Mark Marin, all of his research is done by 0:06:27.760 --> 0:06:32.520 robot in his garage. Yeah yeah, okay, among the cats 0:06:32.560 --> 0:06:37.200 I'm imagining it's a robot cat actually yeah, um, but yeah, 0:06:37.240 --> 0:06:39.080 I kind of feel the same way, especially like when 0:06:39.120 --> 0:06:41.240 I hear these examples. But then at the same time, like, 0:06:41.760 --> 0:06:45.080 so I've done commercial art before. I was a graphic 0:06:45.080 --> 0:06:47.240 designer for a number of years before I got into 0:06:47.279 --> 0:06:50.680 the podcasting business. And I've done creative stuff where I've 0:06:50.720 --> 0:06:54.719 been an illustrator, I've worked as a musician, I write, uh, 0:06:54.760 --> 0:06:57.760 And so the thing that I've learned is that, like 0:06:58.640 --> 0:07:01.279 the one of the most iportant lessons I've learned about 0:07:01.279 --> 0:07:05.720 creativity is that you have to impose limits upon your canvas. 0:07:05.760 --> 0:07:08.840 No matter what medium you're performing in, right or else, 0:07:08.839 --> 0:07:11.720 You're just gonna like stare into this abyss forever with 0:07:11.960 --> 0:07:15.720 like trying to figure out where to place your creative ideas, um, 0:07:15.760 --> 0:07:18.880 because there's this infinite possibility of like what could be right, 0:07:18.920 --> 0:07:21.080 what could be wrong? Which which things should I use? 0:07:21.120 --> 0:07:25.640 Which combination of words, which kind of ink all those things? Um? 0:07:25.800 --> 0:07:29.080 For instance, like writers that I've encountered who have the 0:07:29.080 --> 0:07:33.560 problem where they'll rewrite the first paragraph that they write 0:07:33.600 --> 0:07:36.360 over and over and over again and obsessed on making 0:07:36.400 --> 0:07:39.160 that paragraph perfect before they'll move on, and then they 0:07:39.200 --> 0:07:42.800 never finished. Right. Yeah, I've I've I've done that before. 0:07:43.000 --> 0:07:46.119 Like when I first started righting fiction, I was doing 0:07:46.160 --> 0:07:48.080 that model, and of course the thing I was writing 0:07:48.480 --> 0:07:52.640 is still to this day incomplete from from that totally. Yeah, 0:07:52.680 --> 0:07:54.400 I've been there too, and It's like that's one of 0:07:54.440 --> 0:07:57.000 those things that they don't ever really teach you. At 0:07:57.080 --> 0:08:00.880 least they didn't teach me very well in composition, yes, 0:08:01.040 --> 0:08:03.800 like just move on, um. But when I think about 0:08:03.840 --> 0:08:08.360 creative machines, then I have to wonder, first of all, 0:08:08.400 --> 0:08:11.200 are they aware of what their limitations are or are 0:08:11.240 --> 0:08:15.280 they just executing an algorithm? And how do they do 0:08:15.360 --> 0:08:18.040 they know how to make those limitations work for them 0:08:18.080 --> 0:08:22.160 creatively the same way that human beings do well. It's 0:08:22.200 --> 0:08:25.320 interesting because you know, when you when you start picking 0:08:25.360 --> 0:08:28.600 apart your own creative process, you can you can begin 0:08:28.720 --> 0:08:31.160 to see how a machine could do some of it, 0:08:31.200 --> 0:08:34.240 you know, because you're talking about it imposing limits on yourself. 0:08:34.440 --> 0:08:36.560 Like I'm instantly thinking like, all right, I'm writing, I'm 0:08:36.559 --> 0:08:40.360 gonna write a I'm gonna write this science fiction short story. Okay, 0:08:40.679 --> 0:08:43.560 what kind of ideas can I explore within the confines 0:08:43.600 --> 0:08:46.440 of this story, this setting, maybe this type of character 0:08:46.480 --> 0:08:48.640 that I want to write? And then you start you 0:08:48.679 --> 0:08:51.400 start throwing in another parameter. It's like what kind of 0:08:51.440 --> 0:08:53.960 stories are selling? What? What? What kind of story can 0:08:54.000 --> 