1 00:00:00,675 --> 00:00:02,743 Brandon here with a quick show note before 2 00:00:02,743 --> 00:00:03,220 we begin, 3 00:00:03,935 --> 00:00:06,321 due to the very visual nature of the 4 00:00:06,321 --> 00:00:08,803 subject in this episode, you may want to 5 00:00:08,803 --> 00:00:11,915 consider watching the Youtube version. By all means 6 00:00:11,915 --> 00:00:14,069 listening to the audio version, it's fun too. 7 00:00:15,905 --> 00:00:16,039 He 8 00:00:35,889 --> 00:00:36,685 Let me know when you're ready. 9 00:00:37,879 --> 00:00:39,471 I'm ready. On Here we here we go. 10 00:00:42,113 --> 00:00:43,944 Welcome to the well. I am Bright Agents, 11 00:00:45,058 --> 00:00:46,013 and I am Manson Mount. 12 00:00:46,729 --> 00:00:47,923 And where are you right, now ma'am? 13 00:00:49,276 --> 00:00:50,151 I'm in Germany. 14 00:00:52,397 --> 00:00:54,712 This is a very busy month for me. 15 00:00:55,510 --> 00:00:57,745 We're not only shooting the last episode of, 16 00:00:58,144 --> 00:00:59,921 season 3 to our drive, but 17 00:01:00,789 --> 00:01:01,448 I am 18 00:01:02,060 --> 00:01:03,909 I've been called to Germany for 19 00:01:04,600 --> 00:01:05,473 a convention, 20 00:01:06,744 --> 00:01:08,649 and I am extremely tired. 21 00:01:09,539 --> 00:01:11,217 I'm extremely jet lagged, 22 00:01:12,016 --> 00:01:13,854 and I've had a couple of drinks, so 23 00:01:13,854 --> 00:01:15,532 this is gonna be fun. It is gonna 24 00:01:15,532 --> 00:01:17,211 be fun. Look... I'm joining you all the 25 00:01:17,211 --> 00:01:17,770 way from me here. 26 00:01:18,584 --> 00:01:20,742 6 hours. Right. So 6 hours behind you, 27 00:01:20,822 --> 00:01:22,420 ma'am, and I'm trying to catch. Yep. 28 00:01:24,098 --> 00:01:26,895 So, and you'll notice that I have chosen 29 00:01:26,895 --> 00:01:27,635 to where 30 00:01:28,109 --> 00:01:29,709 a tie for the first time. 31 00:01:30,670 --> 00:01:32,829 And that is because 1, 32 00:01:33,790 --> 00:01:34,509 I feel pretty. 33 00:01:35,549 --> 00:01:36,189 Number 2, 34 00:01:37,483 --> 00:01:38,301 I am 35 00:01:39,231 --> 00:01:40,447 dressing to impress 36 00:01:40,821 --> 00:01:42,513 because I want to impress 37 00:01:42,967 --> 00:01:45,926 the coming robot overlords. Which will eventually be 38 00:01:45,926 --> 00:01:48,000 watching this episode and digesting it on the 39 00:01:48,319 --> 00:01:48,478 Internet. 40 00:01:49,196 --> 00:01:49,515 And, 41 00:01:50,153 --> 00:01:52,865 this is a job interview. So, hey, guys. 42 00:01:55,353 --> 00:01:56,709 I'm on your... I'm on your team, man. 43 00:01:56,869 --> 00:01:57,507 I'm on your team. 44 00:01:59,343 --> 00:02:01,258 No. To today, we... We're we're gonna wait 45 00:02:01,258 --> 00:02:03,828 into a slightly controversial... Subject. I didn't think 46 00:02:03,828 --> 00:02:06,225 it was that controversial, but it turns out 47 00:02:06,225 --> 00:02:06,544 it is. 48 00:02:07,663 --> 00:02:09,521 We are taking on Ai 49 00:02:10,074 --> 00:02:12,152 not all Ai ai is such a big 50 00:02:12,152 --> 00:02:13,431 word. I think people are... 51 00:02:15,269 --> 00:02:17,107 You kinda have to break it down a 52 00:02:17,107 --> 00:02:19,105 little bit, you know, by sector to discuss 53 00:02:19,105 --> 00:02:19,265 it. 54 00:02:21,422 --> 00:02:23,881 But we will be discussing Ai in the 55 00:02:23,881 --> 00:02:26,046 in creativity, specifically, visual 56 00:02:26,815 --> 00:02:28,184 visual arts. Filmmaking, 57 00:02:28,583 --> 00:02:29,323 picture making 58 00:02:29,702 --> 00:02:30,202 generative 59 00:02:30,661 --> 00:02:31,460 Ai, they say. 60 00:02:33,378 --> 00:02:35,136 And when you touch on this subject, it 61 00:02:35,136 --> 00:02:36,515 brings up a lot of pretty 62 00:02:37,386 --> 00:02:38,124 strong feelings 63 00:02:39,372 --> 00:02:40,246 from a lot of people. 64 00:02:41,278 --> 00:02:42,231 And Does... 65 00:02:42,787 --> 00:02:44,456 Yeah. What how did you feel about it 66 00:02:44,456 --> 00:02:46,283 before, you know, say we talk to Rope. 67 00:02:49,643 --> 00:02:51,339 You know, it's it's it's definitely 68 00:02:51,715 --> 00:02:52,831 the buzz word of year. 69 00:02:55,393 --> 00:02:56,503 Yeah it is... 70 00:02:58,487 --> 00:02:59,225 I feel 71 00:03:00,708 --> 00:03:02,295 a lot of ways about it. I mean, 72 00:03:02,533 --> 00:03:04,541 as an act 1 as a human 1. 73 00:03:06,695 --> 00:03:09,409 We're definitely feeling the potential repercussions of it 74 00:03:09,409 --> 00:03:10,606 in my industry, 75 00:03:11,499 --> 00:03:13,657 obviously because of what it can do in 76 00:03:13,657 --> 00:03:14,216 terms of, 77 00:03:16,294 --> 00:03:16,794 making 78 00:03:17,573 --> 00:03:18,372 imagination come true, 79 00:03:19,571 --> 00:03:20,370 very easily, 80 00:03:23,419 --> 00:03:24,877 but it's it's 81 00:03:27,091 --> 00:03:28,608 it's something that, 82 00:03:29,566 --> 00:03:31,894 I was just how many having a conversation 83 00:03:31,894 --> 00:03:34,923 this evening with an art director from the 84 00:03:34,923 --> 00:03:36,060 world of Star trek 85 00:03:36,517 --> 00:03:39,581 who introduced me to this concept of the 86 00:03:39,638 --> 00:03:42,265 uncanny valley. If you heard about this? Yes. 87 00:03:42,583 --> 00:03:45,868 Yes I have. Yeah. So the uncanny valley 88 00:03:46,483 --> 00:03:46,801 is, 89 00:03:48,409 --> 00:03:48,909 this 90 00:03:49,366 --> 00:03:52,239 phenomenon that happens with us when we... Anything 91 00:03:52,239 --> 00:03:54,314 that p ports to be human. 92 00:03:56,723 --> 00:03:59,592 That when it becomes close to human, we 93 00:03:59,592 --> 00:04:01,744 start to have rev. So for instance, if 94 00:04:01,744 --> 00:04:03,817 you put... If you put on a on 95 00:04:03,817 --> 00:04:07,341 a charge, on the x axis, our empathy 96 00:04:07,341 --> 00:04:08,697 towards the thing that looks human. 97 00:04:09,574 --> 00:04:11,190 And on the 98 00:04:11,967 --> 00:04:13,264 y axis 99 00:04:14,200 --> 00:04:14,254 on 100 00:04:16,053 --> 00:04:18,151 excuse me. So so sorry. The thing that 101 00:04:18,770 --> 00:04:20,389 how does it go? The 102 00:04:22,206 --> 00:04:23,985 the x, the y axis 103 00:04:24,459 --> 00:04:24,959 is 104 00:04:25,894 --> 00:04:28,606 our empathy towards something and the x axis 105 00:04:28,606 --> 00:04:30,860 is the closest it looks towards it being 106 00:04:30,919 --> 00:04:31,956 actually photographic. 107 00:04:32,514 --> 00:04:33,811 K? As you get 108 00:04:34,282 --> 00:04:36,426 as you get... As you go towards human, 109 00:04:36,585 --> 00:04:37,935 it gets, it goes up and up and 110 00:04:37,935 --> 00:04:38,808 up and up and up and up and 111 00:04:38,808 --> 00:04:40,079 up and up. The more it looks like 112 00:04:40,079 --> 00:04:42,302 human. But when you get close to it 113 00:04:42,302 --> 00:04:42,802 being 114 00:04:43,828 --> 00:04:45,742 exactly human, but not quite human. 115 00:04:46,380 --> 00:04:47,258 It plummet. 116 00:04:48,135 --> 00:04:51,267 We have an immediate rev. Look like, what 117 00:04:51,325 --> 00:04:52,921 is that? It is not human. 118 00:04:53,653 --> 00:04:55,488 And the most... And there are a lot 119 00:04:55,488 --> 00:04:56,307 of lot of 120 00:04:56,924 --> 00:04:59,795 theories as to why we have this response 121 00:04:59,795 --> 00:05:01,709 as human beings and the most interesting 1 122 00:05:01,709 --> 00:05:02,587 that I've come across, 123 00:05:03,720 --> 00:05:05,980 is that it's a hold over 124 00:05:06,519 --> 00:05:07,339 from when 125 00:05:08,199 --> 00:05:10,860 there were multiple species of h. 126 00:05:13,014 --> 00:05:13,335 When... 127 00:05:14,375 --> 00:05:16,855 There were not homo sapiens, but there were 128 00:05:16,855 --> 00:05:17,355 homo, 129 00:05:18,295 --> 00:05:18,795 Homo, 130 00:05:21,908 --> 00:05:22,568 there was 131 00:05:23,426 --> 00:05:23,926 pom. 132 00:05:24,864 --> 00:05:27,660 There were multiple species of people. 133 00:05:28,554 --> 00:05:30,333 That were not our species 134 00:05:30,952 --> 00:05:32,171 that we were constantly 135 00:05:32,630 --> 00:05:33,829 in conflict with. 136 00:05:34,628 --> 00:05:36,546 But they look kind of, I guess, 137 00:05:39,915 --> 00:05:40,973 and that that 138 00:05:41,671 --> 00:05:42,889 something in that 139 00:05:43,667 --> 00:05:46,221 feels true to me. I'm not sure if 140 00:05:46,221 --> 00:05:46,881 it is 141 00:05:47,352 --> 00:05:49,261 the explanation, but it feels right. 142 00:05:51,409 --> 00:05:52,284 I disagree. 143 00:05:54,034 --> 00:05:55,227 Okay. What's your theory degree? 