0:08:55.920 I write that doesn't have a vampire in it? You know, 0:08:55.960 --> 0:08:59.080 stuff like that, And so there's a there's a certain 0:08:59.120 --> 0:09:04.240 amount of creativity. Is not necessarily this just magical explosion 0:09:05.040 --> 0:09:07.720 that takes place. There's a lot of computation that goes 0:09:07.760 --> 0:09:11.120 into it, and therefore there's an entire side to it 0:09:11.160 --> 0:09:15.360 that could easily be mastered by an inhuman entity, right, 0:09:15.520 --> 0:09:18.800 especially if like we as humans are able to go 0:09:18.960 --> 0:09:22.880 to that machine and input the parameters of those limits 0:09:22.920 --> 0:09:26.000 to it, right and say like, okay, here's your limitations 0:09:26.320 --> 0:09:29.079 now right this story or here's your limitations now create 0:09:29.120 --> 0:09:32.120 this music. And that seems to be what they're trying 0:09:32.160 --> 0:09:35.200 to do. Yeah, And I know when when I think about, 0:09:35.760 --> 0:09:39.320 you know, these various anxieties about you know, future writers 0:09:39.320 --> 0:09:42.640 and podcasters that are entirely machine. You know, we all 0:09:42.640 --> 0:09:45.320 tell ourselves, well, they can replicate a lot of it, 0:09:45.360 --> 0:09:47.560 but there there's a certain spark, there's something I have 0:09:47.920 --> 0:09:50.920 that cannot be replicated. And we we also tell that's 0:09:50.920 --> 0:09:53.400 the same thing we tell ourselves when we're talking about 0:09:53.480 --> 0:09:55.720 human competition. You can say, all right, well this this 0:09:55.800 --> 0:09:59.800 individual may be better at at this skill set than me. 0:10:00.280 --> 0:10:02.079 They maybe have a better voice or what have you, 0:10:02.440 --> 0:10:05.120 but but there's something special about me, and so do 0:10:05.160 --> 0:10:07.760 I extend or is that just sort of a self 0:10:07.760 --> 0:10:12.600 inflating uh fiction that we're telling ourselves all the time anyway. 0:10:12.679 --> 0:10:15.000 Uh And and so maybe the machines will just swoop 0:10:15.040 --> 0:10:18.200 in and take all of our creative jobs. Maybe, Although 0:10:18.400 --> 0:10:20.520 I have to say, like sitting down and listening to 0:10:20.559 --> 0:10:23.040 this conversation and then like filling out notes and doing 0:10:23.040 --> 0:10:25.880 a little extra research for this episode, it made me 0:10:25.960 --> 0:10:30.000 think about my creative process in a systematized way that 0:10:30.080 --> 0:10:33.440 I haven't before, And it made me really realize that 0:10:34.240 --> 0:10:36.160 for me at least, I think like one of the 0:10:37.080 --> 0:10:41.400 big like jumping points for human beings being creative is 0:10:41.760 --> 0:10:46.240 when they start creating new tools to be creative in 0:10:46.280 --> 0:10:48.880 different ways. And we'll talk about that later because this 0:10:48.920 --> 0:10:52.319 is something they discussed in the panel as well. Uh. 0:10:52.360 --> 0:10:57.480 And I can't see a future yet where AI is 0:10:57.559 --> 0:11:02.680 doing those things where AI says, I like a guitar, 0:11:02.880 --> 0:11:05.160 and I can write these songs on guitar, But what 0:11:05.240 --> 0:11:07.000 if I go over to a tree and I carve 0:11:07.040 --> 0:11:10.160 an entirely different kind of instrument so that I can 0:11:10.160 --> 0:11:12.439 get this effect. That's true. That's that's a good point, 0:11:12.720 --> 0:11:16.360 sort of like meta creativity ideas. All right, so we'll 0:11:16.400 --> 0:11:18.520 definitely get into the tools here in a bit, but 0:11:19.080 --> 0:11:20.520 before we roll on, I just want to want to 0:11:21.000 --> 0:11:24.560 really draw attention to two particular ideas that came up 0:11:24.559 --> 0:11:27.760 in the discussion um that that I that I thought 0:11:27.800 --> 0:11:30.720 were particularly striking, and that will come up again and 0:11:30.720 --> 0:11:33.560 again as we uh as we roll through the episode. 