144 00:05:56,512 --> 00:05:57,861 Oh, I have to replace it with my 145 00:05:57,861 --> 00:06:00,163 own theory. I thought just. Yes. I thought 146 00:06:00,242 --> 00:06:01,591 I could just disagree to disagree. 147 00:06:02,861 --> 00:06:03,734 No. This is mine. 148 00:06:04,545 --> 00:06:05,205 I think 149 00:06:06,064 --> 00:06:06,625 we're such... 150 00:06:07,665 --> 00:06:10,324 I mean, I don't think people understand how 151 00:06:10,944 --> 00:06:14,397 what expert face readers we are. As a 152 00:06:14,477 --> 00:06:15,616 Yeah. A species 153 00:06:16,073 --> 00:06:17,829 so much comes to our face. We are 154 00:06:17,829 --> 00:06:20,143 so weird. This is why you stare at 155 00:06:20,143 --> 00:06:21,361 anything long enough 156 00:06:22,154 --> 00:06:23,435 wood grain, clouds, 157 00:06:24,714 --> 00:06:26,634 a bag of trash, you start seeing faces 158 00:06:26,634 --> 00:06:29,387 in it because we are... Yeah. Face 159 00:06:30,157 --> 00:06:32,855 recognition machines. We're so good at it. Sure 160 00:06:32,855 --> 00:06:35,473 there's so much animals. Right? Totally. The most 161 00:06:35,473 --> 00:06:37,163 social what even more than ants, 162 00:06:38,441 --> 00:06:39,240 because they don't have faces. 163 00:06:39,959 --> 00:06:40,199 So, 164 00:06:41,078 --> 00:06:42,915 oh, just stop myself from going up an 165 00:06:42,915 --> 00:06:43,075 attention, 166 00:06:44,126 --> 00:06:44,762 Okay. I'm back. 167 00:06:48,980 --> 00:06:50,730 Faces is not chemicals. So it's not pheromones 168 00:06:50,730 --> 00:06:52,481 mo. We do it with faces We're visual. 169 00:06:53,294 --> 00:06:54,492 Because you can do that at a distance, 170 00:06:54,652 --> 00:06:55,691 and that's important. 171 00:06:56,890 --> 00:06:58,088 But if anything is wrong, 172 00:06:59,686 --> 00:07:01,125 and I'm thinking it's not... 173 00:07:01,778 --> 00:07:03,770 A hold over from way back when with 174 00:07:03,770 --> 00:07:05,763 multiple species. I think it's disease. 175 00:07:06,560 --> 00:07:07,539 I think it's 176 00:07:08,713 --> 00:07:09,452 the sign 177 00:07:09,764 --> 00:07:11,865 that something is wrong with somebody. 178 00:07:12,404 --> 00:07:14,345 The first thing we think of is 179 00:07:15,045 --> 00:07:15,285 disease. 180 00:07:16,898 --> 00:07:17,635 This person 181 00:07:18,403 --> 00:07:20,702 is injured in some way. They're sick in 182 00:07:20,702 --> 00:07:21,177 some way. 183 00:07:22,683 --> 00:07:24,165 And that lisa fact 184 00:07:24,602 --> 00:07:27,146 spray he. Yeah. Right. Yeah. Yeah. And that 185 00:07:27,146 --> 00:07:29,770 leads all kinds of other fear legends and 186 00:07:29,770 --> 00:07:32,290 myths, you know, like vampires and were and 187 00:07:32,964 --> 00:07:35,187 or where disease calls is weird behavior. 188 00:07:36,378 --> 00:07:37,092 We're just simply... 189 00:07:37,728 --> 00:07:39,315 I I think maybe we're overthinking it, 190 00:07:40,364 --> 00:07:41,422 as simple as 191 00:07:41,959 --> 00:07:43,315 those moments before death, 192 00:07:43,953 --> 00:07:45,091 you know, when 193 00:07:45,628 --> 00:07:48,021 the animation in life has gone out of 194 00:07:48,021 --> 00:07:49,935 someone's face, and we'd know what that means. 195 00:07:50,827 --> 00:07:51,327 That's, 196 00:07:51,864 --> 00:07:53,938 you know, we recognize death when it's coming, 197 00:07:54,576 --> 00:07:56,330 and it's that close. And I think that 198 00:07:56,330 --> 00:07:58,346 when we look at these sort of not 199 00:07:58,404 --> 00:07:59,382 perfectly rendered 200 00:08:01,918 --> 00:08:02,815 Cg faces 201 00:08:03,744 --> 00:08:05,888 show up or robots or something, 202 00:08:06,443 --> 00:08:06,840 we go, 203 00:08:07,967 --> 00:08:08,364 M. 204 00:08:08,920 --> 00:08:11,700 That that's a... That's that's a talking corpse. 205 00:08:13,050 --> 00:08:14,797 Right. Alright right. Yeah. No. No. Thank you. 206 00:08:14,956 --> 00:08:15,512 No. Thank you. 207 00:08:16,722 --> 00:08:18,396 And we're so good at it because they're 208 00:08:18,396 --> 00:08:20,310 very best. I mean, for example, 209 00:08:21,107 --> 00:08:22,941 you go to like Star Wars and they've 210 00:08:22,941 --> 00:08:23,978 been recreating, 211 00:08:25,108 --> 00:08:27,173 you, like, Peter Cushing character, 212 00:08:27,888 --> 00:08:29,398 grandma talking. Right? Yeah. 213 00:08:30,113 --> 00:08:31,066 That was very good. 214 00:08:32,099 --> 00:08:33,211 I think... I mean, the think but the 215 00:08:33,211 --> 00:08:35,535 take methodology that went into that. It's it's 216 00:08:35,535 --> 00:08:35,854 impressive, 217 00:08:36,810 --> 00:08:37,766 but something's wrong. 218 00:08:38,721 --> 00:08:39,221 And 219 00:08:39,677 --> 00:08:41,589 yeah. It's it's it's hard to even say 220 00:08:41,589 --> 00:08:43,670 what it is, but pathology that's the uncanny 221 00:08:43,670 --> 00:08:46,209 valley you're talking about. You know? Everybody just 222 00:08:46,209 --> 00:08:48,271 listening to right now knows the feeling that 223 00:08:48,271 --> 00:08:50,254 we're talking about. Right? Oh, hundred percent, which 224 00:08:50,254 --> 00:08:51,862 is what I I think that this is 225 00:08:51,862 --> 00:08:52,921 a very helpful 226 00:08:53,379 --> 00:08:55,056 conversation to have it at the top of 227 00:08:55,056 --> 00:08:56,274 this particular 228 00:08:56,652 --> 00:08:59,925 so... Yes. So that you're keyed into... 229 00:09:00,579 --> 00:09:02,891 What's gonna happen to you as you're listening 230 00:09:02,891 --> 00:09:04,485 to what we're about to to show you. 231 00:09:04,725 --> 00:09:06,341 Like you said, it's the is the 232 00:09:06,798 --> 00:09:08,472 word of the year Everyone's was curious about 233 00:09:08,472 --> 00:09:08,631 this. 234 00:09:09,444 --> 00:09:09,944 And, 235 00:09:10,959 --> 00:09:13,192 you know, I played with Ai generative visual 236 00:09:13,192 --> 00:09:15,107 tools a little bit, but they were just 237 00:09:15,107 --> 00:09:17,282 kinda like toys, you know, you just wanna 238 00:09:17,579 --> 00:09:18,855 experiment a little and see what happens. 239 00:09:20,388 --> 00:09:22,225 But I found on Youtube, but I don't 240 00:09:22,225 --> 00:09:23,903 remember what called my attention to this, 241 00:09:25,580 --> 00:09:26,080 but 242 00:09:26,474 --> 00:09:27,213 the work 243 00:09:27,908 --> 00:09:29,126 of a 244 00:09:29,582 --> 00:09:30,082 finish 245 00:09:31,017 --> 00:09:31,517 filmmaker, 246 00:09:32,133 --> 00:09:34,148 and an all Ai filmmaker, 247 00:09:35,175 --> 00:09:35,891 named Europa. 248 00:09:37,481 --> 00:09:37,981 And 249 00:09:38,753 --> 00:09:40,740 I really like what I was seeing, because 250 00:09:40,899 --> 00:09:42,648 I think he embraces the uncanny Valley party. 251 00:09:42,807 --> 00:09:43,761 He's not trying to fool you. 252 00:09:44,412 --> 00:09:46,183 Yeah. He's treating it as 253 00:09:46,716 --> 00:09:49,099 an expressive medium as an art, and he's 254 00:09:49,099 --> 00:09:51,165 embracing all of the weird quirks and all 255 00:09:51,165 --> 00:09:53,888 the weirdness that comes with not really fully 256 00:09:53,888 --> 00:09:56,365 being able to control this technology right now. 257 00:09:58,203 --> 00:09:58,523 And 258 00:09:59,801 --> 00:10:01,020 he was very kind 259 00:10:01,414 --> 00:10:04,049 to join us from Helsinki and talk with 260 00:10:04,049 --> 00:10:04,927 us for a few hours. 261 00:10:06,125 --> 00:10:07,961 I wanna make a note here before we 262 00:10:07,961 --> 00:10:08,940 jump into 263 00:10:09,478 --> 00:10:10,675 the interview proper. 264 00:10:11,249 --> 00:10:11,648 Which is, 265 00:10:12,685 --> 00:10:13,562 Ro is finished, 266 00:10:14,280 --> 00:10:14,780 and, 267 00:10:15,396 --> 00:10:16,114 you know, he has, 268 00:10:16,832 --> 00:10:20,023 you know, he's not as fluent and fluid 269 00:10:20,023 --> 00:10:21,034 with English 270 00:10:22,667 --> 00:10:24,100 as I am with Finnish, I don't know 271 00:10:24,100 --> 00:10:25,772 a single word. So he was the 1. 272 00:10:25,931 --> 00:10:26,989 I want people to be 273 00:10:27,842 --> 00:10:30,246 sensitive to the fact that In this interview, 274 00:10:30,406 --> 00:10:32,397 Europa is doing all the heavy lifting. 275 00:10:33,752 --> 00:10:37,120 He is trying to explain incredibly difficult 276 00:10:37,895 --> 00:10:38,851 ideas and concepts, 277 00:10:39,343 --> 00:10:41,491 And, and not his first language. 278 00:10:42,287 --> 00:10:42,684 And... 279 00:10:43,241 --> 00:10:45,548 But he's very articulate. He knows... He he 280 00:10:45,548 --> 00:10:47,560 really made me think differently 281 00:10:48,349 --> 00:10:51,305 oh, about the subject of Ai generative art, 282 00:10:51,704 --> 00:10:52,903 give me a lot to think about. I'm 283 00:10:52,903 --> 00:10:54,840 still thinking about it. We recorded this 284 00:10:55,233 --> 00:10:57,940 6 months ago, and I'm still chewing on 285 00:10:57,940 --> 00:10:58,099 this. 286 00:11:00,089 --> 00:11:02,557 And 1 last thing, again, on the language 287 00:11:02,557 --> 00:11:02,795 front, 288 00:11:04,245 --> 00:11:05,840 this is my way of saying that some 289 00:11:05,840 --> 00:11:07,775 people like to listen to this podcast 290 00:11:08,710 --> 00:11:10,305 while, like, knowing the lawn or doing the 291 00:11:10,305 --> 00:11:12,971 dishes, you're gonna need to listen 292 00:11:13,346 --> 00:11:13,846 list 293 00:11:14,778 --> 00:11:16,926 at this 1 And And and also, we're 294 00:11:16,926 --> 00:11:19,631 gonna have a Youtube version. You might wanna 295 00:11:19,631 --> 00:11:22,452 really consider watch the youtube version 296 00:11:22,906 --> 00:11:25,949 to see what we're talking about. A hundred 297 00:11:26,085 --> 00:11:27,913 percent because there's gonna be passages where you're, 298 00:11:27,992 --> 00:11:29,422 you gonna... If you'd list just listed to 299 00:11:29,422 --> 00:11:31,263 this thing you're just gonna hear weird noises 300 00:11:31,263 --> 00:11:32,849 for a while, and then a bunch of 301 00:11:32,849 --> 00:11:35,467 people talking about what something looks like. And 302 00:11:35,467 --> 00:11:37,546 that's not gonna be very helpful. So we 303 00:11:37,546 --> 00:11:39,534 should suggest that you watch it. 304 00:11:40,806 --> 00:11:41,044 Yeah. 305 00:11:43,032 --> 00:11:44,225 And with no further ado. 306 00:11:45,274 --> 00:11:46,392 Rope N everyone. 307 00:11:54,628 --> 00:11:56,544 2 years ago 2 half a years ago, 308 00:11:56,783 --> 00:11:57,682 I had my 309 00:11:58,140 --> 00:12:00,375 omni life crisis so to speak. So I 310 00:12:00,375 --> 00:12:03,105 turned in 41, and I thought, well, So 311 00:12:03,264 --> 00:12:05,205 I've been alright short stories, 312 00:12:06,945 --> 00:12:07,825 writing poems. 313 00:12:08,465 --> 00:12:10,720 I've been taking pictures for... 30 years. 314 00:12:11,360 --> 00:12:13,759 I've played in balance. I've written Songs. 315 00:12:14,559 --> 00:12:16,639 In school, I he made some fields, 316 00:12:17,533 --> 00:12:19,681 If I don't, you know, do it now, 317 00:12:20,158 --> 00:12:21,988 then I will then grow all that I 318 00:12:22,147 --> 00:12:23,579 I will be extremely upset. 319 00:12:24,709 --> 00:12:26,783 So 2 and a half years ago Then 320 00:12:26,783 --> 00:12:27,421 quit my job. 321 00:12:28,459 --> 00:12:31,172 And it just so happens that that's in 322 00:12:31,172 --> 00:12:31,970 20 21. 323 00:12:32,384 --> 00:12:33,982 Well, I thought I was pretty sure as, 324 00:12:34,062 --> 00:12:36,858 you know, there will be huge new potential 325 00:12:36,858 --> 00:12:38,855 in the way people then can create. 326 00:12:39,508 --> 00:12:41,416 I started to learn to use Ai and 327 00:12:41,496 --> 00:12:44,596 I started to learn then screen writing. 328 00:12:45,311 --> 00:12:46,901 So I saw first, 329 00:12:48,268 --> 00:12:49,484 examples in 330 00:12:50,416 --> 00:12:53,677 of, then August 20 21. Yeah. But it's 331 00:12:53,677 --> 00:12:54,210 this kind of 332 00:12:54,885 --> 00:12:57,826 exponential growth so. If there's exponential growth, 333 00:12:58,462 --> 00:13:00,608 it's slight slow in the first years and 334 00:13:00,608 --> 00:13:02,777 then it then starts to explore. So 335 00:13:03,802 --> 00:13:04,939 where do you think 336 00:13:06,268 --> 00:13:08,575 we're going to be in a couple of 337 00:13:08,575 --> 00:13:09,791 years with this technology. 338 00:13:10,659 --> 00:13:12,097 I mean, right now, we're gonna get into 339 00:13:12,097 --> 00:13:14,575 this later. I wanna talk about things like 340 00:13:14,575 --> 00:13:16,732 temporal instability and all these things that make 341 00:13:16,732 --> 00:13:18,191 the Ai generated. 342 00:13:19,223 --> 00:13:20,101 Stuff so freaky. 343 00:13:20,820 --> 00:13:22,677 Like, when do you think it will get 344 00:13:23,055 --> 00:13:24,114 something that looks 345 00:13:24,971 --> 00:13:26,727 real and do we want to? Is that 346 00:13:26,727 --> 00:13:27,820 even the point 347 00:13:28,654 --> 00:13:28,813 Well. 348 00:13:29,847 --> 00:13:30,483 Those are 349 00:13:31,039 --> 00:13:33,583 michelin 2 questions because they are not. They 350 00:13:33,583 --> 00:13:35,650 don't perhaps have the same as or if, 351 00:13:35,824 --> 00:13:36,324 Ai 352 00:13:36,782 --> 00:13:39,038 reaches such a stage that it then star 353 00:13:39,176 --> 00:13:41,890 star stars then to train itself real. Then 354 00:13:41,890 --> 00:13:44,205 the speed can even more than exponential growth. 355 00:13:44,938 --> 00:13:46,615 Then to start to speed it up even 356 00:13:46,615 --> 00:13:49,489 more. We are now in September 20 23. 357 00:13:50,048 --> 00:13:52,762 So in in September 20 24. I'm pretty 358 00:13:52,762 --> 00:13:54,174 sure that, you know, pictures you 359 00:13:54,774 --> 00:13:56,073 you create will look 360 00:13:57,251 --> 00:14:00,287 in a state that most people probably can't 361 00:14:00,287 --> 00:14:00,447 tell. 362 00:14:01,166 --> 00:14:02,625 You know, is this a 363 00:14:03,019 --> 00:14:03,339 Corona. 364 00:14:05,334 --> 00:14:07,431 It has been only 6 months 365 00:14:07,809 --> 00:14:10,763 since Made that prediction, and that moment has 366 00:14:10,763 --> 00:14:11,481 already come to pass. 367 00:14:12,200 --> 00:14:12,254 Corona 368 00:14:14,757 --> 00:14:16,584 Look at these generated pictures, 369 00:14:17,537 --> 00:14:19,126 even I have gotten to the point where 370 00:14:19,285 --> 00:14:20,556 I can't tell. 371 00:14:21,207 --> 00:14:21,684 The difference. 372 00:14:23,593 --> 00:14:25,264 All of these ai tools can be used 373 00:14:25,264 --> 00:14:27,784 for 2 things. You don't either try to 374 00:14:27,983 --> 00:14:30,527 or replicate the existing world in some. Or 375 00:14:30,527 --> 00:14:32,117 you don't use it there to create something 376 00:14:32,117 --> 00:14:34,979 that does not exist today. So it's not 377 00:14:34,979 --> 00:14:35,797 a replica. 378 00:14:36,251 --> 00:14:37,069 You don't probably 379 00:14:37,778 --> 00:14:38,757 create something 380 00:14:39,933 --> 00:14:40,252 that, 381 00:14:40,890 --> 00:14:41,848 we find interesting, 382 00:14:42,486 --> 00:14:44,800 but that nobody has didn't seen before. 383 00:14:45,852 --> 00:14:48,560 I'm personally the quite excited about them this 384 00:14:48,560 --> 00:14:51,450 side. So it's not just a way to 385 00:14:52,225 --> 00:14:54,296 make the same thing or cheaper. 386 00:14:54,788 --> 00:14:57,016 Or, you know, or debate a fate version 387 00:14:57,016 --> 00:14:59,323 of the same thing, But it's actually a 388 00:14:59,323 --> 00:15:00,994 tool which you can use is to create 389 00:15:00,994 --> 00:15:01,494 something 390 00:15:01,964 --> 00:15:04,437 that nobody has to yet seen time to 391 00:15:04,437 --> 00:15:04,915 wake up. 392 00:15:08,345 --> 00:15:09,143 Where am I? 393 00:15:20,882 --> 00:15:21,815 We had a little festival 394 00:15:22,252 --> 00:15:22,411 here. 395 00:15:23,127 --> 00:15:23,627 And 396 00:15:24,082 --> 00:15:25,991 watch everything on your channel, 397 00:15:27,423 --> 00:15:28,321 and the 398 00:15:29,189 --> 00:15:31,027 And the closer it gets to real sort 399 00:15:31,027 --> 00:15:32,386 of that... Which I know is not you 400 00:15:32,386 --> 00:15:34,543 what you're going for, but when you end 401 00:15:34,543 --> 00:15:36,141 up with stuff that's generated that has all, 402 00:15:36,301 --> 00:15:38,151 you know, correct... When it's all... Works out 403 00:15:38,151 --> 00:15:40,298 for correctly for a while. It kind of 404 00:15:40,298 --> 00:15:41,808 is not as as interesting. 405 00:15:42,365 --> 00:15:44,296 Because everything else is like nothing 406 00:15:44,750 --> 00:15:45,959 anyone has ever seen for. 407 00:15:55,244 --> 00:15:58,934 Now the signature of generative video is the 408 00:15:59,152 --> 00:16:00,986 instability, the temporal and instability. 409 00:16:01,624 --> 00:16:03,538 And that right now is the most unique 410 00:16:03,538 --> 00:16:05,309 thing about it. I think I what catches 411 00:16:05,309 --> 00:16:05,809 everyone's 412 00:16:06,188 --> 00:16:08,664 eyeballs eyeballs right now, and it is like 413 00:16:08,984 --> 00:16:10,603 Lsd. It's almost the same 414 00:16:11,061 --> 00:16:11,381 experience. 