0:11:33.600 --> 0:11:36.079 First of all, they kept to bring up this analogy 0:11:36.160 --> 0:11:39.200 of the career of creative technology compared to the airplane. 0:11:39.640 --> 0:11:42.400 So obviously we make airplanes that fly faster and higher 0:11:42.480 --> 0:11:45.360 than any bird, but that doesn't mean we fully understand 0:11:45.360 --> 0:11:48.439 how birds fly, nor does it mean we can replicate 0:11:48.520 --> 0:11:52.360 the evolved perfection of their powered flight. So technology essentially 0:11:52.400 --> 0:11:56.040 gives us a version of the same thing that outperforms 0:11:56.080 --> 0:12:00.480 the organic in many respects, but also underperforms and other areas. 0:12:00.559 --> 0:12:03.719 And this was al rix Says primary argument sort of 0:12:03.840 --> 0:12:08.199 against the idea that AI could be quote unquote creative, 0:12:08.360 --> 0:12:13.280 right because birds airplanes are not birds, right, And so 0:12:13.320 --> 0:12:18.080 he was saying basically that then subsequently the software couldn't 0:12:18.080 --> 0:12:22.200 necessarily be human in its creativity, But it depends on 0:12:22.240 --> 0:12:25.080 how you define creativity too. Yeah, it's a great metaphor, 0:12:25.200 --> 0:12:28.160 and I think all the panelists ended up, picking up 0:12:28.200 --> 0:12:29.559 and playing with it a little bit, and well we'll 0:12:29.600 --> 0:12:32.360 play with it here as well, now chewing. For her part, 0:12:32.440 --> 0:12:36.800 she highlighted her own projects with creative machine learning, in 0:12:36.840 --> 0:12:41.360 which she and there's some wonderful like overhead video examples 0:12:41.360 --> 0:12:45.040 of this, but she draws, uh with or paints with 0:12:45.080 --> 0:12:47.320 a piano. Believe it was a piano. It looked to 0:12:47.360 --> 0:12:51.160 me like she was using some kind of stylist tablet thing. 0:12:51.440 --> 0:12:54.560 But yeah, like basically she from what I could tell 0:12:54.760 --> 0:12:58.120 as somebody who does illustration, she was drawing on the 0:12:58.240 --> 0:13:02.160 right half of a of a canvas, and what she 0:13:02.240 --> 0:13:05.320 would draw, the computer would then draw on the left 0:13:05.320 --> 0:13:07.839 half of the canvas. But it wouldn't just imitate her. 0:13:07.880 --> 0:13:09.520 It wasn't like, I mean, you can do this in 0:13:09.520 --> 0:13:12.040 like Adobe Illustrator where it just like draws the exact 0:13:12.120 --> 0:13:14.400 same thing you draw, and you've got like a perfectly 0:13:14.400 --> 0:13:17.520 symmetrical drawing. So it's not a mirror image though. This 0:13:17.600 --> 0:13:21.720 was this was like creating its own thing, learning from 0:13:21.840 --> 0:13:25.120 her curves and lines. Yeah, I was playing off her movement, 0:13:25.200 --> 0:13:28.360 sort of improvising with her to create a work of 0:13:28.400 --> 0:13:31.040 collaborative art. And so she argues that the notion of 0:13:31.080 --> 0:13:35.040 a robotic agent that creates with you the dreams with 0:13:35.120 --> 0:13:39.360 you in harmony is quote an underserved narrative in our culture, 0:13:39.440 --> 0:13:42.240 and I and I definitely agree with that. She was 0:13:42.960 --> 0:13:45.240 to me the most compelling person on the panel. I 0:13:45.280 --> 0:13:48.240 thought they were all really interesting. There was something about 0:13:48.280 --> 0:13:49.839 her that I was like, this person is going to 0:13:49.960 --> 0:13:53.160 go on to do like amazing work, and she's gonna 0:13:53.240 --> 0:13:55.680 come back. Because she was relatively young, I can see 0:13:55.679 --> 0:13:57.679 her in like ten twenty years coming back to us 0:13:57.679 --> 0:14:00.880 and just like dropping like some huge revelation on us 0:14:00.920 --> 0:14:08.360 about creativity and computer learning after the Singularity, after shelded 0:14:08.400 --> 0:14:11.600 into the mind. They chose me I'm going to be 0:14:11.679 --> 0:14:13.520 the go between. But she was like she was dropping 0:14:13.559 --> 0:14:17.840 some really interesting tidbits into the conversation, Like she was 0:14:17.920 --> 0:14:23.440 thinking about the way that she draws based on competitive gaming. 0:14:24.000 --> 0:14:26.360 Like she kept bringing up like, and by this we 0:14:26.400 --> 0:14:29.000 mean like video games where people are playing I don't know, 0:14:29.080 --> 0:14:33.680 League of Legends or um Overwatch, and so just the 0:14:33.720 --> 0:14:36.360 idea that she was thinking about how like the AI 0:14:36.480 --> 0:14:40.600 in those games was compensating for the human movements or 0:14:40.720 --> 0:14:45.000 or play styles as opposed to how she drew and 0:14:45.040 --> 0:14:48.040 how the computer drew along with her. That's an very 0:14:48.120 --> 0:14:52.040 interesting connection that I would never have thought of before. Yeah, 0:14:52.080 --> 0:14:54.480 that's that's that that is that's a That's a really 0:14:54.520 --> 0:14:56.640 a really great example because you don't think about like 0:14:56.720 --> 0:15:01.520 this competitive model being applicable to a creative collaboration. And 0:15:01.640 --> 0:15:03.840 you and I play a lot of video games, you know, 0:15:03.920 --> 0:15:06.080 but I don't spend a lot of time thinking about 0:15:06.080 --> 0:15:10.040 how the AI in the game is compensating for my movements. 0:15:10.720 --> 0:15:12.560 I have the only time I've thought about this, and 0:15:12.600 --> 0:15:15.520 this is going to be kind of geeky and and silly, 0:15:15.640 --> 0:15:19.560 but in playing uh, pro wrestling video games, I've thought 0:15:19.600 --> 0:15:23.560 about this because because there's like certain wrestling styles that 0:15:23.640 --> 0:15:27.040 the characters have, well, it basically comes down to this weirdness. 0:15:27.640 --> 0:15:34.080 Professional wrestling is a simulated combat sport. So yes, sorry spoilers, 0:15:34.120 --> 0:15:39.479 but as you see it on TV, it's two humans pretending, 0:15:40.120 --> 0:15:42.800 you know, acting out a combat scenario. We actually have 0:15:42.840 --> 0:15:44.920 a video about this on KFE that you did a 0:15:44.920 --> 0:15:46.920 couple of years ago. Yeah, and so when people make 0:15:46.920 --> 0:15:49.680 a video game of that, they make a fighting game 0:15:50.000 --> 0:15:53.560 in which too like a huge too humans or a 0:15:53.640 --> 0:15:55.680 human in an AI or two AI s if you're 0:15:55.680 --> 0:15:58.400 doing a demo mode, are fighting each other in a 0:15:58.440 --> 0:16:03.840 competitive in calendar using the moves of professional wrestling. And 0:16:03.880 --> 0:16:06.560 then on top of that, some players are going to 0:16:06.760 --> 0:16:09.120 play that video game just like a fighting game. I'm 0:16:09.120 --> 0:16:12.520 gonna beat my opponent, but others will play it attempting 0:16:12.560 --> 0:16:15.880 to tell a story, attempting to get the most dramatic match. 0:16:16.400 --> 0:16:18.720 And in those cases, I don't know that any programmers 0:16:18.720 --> 0:16:24.560 have really gained that really strange area of video game playing, 0:16:25.160 --> 0:16:27.720 but like, like that's that's what I think of when 0:16:27.760 --> 0:16:29.840 she brings up this example, Like, here's an example of 0:16:29.920 --> 0:16:34.520 somebody trying to not only defeat an AI opponent, but 0:16:34.600 --> 0:16:38.120 to try and tell a story with an AI opponent. Okay, 0:16:38.120 --> 0:16:39.640 so this is an episode where we're gonna have a 0:16:39.640 --> 0:16:42.600 little bit of more digressions than we normally do because 0:16:42.720 --> 0:16:45.080 of our creative backgrounds and sort of things we can 0:16:45.120 --> 0:16:48.040 