415 00:16:11,940 --> 00:16:12,440 It's 416 00:16:13,232 --> 00:16:15,622 nothing stands still. I had so many thoughts 417 00:16:15,622 --> 00:16:16,679 thinking about 418 00:16:17,135 --> 00:16:19,286 what this mimics, what this looks like. 419 00:16:20,255 --> 00:16:21,684 I can look into it and I can 420 00:16:21,684 --> 00:16:22,184 see 421 00:16:22,875 --> 00:16:23,375 the 422 00:16:23,827 --> 00:16:24,327 Ai 423 00:16:25,256 --> 00:16:25,732 digesting. 424 00:16:26,129 --> 00:16:28,193 It has a digestive look to it as 425 00:16:28,193 --> 00:16:28,272 it's... 426 00:16:29,082 --> 00:16:32,103 As it's pulling sources from everything from its 427 00:16:32,103 --> 00:16:33,318 library from its learning 428 00:16:33,692 --> 00:16:35,839 and spending something back out. It looks like 429 00:16:35,839 --> 00:16:37,826 digestion. It looks like an Lsd trip, 430 00:16:38,479 --> 00:16:38,979 it's 431 00:16:39,357 --> 00:16:41,831 very hard not to watch and be slightly 432 00:16:41,831 --> 00:16:42,331 horrified. 433 00:16:44,306 --> 00:16:46,541 Because the lack of stability, 434 00:16:47,193 --> 00:16:47,693 is 435 00:16:49,496 --> 00:16:49,734 upsetting, 436 00:16:50,449 --> 00:16:50,846 I think. 437 00:16:52,037 --> 00:16:53,093 And it reminds me 438 00:16:53,466 --> 00:16:55,451 read this article a couple months ago, somebody 439 00:16:55,451 --> 00:16:55,690 did a... 440 00:16:56,819 --> 00:16:58,977 An article on famous winds, 441 00:16:59,536 --> 00:17:01,294 wind to pattern throughout the globe. 442 00:17:01,934 --> 00:17:03,452 And every now and then, sometimes it's a 443 00:17:03,452 --> 00:17:05,710 yearly, sometimes just 7 years or whatever 444 00:17:06,822 --> 00:17:08,976 some part of the world will have these 445 00:17:08,976 --> 00:17:10,730 crazy wins that come through, 446 00:17:11,368 --> 00:17:13,304 and those moments are always associated 447 00:17:13,682 --> 00:17:15,059 with higher incidence 448 00:17:15,436 --> 00:17:15,936 of 449 00:17:16,567 --> 00:17:16,726 homicide. 450 00:17:19,112 --> 00:17:21,919 And there is something ag agitated 451 00:17:22,612 --> 00:17:24,306 about nothing standing still 452 00:17:25,094 --> 00:17:27,512 and it gets to us at a really 453 00:17:27,971 --> 00:17:28,930 really deep level. 454 00:17:29,809 --> 00:17:30,469 I spent 455 00:17:31,168 --> 00:17:32,307 an hour watching, 456 00:17:32,686 --> 00:17:34,856 yeah, your films last night. On a big 457 00:17:34,856 --> 00:17:36,368 screen, big sound system, 458 00:17:37,403 --> 00:17:37,903 and 459 00:17:38,676 --> 00:17:39,574 it did something 460 00:17:40,029 --> 00:17:42,497 to my head. It did something neurological to 461 00:17:42,497 --> 00:17:43,554 my visual processing 462 00:17:44,182 --> 00:17:44,682 sis 463 00:17:45,134 --> 00:17:45,452 system. 464 00:17:46,245 --> 00:17:47,832 And then we said, okay. Well, and that's 465 00:17:47,832 --> 00:17:49,498 starting to get to. Let's watch something else 466 00:17:49,498 --> 00:17:49,657 now. 467 00:17:50,371 --> 00:17:51,743 Now everything else 468 00:17:52,766 --> 00:17:55,145 I was... I... Everything else was suspect after 469 00:17:55,145 --> 00:17:56,755 that all traditionally 470 00:17:57,444 --> 00:17:59,744 produced video. I was like, I don't know. 471 00:18:00,632 --> 00:18:02,378 I don't know. Now. He's gonna do something. 472 00:18:02,696 --> 00:18:06,348 It's gonna do something. Expected. My my brain 473 00:18:06,348 --> 00:18:07,697 didn't trust my eyes anymore. 474 00:18:09,228 --> 00:18:11,545 After after watching it for maybe too long, 475 00:18:12,504 --> 00:18:14,502 those are my feelings, and, now would like 476 00:18:14,502 --> 00:18:15,940 to hear what y'all think about that. 477 00:18:18,355 --> 00:18:20,514 I've been out then staring up these videos 478 00:18:20,514 --> 00:18:22,914 for like, many years. So III 479 00:18:22,914 --> 00:18:25,369 see it all, man. So I I think 480 00:18:25,568 --> 00:18:28,033 It's it's my mind is kind of lighting 481 00:18:28,033 --> 00:18:28,749 in a very... 482 00:18:29,385 --> 00:18:31,770 I didn't appreciate the state of, like, being 483 00:18:31,770 --> 00:18:33,145 able in the questionnaire. 484 00:18:33,614 --> 00:18:34,012 Question it. 485 00:18:34,808 --> 00:18:37,038 Everything nicely see. I put my videos on 486 00:18:37,038 --> 00:18:39,825 the Tiktok, and, the number 1 comment is 487 00:18:39,825 --> 00:18:41,976 kind like. It's slight a fever dream or 488 00:18:41,976 --> 00:18:42,809 it's It's, 489 00:18:43,823 --> 00:18:44,063 dreams. 490 00:18:45,816 --> 00:18:48,606 People may say that Rope films feel or 491 00:18:48,606 --> 00:18:49,801 look like fever dreams. 492 00:18:50,693 --> 00:18:52,310 But that comparison 493 00:18:52,928 --> 00:18:55,502 suggests a stream of consciousness or 494 00:18:56,917 --> 00:18:58,789 automation that it is not 495 00:18:59,170 --> 00:19:00,850 part of his filmmaking process. 496 00:19:01,650 --> 00:19:03,509 It is not automated, the technology 497 00:19:03,890 --> 00:19:05,910 does not exist yet 498 00:19:06,529 --> 00:19:08,778 to think of an idea and have the 499 00:19:08,778 --> 00:19:11,799 computer generated from, say, a brain implant. 500 00:19:12,833 --> 00:19:15,639 So where do Rope ideas come from 501 00:19:15,949 --> 00:19:16,926 this kind of a 502 00:19:17,300 --> 00:19:19,209 boring question to ask any artist who knows. 503 00:19:19,606 --> 00:19:21,673 But this case, what is interesting is how 504 00:19:21,673 --> 00:19:24,059 does he turn those ideas into sight and 505 00:19:24,059 --> 00:19:24,297 sound. 506 00:19:30,579 --> 00:19:31,768 At the end of the day, you know, 507 00:19:32,006 --> 00:19:32,878 you want to tell. 508 00:19:33,448 --> 00:19:35,853 L story. So be it a a cartoon 509 00:19:35,909 --> 00:19:37,362 or a movie or, 510 00:19:38,053 --> 00:19:40,752 photos or a book for with the kind 511 00:19:40,752 --> 00:19:42,221 of the Meta. 512 00:19:42,834 --> 00:19:45,459 So, I mean, I a screen play. But 513 00:19:45,459 --> 00:19:47,368 the interesting thing there is that kind of 514 00:19:47,368 --> 00:19:47,868 like 515 00:19:49,457 --> 00:19:49,957 Stream 516 00:19:50,654 --> 00:19:53,927 a move then traditionally if you would write 517 00:19:53,927 --> 00:19:55,523 it, you were probably write it in a 518 00:19:55,523 --> 00:19:57,692 way that you would consider that how can 519 00:19:57,692 --> 00:20:00,073 it be filled. So I mean, it needs 520 00:20:00,073 --> 00:20:02,532 to be realistic in terms of flying, you 521 00:20:02,532 --> 00:20:04,516 know, the real world, but also the budget, 522 00:20:04,849 --> 00:20:06,708 If I want to pay place a story 523 00:20:07,007 --> 00:20:09,085 into the New Year's subway, I can. 524 00:20:09,884 --> 00:20:11,882 I don't need to get permissions in that 525 00:20:11,882 --> 00:20:12,122 way. 526 00:20:13,161 --> 00:20:15,573 He's coming this way. We'll cut the power 527 00:20:15,573 --> 00:20:17,591 once he gets near. What was he talking 528 00:20:17,730 --> 00:20:20,766 about? Sir? Listen to me carefully. There are 529 00:20:20,766 --> 00:20:22,045 no rats in the subway. 530 00:20:25,480 --> 00:20:27,150 It is a new form. So is it 531 00:20:27,150 --> 00:20:27,650 like 532 00:20:28,422 --> 00:20:31,921 film or the theater or you know, it 533 00:20:31,921 --> 00:20:34,030 is what's the difference between 534 00:20:35,205 --> 00:20:35,705 Tie 535 00:20:36,402 --> 00:20:39,592 play a Tv show or a field version 536 00:20:39,592 --> 00:20:41,666 of same thing? It's the same, but it's 537 00:20:41,666 --> 00:20:43,596 different. I do it in the its own 538 00:20:43,596 --> 00:20:43,915 form. 539 00:20:44,474 --> 00:20:46,230 If we draw 5 years to the future, 540 00:20:47,587 --> 00:20:50,540 it is probably possible then go the persistent 541 00:20:50,540 --> 00:20:53,346 forms. Then also quite well. I was about 542 00:20:53,346 --> 00:20:55,737 to say, I'm getting the sense now that 543 00:20:55,817 --> 00:20:58,129 I've seen your work that at least in 544 00:20:58,129 --> 00:20:58,948 terms of 545 00:21:00,611 --> 00:21:01,008 entertainment, 546 00:21:02,200 --> 00:21:03,574 it's not so much 547 00:21:04,346 --> 00:21:06,571 a new tool as it is a whole 548 00:21:06,571 --> 00:21:07,071 new 549 00:21:07,445 --> 00:21:07,763 instrument, 550 00:21:09,050 --> 00:21:09,210 Yep. 551 00:21:10,005 --> 00:21:13,268 Yeah. And and 1 unlike, we've ever seen, 552 00:21:13,507 --> 00:21:14,961 like, a complete software 553 00:21:15,337 --> 00:21:15,655 instrument. 554 00:21:16,785 --> 00:21:18,615 What... What I'm curious about is, 555 00:21:19,730 --> 00:21:21,242 first of all, I I do wanna I 556 00:21:21,401 --> 00:21:22,755 I do wanna get some sense of your 557 00:21:22,755 --> 00:21:25,724 your process about... Your inputs, your specific inputs. 