bring to this. I have a story that I think 0:16:48.080 --> 0:16:50.160 can tie into this that we can then maybe bring 0:16:50.200 --> 0:16:52.320 back to this AI. Okay, So a friend of mine 0:16:52.400 --> 0:16:55.440 is currently working on a comic book for a w 0:16:55.440 --> 0:16:57.680 W E comic. He's just doing like a short story 0:16:58.120 --> 0:17:05.560 and it's about some missfight between the Undertaker and Mankind 0:17:05.720 --> 0:17:10.199 is the guy that would okay, and uh so I 0:17:10.240 --> 0:17:12.399 never saw this. I'm not as bigger a wrestling fan 0:17:12.480 --> 0:17:15.000 as Robert, but uh when he told me about it, 0:17:15.040 --> 0:17:17.359 was like, oh, cool, like you could you could do 0:17:17.440 --> 0:17:19.359 some fun stuff like that, like have them being a 0:17:19.400 --> 0:17:21.800 slaughterhouse or have them be in a graveyard and they're 0:17:21.840 --> 0:17:24.320 hitting each other with gravestones or something, right, and he 0:17:24.400 --> 0:17:27.040 was like, no, we have to do an exact replica 0:17:27.280 --> 0:17:30.760 in comic form of this infamous fight they hadn't like 0:17:30.760 --> 0:17:34.480 I don't know, two thousand or something. And I was like, huh, 0:17:34.520 --> 0:17:38.040 that's that's interesting. But to me that like here we 0:17:38.080 --> 0:17:42.200 go with limitations, right, Like the limitation there is so 0:17:42.320 --> 0:17:45.840 limiting that I don't see what would be creative about 0:17:45.840 --> 0:17:48.760 it or interesting other than obviously my friend and his 0:17:48.840 --> 0:17:51.520 style and how he draws the thing. But he's not 0:17:51.920 --> 0:17:58.240 being able to express anything. Mhmm. Yeah, yeah, yeah, I 0:17:58.280 --> 0:17:59.520 see what you mean. Because the one thing if you 0:17:59.520 --> 0:18:02.560 could say, I tell this, tell the story, but they 0:18:02.560 --> 0:18:06.520 actually have magical powers, like they're Undertakers, really an undead 0:18:06.520 --> 0:18:09.000 creature exactly, That's how I was thinking it would be. 0:18:09.040 --> 0:18:11.320 I was like, oh, that this is the benefit of 0:18:11.359 --> 0:18:13.320 doing it in the comics medium is that you can 0:18:13.359 --> 0:18:16.679 actually give them weird powers and you can like really 0:18:16.760 --> 0:18:22.440 play with the the goofy narrative of like their their roles. Right. Um. 0:18:22.600 --> 0:18:24.679 But then this this also leads to the creative challenge, 0:18:24.760 --> 0:18:28.480 Like sometimes just because the constraints are there doesn't mean 0:18:28.520 --> 0:18:31.280 there's there's not an opportunity exactly. And that's what I'm 0:18:31.320 --> 0:18:34.640 looking forward to it because when I hear that, I go, man, 0:18:34.680 --> 0:18:38.120 there's there's too many limitations there. But you know, who knows, 0:18:38.240 --> 0:18:41.120 maybe there's a way within all those limitations. I trust 0:18:41.200 --> 0:18:42.520 that my friend is gonna be able to do this, 0:18:43.000 --> 0:18:45.240 that that they'll be like a really cool way to 0:18:45.280 --> 0:18:48.520 depict that fight. All right, but it's gonna take creativity. 0:18:48.600 --> 0:18:50.679 And so we're gonna we're gonna roll it back a 0:18:50.680 --> 0:18:54.879 little bit and talk about what creativity actually is, like 0:18:55.000 --> 0:18:58.720 boiling it down, reconstructing it so that we can understand 0:18:58.800 --> 0:19:01.119 what we're doing when we're create and in order to 0:19:01.600 --> 0:19:05.240 uh make a machine engage in the same process. Right, 0:19:05.600 --> 0:19:09.760 So they come up with some of their uh kind 0:19:09.800 --> 0:19:14.119 of limited limitations on what creativity can or cannot be. 0:19:14.560 --> 0:19:16.679 And this is where I think, like Ultimately, like the 0:19:16.720 --> 0:19:19.680 split between the four of them, or maybe