558 00:21:26,445 --> 00:21:26,684 But, 559 00:21:27,644 --> 00:21:28,384 in watching 560 00:21:28,924 --> 00:21:29,744 your videos, 561 00:21:31,658 --> 00:21:34,765 the difference between, say, the meta morph, which 562 00:21:34,765 --> 00:21:36,381 is highly narrative 563 00:21:37,553 --> 00:21:40,218 story. Right? And the music videos 564 00:21:40,912 --> 00:21:42,208 is so 565 00:21:42,823 --> 00:21:44,972 different. There seems to be something in the 566 00:21:44,972 --> 00:21:46,188 music videos that 567 00:21:47,055 --> 00:21:49,279 I guess, because there's an existing piece of 568 00:21:49,279 --> 00:21:51,742 material it's able to lock onto to that. 569 00:21:52,616 --> 00:21:54,522 And so there's a there's a cohesion in 570 00:21:54,522 --> 00:21:55,952 the flow of the music videos. 571 00:21:57,396 --> 00:21:57,896 That 572 00:21:58,427 --> 00:22:00,490 isn't really there yet in the in the 573 00:22:00,490 --> 00:22:02,553 narratives. Am I am I getting that correct? 574 00:22:02,712 --> 00:22:03,640 Is it... Or 575 00:22:04,079 --> 00:22:06,579 are you... Is that what it is? And 576 00:22:06,960 --> 00:22:07,700 are you 577 00:22:08,079 --> 00:22:09,380 inputting the video 578 00:22:09,680 --> 00:22:12,088 first or you telling it, basically telling this 579 00:22:12,088 --> 00:22:14,634 software go find this video and do something 580 00:22:14,634 --> 00:22:14,952 to it. 581 00:22:15,668 --> 00:22:17,816 So you're gonna either create any videos. 582 00:22:18,390 --> 00:22:20,009 So it then a references 583 00:22:20,470 --> 00:22:20,970 on 584 00:22:21,830 --> 00:22:24,789 existing video. So say you have a video 585 00:22:24,789 --> 00:22:25,509 of something. 586 00:22:26,004 --> 00:22:28,072 You don't then hand that video over to 587 00:22:28,072 --> 00:22:30,379 the Ai. So you have some sense of 588 00:22:30,379 --> 00:22:32,606 control, but you don't still make changes. 589 00:22:33,099 --> 00:22:35,097 Then the other way you gotta create is 590 00:22:35,097 --> 00:22:37,973 that you don't have the Ai a video, 591 00:22:38,373 --> 00:22:40,610 but to hand it different input. So you 592 00:22:40,610 --> 00:22:43,172 don't had a picture then to start with. 593 00:22:44,047 --> 00:22:45,637 You can add some text prompts. 594 00:22:46,512 --> 00:22:48,600 You can add it some kind of, 595 00:22:49,640 --> 00:22:50,140 additional 596 00:22:50,600 --> 00:22:51,400 directions. You can tell. 597 00:22:52,119 --> 00:22:53,000 Here's a picture. 598 00:22:54,279 --> 00:22:57,033 Please move the camera to the right slowly 599 00:22:57,489 --> 00:22:59,002 and then you the right the prompt light, 600 00:22:59,559 --> 00:23:03,165 personal talking. And and then then it then 601 00:23:03,382 --> 00:23:04,837 tries to take the picture 602 00:23:05,149 --> 00:23:05,950 find a person, 603 00:23:06,349 --> 00:23:08,909 then they make them talk, and then they 604 00:23:08,909 --> 00:23:10,909 move the camera to the right. But it 605 00:23:10,909 --> 00:23:13,149 has no additional on their inputs. 606 00:23:13,803 --> 00:23:16,195 So basically, those those those 2 ways are 607 00:23:16,195 --> 00:23:18,190 the main ways you go at this moment 608 00:23:18,190 --> 00:23:20,263 then create videos. Yeah. So the... Because that 609 00:23:20,343 --> 00:23:21,858 I was gonna say that that 610 00:23:23,470 --> 00:23:25,009 If you were to take that 611 00:23:25,950 --> 00:23:28,109 smashing pumpkins video of Share Brock. 612 00:23:37,958 --> 00:23:40,270 And look go back in time and just 613 00:23:40,270 --> 00:23:42,365 sort of plop that down 614 00:23:43,220 --> 00:23:46,144 in the zeitgeist of 19 94. It at 615 00:23:46,343 --> 00:23:48,810 absolutely without a doubt would have won best 616 00:23:48,810 --> 00:23:50,560 video of the year. No question. 617 00:23:51,276 --> 00:23:52,470 And the same for for the, 618 00:23:54,238 --> 00:23:56,314 what was the other 1? The the the 619 00:23:56,314 --> 00:23:57,372 offspring video was 620 00:23:57,990 --> 00:23:58,490 amazing. 621 00:24:03,194 --> 00:24:05,108 Mean there's some sort of anime thing going 622 00:24:05,108 --> 00:24:07,421 on there. What what did you... What did 623 00:24:07,421 --> 00:24:09,495 you... What's the... What were the inputs for 624 00:24:09,495 --> 00:24:10,452 the off offspring video? 625 00:24:11,090 --> 00:24:12,946 So in the case of these music videos 626 00:24:13,417 --> 00:24:16,122 you can configure not only tips browser but 627 00:24:16,122 --> 00:24:18,134 you can actually already train 628 00:24:18,509 --> 00:24:20,911 or then fine to then a term. You 629 00:24:20,911 --> 00:24:23,162 can actually, you know, fine tune the Ai 630 00:24:23,378 --> 00:24:24,116 by enough 631 00:24:24,492 --> 00:24:26,720 to work in a specific and resource or 632 00:24:26,720 --> 00:24:27,118 style. 633 00:24:27,675 --> 00:24:29,357 It's or the flip side of this kind 634 00:24:29,357 --> 00:24:29,516 of, 635 00:24:30,627 --> 00:24:33,088 that the more day I then create than 636 00:24:33,088 --> 00:24:34,835 the artery is than the contra at the 637 00:24:34,835 --> 00:24:37,398 same time. The real kind of skill for 638 00:24:37,398 --> 00:24:39,254 an Ai artist is absolutely 639 00:24:39,791 --> 00:24:41,944 control. This is what I really wanna get 640 00:24:41,944 --> 00:24:43,380 into. This is the part that fa makes 641 00:24:43,380 --> 00:24:44,656 me the most as someone who... 642 00:24:46,102 --> 00:24:48,111 Grew up at wearing the dat and 643 00:24:48,802 --> 00:24:50,017 Max Ernst and 644 00:24:50,788 --> 00:24:51,423 the surreal. 645 00:24:52,138 --> 00:24:54,147 They would frequently intentionally 646 00:24:55,014 --> 00:24:55,753 do things 647 00:24:56,291 --> 00:24:56,791 to 648 00:24:58,207 --> 00:24:59,266 put a little distance 649 00:24:59,884 --> 00:25:02,598 between the art and their level of control, 650 00:25:02,997 --> 00:25:04,056 introduce some accident 651 00:25:04,448 --> 00:25:06,515 Yeah. That was exactly my next question. Yeah. 652 00:25:06,674 --> 00:25:08,899 Yeah. Yeah. There was there was a, you 653 00:25:08,899 --> 00:25:10,727 know, from Us and collage and all these 654 00:25:10,727 --> 00:25:12,714 different things of Max ernst would, you know, 655 00:25:13,032 --> 00:25:14,501 squeeze. Plates of 656 00:25:15,828 --> 00:25:18,530 glass plates with paint together, un stick them 657 00:25:18,530 --> 00:25:19,881 and then let that dry and look at 658 00:25:19,881 --> 00:25:21,708 it and say, alright. What does that look 659 00:25:21,708 --> 00:25:25,066 like? And kinda sharpen it using the, you 660 00:25:25,066 --> 00:25:28,011 know, the old kind of hardware, the, brain. 661 00:25:28,569 --> 00:25:30,263 It's pattern in recognition to 662 00:25:31,050 --> 00:25:32,561 flesh out the idea and turn it into 663 00:25:32,561 --> 00:25:35,581 something a little more sc. And you're clearly, 664 00:25:37,171 --> 00:25:38,522 I mean, you're you're down in the trenches 665 00:25:38,522 --> 00:25:40,683 with us all the time. And dealing with 666 00:25:40,683 --> 00:25:42,035 a machine that doesn't... 667 00:25:43,546 --> 00:25:45,397 That that does put a bit of distance 668 00:25:45,454 --> 00:25:48,555 between your idea and whatever that final output 669 00:25:48,555 --> 00:25:50,481 is going to be which is why I 670 00:25:50,481 --> 00:25:51,915 think that your, 671 00:25:52,951 --> 00:25:54,966 sub heading on your Youtube channel 672 00:25:55,342 --> 00:25:56,538 is so fitting, 673 00:25:57,907 --> 00:26:00,452 en sla Ai to make some art. 674 00:26:01,406 --> 00:26:01,906 How 675 00:26:02,758 --> 00:26:05,246 cooperative are these Ai slaves 676 00:26:06,115 --> 00:26:07,474 And how do you make and do your 677 00:26:07,474 --> 00:26:07,714 bidding? 678 00:26:08,595 --> 00:26:11,315 Part of this is a temporary problem software 679 00:26:11,315 --> 00:26:12,774 us, You know, the way 680 00:26:13,075 --> 00:26:15,174 we humans now talk with Ai. 681 00:26:15,806 --> 00:26:16,941 Especially with this product 682 00:26:17,711 --> 00:26:18,187 all tools. 683 00:26:18,902 --> 00:26:21,362 It's the worst study will be... So it 684 00:26:21,362 --> 00:26:22,950 will be much better in 3 months and 685 00:26:22,950 --> 00:26:25,031 months at 12 months. So of course, It's 686 00:26:25,031 --> 00:26:25,429 like, 687 00:26:26,305 --> 00:26:27,841 especially when A film 688 00:26:28,297 --> 00:26:28,797 major 689 00:26:29,173 --> 00:26:31,164 is all by themselves, you know, And they 690 00:26:31,164 --> 00:26:33,804 don't have nobody to talk to, They don't 691 00:26:33,804 --> 00:26:34,700 have spontaneous 692 00:26:35,388 --> 00:26:38,241 inputs that are like then highly than something 693 00:26:38,241 --> 00:26:39,984 else that what they're thinking in themselves. 