the three 0:19:19.680 --> 0:19:24.040 of them versus Alri say came down was that Ria 0:19:24.160 --> 0:19:28.120 has this sort of broader definition of creativity as being 0:19:28.119 --> 0:19:30.840 this sort of magical thing that happens in the human brain. Right, 0:19:31.000 --> 0:19:33.520 like he talks about at different times in the discussion, 0:19:33.520 --> 0:19:37.040 talks about the importance of pain, of intent, of consciousness, 0:19:37.720 --> 0:19:40.919 various attributes that are currently beyond the scope of of 0:19:41.000 --> 0:19:43.840 AI and would be problematic if we had them, you know, 0:19:43.880 --> 0:19:47.080 if you could, would it be It's a whole separate issue, 0:19:47.119 --> 0:19:49.760 but it would it be ethical to make a you know, 0:19:49.800 --> 0:19:53.920 a depressed artist robot because why what kind of cruel 0:19:53.920 --> 0:19:55.800 god are you to make a robot so depressed? You know, 0:19:55.800 --> 0:19:57.520 it's funny you mentioned that later on in the notes. 0:19:57.560 --> 0:20:00.680 I have a question on whether robots can be crazy 0:20:00.760 --> 0:20:04.640 enough to be creative, because, as we know, many creative 0:20:04.800 --> 0:20:08.320 people suffer mental health issues. Yeah, well that's how we 0:20:08.400 --> 0:20:13.080 got David and in the Covenant. Well they already answered 0:20:13.119 --> 0:20:16.240 it for us, Thanks for at le Scott. Alright, So 0:20:16.600 --> 0:20:18.560 one of the one of these sort of loose models 0:20:18.560 --> 0:20:21.879 though for creativity that that is thrown out here is 0:20:22.119 --> 0:20:25.560 it comes from engineering theorist lab Varshny, and he breaks 0:20:25.560 --> 0:20:28.360 it all down to a comp to a combination of 0:20:28.840 --> 0:20:32.880 novelty and high quality that comes together in a way 0:20:32.920 --> 0:20:36.040 that changes your belief. So like a Venn diagram, I'm 0:20:36.080 --> 0:20:39.080 imagining here of is it novel, is it knew? Is 0:20:39.080 --> 0:20:41.920 it something different? And is it of high enough quality? 0:20:42.119 --> 0:20:44.080 And when it comes and does it come together in 0:20:44.119 --> 0:20:46.760 a way that actually have some sort of meaning to me? 0:20:46.960 --> 0:20:50.480 I think this is a really interesting way to model 0:20:51.560 --> 0:20:54.359 aesthetics and that I had not thought of before. And 0:20:54.400 --> 0:20:56.560 when you try to like sort of lay this on 0:20:56.640 --> 0:21:00.240 top of like creative artifacts and see like how that works, 0:21:00.560 --> 0:21:02.560 it's interesting, right, like the first thing I thought of. 0:21:02.600 --> 0:21:04.000 I was talking to you about this before we went 0:21:04.000 --> 0:21:06.160 in to the studio, and no spoilers, but the new 0:21:06.200 --> 0:21:08.359 Twin Peaks is on TV and I've been watching it 0:21:08.400 --> 0:21:11.040 every week, and as you would expect from David Lynch, 0:21:11.400 --> 0:21:15.960 it's wacky, right, But in terms of lab varsenyse ven 0:21:15.960 --> 0:21:19.240 diagram here of novelty and high quality, it's just the 0:21:19.359 --> 0:21:22.159 right amount of novelty and just the right amount of 0:21:22.280 --> 0:21:26.160 quality that it works for me. Now maybe for other people, 0:21:26.160 --> 0:21:31.080 because obviously creativity and artist objective. It doesn't, but but yeah, 0:21:31.160 --> 0:21:33.280 I was seeing it like, oh, that is the perfect 0:21:33.359 --> 0:21:35.360 David Lynch is like the perfect metaphor for this, because, 0:21:35.440 --> 0:21:38.080 let's be honest, like some of David Lynch's stuff, the 0:21:38.160 --> 0:21:41.080 novelty factor can get a little too big in that 0:21:41.160 --> 0:21:44.480 ven diagramic situation and you're like and and Varsny actually 0:21:44.560 --> 0:21:47.520 describes it as noise. He says, like, when either one 0:21:47.520 --> 0:21:50.480