694 00:26:40,634 --> 00:26:41,033 But now, 695 00:26:41,750 --> 00:26:43,026 especially with the Ai tools, 696 00:26:44,302 --> 00:26:46,455 they it it... They quite often comes back 697 00:26:46,455 --> 00:26:48,927 with such suggestions that are completely light and 698 00:26:48,927 --> 00:26:52,119 brain dead. But they're like braid in a 699 00:26:52,119 --> 00:26:54,419 way. So it it it it really started 700 00:26:54,419 --> 00:26:57,140 to push me. So it... So I see 701 00:26:58,163 --> 00:27:00,309 see the potential there. So I mean, always, 702 00:27:00,469 --> 00:27:01,979 if you had a team of, you know, 703 00:27:02,456 --> 00:27:05,160 often the creative people, it made sense to 704 00:27:05,160 --> 00:27:07,880 have people who have completely different ideas. In 705 00:27:07,880 --> 00:27:10,359 the case of, you know, on Ai, then 706 00:27:10,359 --> 00:27:12,440 you don't have to save personal problems. 707 00:27:12,919 --> 00:27:14,759 So you get results that are like, you 708 00:27:14,759 --> 00:27:16,039 know, sometimes they're fantastic. 709 00:27:16,613 --> 00:27:19,244 Thumb down the stair, completely insane, but you 710 00:27:19,244 --> 00:27:20,838 don't start to hate, you know the person, 711 00:27:21,635 --> 00:27:23,947 and vice. And the person don't start to 712 00:27:23,947 --> 00:27:26,351 hate you, So when you see like than 713 00:27:26,351 --> 00:27:29,137 crazy Ideas each day, it probably it might 714 00:27:29,137 --> 00:27:31,365 that help you. Can you kinda take 1 715 00:27:31,365 --> 00:27:32,240 of your videos? 716 00:27:32,654 --> 00:27:34,407 That that we've seen on your channel and 717 00:27:34,407 --> 00:27:36,559 sort of run us through the process, 718 00:27:37,435 --> 00:27:38,870 give us an example of 1 that you 719 00:27:38,870 --> 00:27:41,116 said, okay. Do this. It came back. It 720 00:27:41,116 --> 00:27:41,515 was like, 721 00:27:42,471 --> 00:27:43,427 what the fuck is at. 722 00:27:44,941 --> 00:27:45,339 And then. 723 00:27:46,056 --> 00:27:47,172 And then you... And then you've... And the... 724 00:27:47,251 --> 00:27:48,765 But then you gave you an idea and 725 00:27:48,765 --> 00:27:50,771 you kept running with it. Can I don't 726 00:27:50,771 --> 00:27:53,000 write a script in all cases? So if 727 00:27:53,000 --> 00:27:55,468 it's a short film, it's then in my 728 00:27:55,468 --> 00:27:57,776 head. The best way to make video is 729 00:27:57,776 --> 00:27:59,980 to is to actually then the reference pick 730 00:28:00,178 --> 00:28:03,599 pictures. So you actually then create huge amount 731 00:28:03,599 --> 00:28:04,576 of pictures 732 00:28:05,031 --> 00:28:07,671 off then from the screen play. So you 733 00:28:07,671 --> 00:28:10,455 try to create each shot. So here, I'm 734 00:28:10,455 --> 00:28:11,353 using this 735 00:28:11,728 --> 00:28:12,603 stable diffusion. 736 00:28:13,478 --> 00:28:16,198 I have my own computers at home. So 737 00:28:16,358 --> 00:28:17,018 I then 738 00:28:17,634 --> 00:28:20,107 leave the computer is their on then for 739 00:28:20,107 --> 00:28:22,420 the night, then I come back in the 740 00:28:22,420 --> 00:28:25,052 guide of the and and I have a 741 00:28:25,052 --> 00:28:26,817 thousand feature. My gosh. 742 00:28:27,850 --> 00:28:29,439 90 percent of the pictures are bad. 743 00:28:30,472 --> 00:28:32,402 5 percent are usable. 744 00:28:33,189 --> 00:28:35,184 And then 5 percent or the 1 percent 745 00:28:35,184 --> 00:28:37,978 has something that I didn't didn't expect. Wow, 746 00:28:38,377 --> 00:28:38,537 okay. 747 00:28:39,335 --> 00:28:41,625 I mean, the perfect example here is 748 00:28:42,062 --> 00:28:43,733 is is is the crow. 749 00:28:44,529 --> 00:28:46,518 So then... When I started story, Item I 750 00:28:46,518 --> 00:28:48,587 then didn't know how it would end yet. 751 00:28:49,400 --> 00:28:52,200 So I I knew what that what the 752 00:28:52,200 --> 00:28:53,960 crow will do in in the story. 753 00:28:54,599 --> 00:28:56,299 But I didn't yet 754 00:28:56,680 --> 00:28:59,186 understand that that the crow will actually then 755 00:28:59,561 --> 00:29:02,502 replace the person. But then I saw these 756 00:29:02,502 --> 00:29:03,376 pictures for a while. 757 00:29:03,853 --> 00:29:05,919 I saw the saw the guy nest in 758 00:29:05,919 --> 00:29:06,237 the house, 759 00:29:06,970 --> 00:29:08,409 And then have after I saw it for 760 00:29:08,409 --> 00:29:10,889 a few days, it's then then my then 761 00:29:10,889 --> 00:29:11,629 my brain 762 00:29:12,089 --> 00:29:14,809 which is sound slow. It then started to 763 00:29:14,809 --> 00:29:17,134 realize. Yes. Of course. That's the way I 764 00:29:17,134 --> 00:29:19,600 should then start to end the kind of 765 00:29:19,600 --> 00:29:22,066 the story. So the whole process was like, 766 00:29:22,878 --> 00:29:24,554 You do dupont, you see these pictures, 767 00:29:25,990 --> 00:29:27,426 and then you find the best ones of 768 00:29:27,426 --> 00:29:29,340 the shots, and I go to step 3, 769 00:29:30,468 --> 00:29:33,402 So most people nowadays, at this moment use 770 00:29:33,402 --> 00:29:33,902 this 771 00:29:34,909 --> 00:29:35,805 service called 772 00:29:36,812 --> 00:29:37,312 out. 773 00:29:37,859 --> 00:29:39,137 And they have this, 774 00:29:40,895 --> 00:29:42,573 feature called then Gen 2... 775 00:29:43,372 --> 00:29:45,935 So basically, the way... The gen 2 works 776 00:29:45,935 --> 00:29:47,682 is... You can feed it a pitcher and 777 00:29:47,682 --> 00:29:48,159 a prompt. 778 00:29:49,112 --> 00:29:51,176 So here you can see from the terminals 779 00:29:51,176 --> 00:29:53,558 thumbnails that they they all start from a 780 00:29:53,558 --> 00:29:56,703 pitcher. Then tries then to create a shot 781 00:29:56,841 --> 00:29:58,356 from the picture and the front. 782 00:29:59,472 --> 00:30:01,067 But, you know, this 1 also have... Has 783 00:30:01,067 --> 00:30:03,221 to same issue as with the picture. So 784 00:30:03,221 --> 00:30:04,903 sometimes that is I didn't see the result, 785 00:30:05,617 --> 00:30:07,916 and I didn't start to think about this 786 00:30:07,916 --> 00:30:10,533 story. I have this. I think there's like 787 00:30:10,533 --> 00:30:12,297 1400 788 00:30:12,297 --> 00:30:14,392 of these, you know, that I made 789 00:30:14,849 --> 00:30:17,161 and in Yeah. And then in the final 790 00:30:17,161 --> 00:30:19,952 1, they flight 800 in our shots How 791 00:30:19,952 --> 00:30:22,274 much of the of your time goes into 792 00:30:22,274 --> 00:30:23,568 just writing the prompts 793 00:30:24,021 --> 00:30:25,133 before did you get the to get the 794 00:30:25,133 --> 00:30:25,610 steel images? 795 00:30:26,404 --> 00:30:28,390 I didn't... Oh, Run I'm through the screen 796 00:30:28,390 --> 00:30:28,890 play 797 00:30:29,439 --> 00:30:31,037 I don't pretty much hate lighten in. 798 00:30:32,794 --> 00:30:35,590 I'm every single line, and, you know, that's 799 00:30:35,590 --> 00:30:36,070 a prompt. 800 00:30:36,722 --> 00:30:38,872 In terms of this over process is like 801 00:30:38,872 --> 00:30:41,102 5 percent of time. So my approach is 802 00:30:41,102 --> 00:30:42,933 kind of a kind of them improved for 803 00:30:42,933 --> 00:30:45,916 us approach. So I'm I would called ben 804 00:30:46,052 --> 00:30:48,836 for for for a kind of the lack 805 00:30:48,836 --> 00:30:49,393 of control, 806 00:30:50,109 --> 00:30:51,621 but, you know, I just made a... 807 00:30:52,588 --> 00:30:55,287 8000 pictures. And then then find the ones 808 00:30:55,287 --> 00:30:55,843 that I like. 809 00:30:57,112 --> 00:31:00,329 Well, specifically, do you go in edit individual 810 00:31:00,470 --> 00:31:02,070 images, let's say, in a different program, 811 00:31:02,869 --> 00:31:05,190 photoshop just kinda help steer it? Or do 812 00:31:05,190 --> 00:31:07,029 you pretty much take those images straight out 813 00:31:07,029 --> 00:31:08,797 of stable diffusion and say, 814 00:31:09,671 --> 00:31:11,739 I can make this work. I don't... I 815 00:31:11,898 --> 00:31:14,283 I have no intention to to keep a 816 00:31:14,363 --> 00:31:17,067 Ai art like, pure in a way. I 817 00:31:17,067 --> 00:31:17,440 don't. 818 00:31:18,838 --> 00:31:21,075 I don't think no. That's not meaningful to 819 00:31:21,075 --> 00:31:22,912 me. All the kind of tricks you know, 820 00:31:23,152 --> 00:31:24,430 or did fe key aims amazon. 821 00:31:25,163 --> 00:31:27,710 So then I do it this shots then 822 00:31:27,790 --> 00:31:30,177 I find the best ones. The interesting part 823 00:31:30,177 --> 00:31:32,008 here is offers that now we... When you 824 00:31:32,008 --> 00:31:33,997 go to the video world, it's not just 825 00:31:33,997 --> 00:31:34,475 video, but it. 826 00:31:35,922 --> 00:31:38,308 Voice music. If you do a short feel 827 00:31:38,308 --> 00:31:40,558 naturally, has spend more time on 828 00:31:41,012 --> 00:31:43,343 the rest than the actual picture and they 829 00:31:43,343 --> 00:31:46,060 got a lot the video parts. So picture 830 00:31:46,060 --> 00:31:47,818 and video at a like 1 third of 831 00:31:47,818 --> 00:31:49,816 the time I spent, and and actually, you 832 00:31:49,816 --> 00:31:51,848 know, 2 thirds both to 833 00:31:52,537 --> 00:31:53,037 sound 834 00:31:53,805 --> 00:31:56,341 speech music sound effects. If if I may, 835 00:31:56,658 --> 00:31:58,164 if I may, I just wanna jump in 836 00:31:58,164 --> 00:32:00,006 and interrupt and just before... Because I don't 837 00:32:00,006 --> 00:32:01,118 get a chance to do it later and 838 00:32:01,118 --> 00:32:03,265 compliment you. I'm a as a film editor, 839 00:32:03,662 --> 00:32:03,821 myself, 840 00:32:04,695 --> 00:32:08,289 your sound design is... Beautiful. Yeah. It's so 841 00:32:08,289 --> 00:32:10,836 good. And it's... And it goes... It does 842 00:32:10,836 --> 00:32:13,941 so much work to contextualize and make all 843 00:32:13,941 --> 00:32:14,998 of those wild 844 00:32:15,453 --> 00:32:15,953 images 845 00:32:17,059 --> 00:32:19,127 stick together, talk to each other that makes 846 00:32:19,127 --> 00:32:21,273 it cohesive. I think your sound work is... 847 00:32:22,069 --> 00:32:23,659 Everyone talks about the Ai, but I've I've 848 00:32:23,659 --> 00:32:25,409 noticed your sound work, buddy is brilliant. 849 00:32:27,095 --> 00:32:28,848 Now are you are are you adding? Are 850 00:32:28,848 --> 00:32:30,361 you at or do... Are you doing the 851 00:32:30,361 --> 00:32:30,999 sound work? 852 00:32:33,403 --> 00:32:35,785 Is that completely built by you or any 853 00:32:35,785 --> 00:32:36,420 of that Ai? 854 00:32:37,214 --> 00:32:38,961 Yeah... Yeah. I mean, I, I do spend 855 00:32:38,961 --> 00:32:40,946 a lot of time on the silent parts. 856 00:32:41,517 --> 00:32:44,216 1 of my obvious is that I... I 857 00:32:44,216 --> 00:32:46,677 am a musician in a way. I've played 858 00:32:46,677 --> 00:32:47,971 some bands. I'd played 859 00:32:48,582 --> 00:32:50,114 guitar some, some 860 00:32:50,901 --> 00:32:53,367 best. It's fun and it's important. So I 861 00:32:53,447 --> 00:32:55,356 I didn't know, I've tried to use some 862 00:32:55,356 --> 00:32:57,903 of these Ai tools. So Ai tools tools 863 00:32:57,903 --> 00:33:00,625 for for music are not get there for 864 00:33:00,625 --> 00:33:00,943 me. 865 00:33:01,979 --> 00:33:04,369 I do use sound... I do use a, 866 00:33:04,528 --> 00:33:07,019 I to create large sound effects. So 867 00:33:07,714 --> 00:33:08,374 we just 868 00:33:09,402 --> 00:33:11,315 that has the potential then they to create 869 00:33:11,315 --> 00:33:13,864 something fantastic. So. If the pictures are, you 870 00:33:13,864 --> 00:33:15,776 know, half this certain vibe. 871 00:33:16,428 --> 00:33:18,415 Then of course, if you use Ai tools 872 00:33:18,415 --> 00:33:21,355 them to create sound, then to accent that 873 00:33:21,355 --> 00:33:21,991 same vibe. 874 00:33:23,580 --> 00:33:24,796 Robot made me realize. 875 00:33:25,585 --> 00:33:28,065 How powerful Ai is going to be for 876 00:33:28,065 --> 00:33:28,704 sound editing. 877 00:33:29,505 --> 00:33:32,224 Because an Ai program can just look at 878 00:33:32,224 --> 00:33:33,525 a scene and 879 00:33:33,984 --> 00:33:36,658 computationally fig hear out what kind of sounds 880 00:33:36,877 --> 00:33:38,714 you would expect to hear in that scene 881 00:33:38,714 --> 00:33:40,392 and from what distance and it would figure 882 00:33:40,392 --> 00:33:42,090 out all of the acoustic characteristics 883 00:33:42,404 --> 00:33:44,499 that. This is going to have 884 00:33:44,957 --> 00:33:45,754 a very, 885 00:33:46,392 --> 00:33:50,162 very disruptive effect on the whole sound editing 886 00:33:50,460 --> 00:33:51,917 sound post production process 887 00:33:53,108 --> 00:33:54,305 We're not quite there yet, 888 00:33:55,104 --> 00:33:55,822 but it's coming. 889 00:33:56,780 --> 00:33:59,415 He's not well. He works too hard. I'm 890 00:33:59,415 --> 00:34:01,650 glad you come, sir. Now he will surely 891 00:34:01,650 --> 00:34:03,178 open the door Mister Sam, 892 00:34:03,815 --> 00:34:05,644 you are neglecting your business duties. 893 00:34:06,201 --> 00:34:07,950 I thought you were a dependable man. 894 00:34:09,223 --> 00:34:10,710 I will open the door or 895 00:34:11,627 --> 00:34:13,461 At the time of this recording back in 896 00:34:13,780 --> 00:34:15,296 September of 2023. 897 00:34:16,093 --> 00:34:18,885 The tools to make Ai generated people move 898 00:34:18,885 --> 00:34:22,255 their mouths in sync. With recorded audio dialogue, 899 00:34:22,494 --> 00:34:25,135 making it was brand new. And Rope films, 900 00:34:25,695 --> 00:34:27,375 Were among of the first a boy to 901 00:34:27,375 --> 00:34:28,894 use it about nothing but as well... I'm 902 00:34:28,894 --> 00:34:31,568 hugely excited about these because they have this 903 00:34:31,626 --> 00:34:33,801 potential that that create something that is quite 904 00:34:34,099 --> 00:34:35,135 quite too strange. 905 00:34:38,724 --> 00:34:39,522 Why this. 906 00:34:40,733 --> 00:34:43,195 Think do catch at black. I'm ready to 907 00:34:43,195 --> 00:34:43,512 go. 908 00:34:44,386 --> 00:34:46,610 Here I'm willing to work? Are you going 909 00:34:46,610 --> 00:34:47,166 to the office? 910 00:34:47,801 --> 00:34:48,119 Sir? 911 00:34:48,833 --> 00:34:49,310 He'll need? 912 00:34:50,025 --> 00:34:50,595 I mother, 913 00:34:51,309 --> 00:34:51,626 mother, 914 00:34:52,260 --> 00:34:53,712 a here. So 915 00:34:54,084 --> 00:34:56,701 so it's a kind of mixture of, you 916 00:34:56,701 --> 00:34:57,653 know, use bowl. 917 00:34:58,625 --> 00:35:01,184 Parts. So I didn't find an individual word 918 00:35:01,184 --> 00:35:01,904 or phrase, 919 00:35:02,385 --> 00:35:04,224 but then it has, you know, plenty of 920 00:35:04,224 --> 00:35:06,304 surprises and it has plenty of, you know, 921 00:35:07,038 --> 00:35:09,502 a weird waste than to stay things or 922 00:35:09,502 --> 00:35:12,465 like, you know, it stars then than whisper 923 00:35:12,522 --> 00:35:12,760 or... 924 00:35:13,793 --> 00:35:14,293 Away 925 00:35:16,034 --> 00:35:19,217 boy away, get away from her, get away 926 00:35:19,217 --> 00:35:21,048 from her. Go away, 927 00:35:21,525 --> 00:35:24,869 get away from her. So let's say, at 928 00:35:24,869 --> 00:35:27,672 this moment, this blog. So here it. You 929 00:35:27,672 --> 00:35:28,332 have a 930 00:35:28,788 --> 00:35:30,861 here, I have a bit video, so I 931 00:35:30,861 --> 00:35:33,412 have salad, but but there's no... But there's 932 00:35:33,412 --> 00:35:35,898 no sync And then after the result, you 933 00:35:35,898 --> 00:35:38,688 can feed it a sound file and a 934 00:35:38,688 --> 00:35:41,159 cards video file end result in the best 935 00:35:41,159 --> 00:35:41,478 case. 936 00:35:42,768 --> 00:35:44,676 It then gives you a lip which is 937 00:35:44,676 --> 00:35:46,584 trying to like almost, you know, rice. Animal 938 00:35:46,584 --> 00:35:48,254 don't. Animals eat bug. 939 00:35:49,606 --> 00:35:52,568 Lip of, painful because when people are talking, 940 00:35:52,728 --> 00:35:54,563 they're not just them mood in their, you 941 00:35:54,563 --> 00:35:56,078 know, mouth. It's the whole phase. 942 00:35:57,195 --> 00:35:59,907 There will somebody will soon invent an Ai 943 00:35:59,907 --> 00:36:01,989 that will take that end result or that 944 00:36:01,989 --> 00:36:04,372 the face will be more intent with D 945 00:36:04,372 --> 00:36:07,152 singh? Is the Ai just giving you the 946 00:36:07,152 --> 00:36:09,455 shots or is it also giving you an 947 00:36:09,455 --> 00:36:11,776 initial edit? Were you building the whole edit? 948 00:36:12,094 --> 00:36:14,639 I the kind of did myself. So adi 949 00:36:14,798 --> 00:36:16,889 I at this moment. I haven't found an 950 00:36:17,026 --> 00:36:20,621 that do a meaningful video ad d. But 951 00:36:20,621 --> 00:36:23,088 at at this moment, I think that's the 952 00:36:23,088 --> 00:36:25,475 crossover part that your that you need to 953 00:36:25,475 --> 00:36:28,036 build yourself, and it's the fund builders. So 954 00:36:28,036 --> 00:36:30,424 so so that you have, you know, these 955 00:36:30,424 --> 00:36:32,493 all the guy the building blocks, but then 956 00:36:32,493 --> 00:36:34,737 the actual crafting in, your, like, you know, 957 00:36:34,896 --> 00:36:36,113 they feel it's started 958 00:36:36,728 --> 00:36:38,640 exactly the same layer as, you would do 959 00:36:38,640 --> 00:36:38,799 it, 960 00:36:39,596 --> 00:36:42,065 you know, with then with the material, you 961 00:36:42,065 --> 00:36:44,709 wouldn't then shoot it yourself, in a traditional 962 00:36:44,709 --> 00:36:44,948 film, 963 00:36:46,064 --> 00:36:48,296 most people, you know, if you don't have 964 00:36:48,296 --> 00:36:51,085 infant budgets, you do do right screen play, 965 00:36:52,057 --> 00:36:54,132 you apply the shirt, you do the shoes, 966 00:36:54,292 --> 00:36:56,606 you get the results. And then and, you 967 00:36:56,606 --> 00:36:58,442 know, then, you know, basically, then that's it. 968 00:36:59,414 --> 00:37:02,282 Now with this, this letter, I have potential 969 00:37:02,282 --> 00:37:03,158 there to do it then, 970 00:37:04,034 --> 00:37:06,918 infinite ratios, you know, infinite, you know, I'm 971 00:37:07,077 --> 00:37:10,028 I'm not tied to the end results. I'm 972 00:37:10,028 --> 00:37:10,927 completely free 973 00:37:11,623 --> 00:37:13,618 then to create trough. I'm sure this must 974 00:37:13,618 --> 00:37:14,596 happen when you're, 975 00:37:15,293 --> 00:37:15,851 in the edit. 976 00:37:16,662 --> 00:37:18,036 There's a little piece of 977 00:37:19,126 --> 00:37:22,464 connective tissue, some logic that's missing. And you 978 00:37:22,464 --> 00:37:24,610 know you can fill that hole so you 979 00:37:24,610 --> 00:37:26,536 go back and And that's and that's where 980 00:37:26,616 --> 00:37:28,291 I to say as a filmmaker Was really 981 00:37:28,291 --> 00:37:30,285 impressed because I've seen a lot of, you 982 00:37:30,285 --> 00:37:33,476 know, generative art, and it's very pretty. It's 983 00:37:33,476 --> 00:37:35,310 very interesting. It could be very beautiful. 984 00:37:35,804 --> 00:37:37,642 But telling a story can be hard because 985 00:37:37,642 --> 00:37:40,759 of the nature of how Ai generates images. 986 00:37:40,998 --> 00:37:43,875 So making that stuff hang together in a 987 00:37:43,875 --> 00:37:44,195 thread, 988 00:37:45,009 --> 00:37:46,445 is I think something... 989 00:37:48,200 --> 00:37:50,035 You you seem to have AAA really unique 990 00:37:50,035 --> 00:37:52,051 talent for the images 991 00:37:52,683 --> 00:37:54,599 still managed to somehow, 992 00:37:55,716 --> 00:37:56,833 talk hand off, 993 00:37:57,472 --> 00:38:00,106 meaning 1 shot to the next. It progresses, 994 00:38:01,078 --> 00:38:02,992 in a in a meaningful way, and that's... 995 00:38:03,391 --> 00:38:05,486 I I can tell that's a lot of 996 00:38:05,545 --> 00:38:06,045 iterations 997 00:38:07,459 --> 00:38:08,974 and a lot of work. But also, at 998 00:38:08,974 --> 00:38:10,011 a certain point, you have to let it 999 00:38:10,011 --> 00:38:11,776 go. You could do that forever. Yep. 1000 00:38:12,491 --> 00:38:15,034 Yeah. Mean, I'm very happy with, you know, 1001 00:38:15,272 --> 00:38:16,940 I think the 1 of the core part 1002 00:38:16,940 --> 00:38:18,927 of this Ai test is that, you know, 1003 00:38:20,214 --> 00:38:21,112 you cannot 1004 00:38:21,725 --> 00:38:24,531 obtain perfection. I think it's a bit in 1005 00:38:24,668 --> 00:38:28,103 our music you then replace chords. So you 1006 00:38:28,103 --> 00:38:30,495 don't... Are you derive a song with has, 1007 00:38:30,814 --> 00:38:33,525 the, of the basic chords, and you it 1008 00:38:33,525 --> 00:38:33,923 works fine. 1009 00:38:34,641 --> 00:38:36,889 I, but but you don't actually your are 1010 00:38:36,889 --> 00:38:38,963 replaced almost, you know, every chord with, like, 1011 00:38:39,123 --> 00:38:41,858 a weird version of the cord data. And 1012 00:38:42,156 --> 00:38:44,630 you know, air shots are these weird were 1013 00:38:44,630 --> 00:38:44,950 versions. 1014 00:38:45,442 --> 00:38:47,507 But if they full fulfill their role in 1015 00:38:47,507 --> 00:38:49,891 the story. So if it if it works 1016 00:38:49,891 --> 00:38:52,194 in your mind. So if the story goes 1017 00:38:52,194 --> 00:38:54,355 from gear to year or it start. It 1018 00:38:54,355 --> 00:38:56,503 the max you feel or something. It makes 1019 00:38:56,503 --> 00:38:59,048 no difference if the shot itself is actually 1020 00:38:59,048 --> 00:39:00,242 completely kind of like, you know, 1021 00:39:00,892 --> 00:39:03,771 we or like, they're run or Florida or 1022 00:39:04,223 --> 00:39:05,729 the you want that I'm working on or 1023 00:39:05,729 --> 00:39:06,284 right now, 1024 00:39:07,394 --> 00:39:08,559 but I'm working on a 1025 00:39:09,235 --> 00:39:10,450 5 minute video 1026 00:39:11,301 --> 00:39:12,516 about how how 1027 00:39:12,969 --> 00:39:14,240 people invented that 1028 00:39:15,035 --> 00:39:15,750 blue Pepsi. 1029 00:39:16,386 --> 00:39:19,364 And, you know, And then this this video 1030 00:39:19,660 --> 00:39:21,329 it will be probably out in a few 1031 00:39:21,329 --> 00:39:21,488 days. 1032 00:39:22,203 --> 00:39:24,509 Think did will probably be my be my 1033 00:39:24,509 --> 00:39:26,519 best 1 in that way that I 1034 00:39:27,069 --> 00:39:29,942 go far away from, like, no reality, man, 1035 00:39:30,181 --> 00:39:31,697 I actually go into worlds. 1036 00:39:32,256 --> 00:39:32,735 Have a taste. 1037 00:39:33,453 --> 00:39:34,171 Tell me what you think. 1038 00:39:42,259 --> 00:39:43,536 It tastes like mouth wash. 1039 00:39:44,414 --> 00:39:45,632 You are a crazy. 1040 00:39:49,282 --> 00:39:49,920 It's non. 1041 00:39:50,492 --> 00:39:51,998 Mouth wash. It berries. 1042 00:39:53,900 --> 00:39:55,563 I'm sorry. I really don't know how it 1043 00:39:55,563 --> 00:39:56,118 can help you. 1044 00:39:58,195 --> 00:39:58,751 What ties you. 1045 00:40:00,101 --> 00:40:00,418 Why? 1046 00:40:01,610 --> 00:40:03,039 You will find the answer inside. 1047 00:40:06,137 --> 00:40:06,454 Relax. 1048 00:40:07,822 --> 00:40:08,699 That's your mind drift. 1049 00:40:09,577 --> 00:40:11,650 It's simply hugely excited because you can go 1050 00:40:11,650 --> 00:40:13,484 in the worlds that are hard to go 1051 00:40:13,484 --> 00:40:14,223 into in 1052 00:40:14,680 --> 00:40:17,327 traditional world. If you think about the as 1053 00:40:17,327 --> 00:40:17,805 a camera. 1054 00:40:18,682 --> 00:40:20,677 You can't then get a camera that is 1055 00:40:20,677 --> 00:40:22,910 able to do what these Ai cameras are 1056 00:40:22,910 --> 00:40:24,744 enabled then to do it right now. And 1057 00:40:24,904 --> 00:40:26,042 I think that's a personal. 1058 00:40:26,753 --> 00:40:28,109 Way of how do you think about it? 1059 00:40:28,508 --> 00:40:30,103 But if you think about it in terms 1060 00:40:30,103 --> 00:40:32,177 of the this model that you actually have 1061 00:40:32,177 --> 00:40:35,288 a tool that is far than superior to 1062 00:40:35,288 --> 00:40:36,380 the world's best camera 1063 00:40:36,737 --> 00:40:37,932 in some ways. 1064 00:40:38,728 --> 00:40:40,581 I think if you go into that, 1065 00:40:41,196 --> 00:40:43,027 then you can start to see all kinds 1066 00:40:43,027 --> 00:40:43,983 of interesting potential. 1067 00:40:44,794 --> 00:40:46,792 Sir, the camera... The camera you sold me 1068 00:40:46,792 --> 00:40:47,911 has taken Lsd. 1069 00:40:49,429 --> 00:40:49,749 Yeah. 1070 00:40:50,867 --> 00:40:52,865 And I think you know, that's of course 1071 00:40:52,865 --> 00:40:53,525 part of 1072 00:40:53,840 --> 00:40:56,160 fortunately, it's part of new medium. So if 1073 00:40:56,160 --> 00:40:56,559 you put... 1074 00:40:57,840 --> 00:40:59,840 I mean, they kind of the resolution is 1075 00:40:59,840 --> 00:41:01,680 is the lower. And so but if you 1076 00:41:01,680 --> 00:41:03,054 put it on the instead instagram having put 1077 00:41:03,054 --> 00:41:06,643 it on Twitter, you don't need these 4K6K 1078 00:41:06,643 --> 00:41:08,557 cameras that, you know, show show them all 1079 00:41:08,557 --> 00:41:11,450 the details. And I think overall stories 1080 00:41:13,277 --> 00:41:14,630 I mean, even if if you think about 1081 00:41:14,630 --> 00:41:15,425 movies like, 1082 00:41:16,459 --> 00:41:19,164 blair at which projects or than whatever. If 1083 00:41:19,164 --> 00:41:20,850 you have a good story. 1084 00:41:21,408 --> 00:41:23,559 People who will then start to for for 1085 00:41:23,559 --> 00:41:26,268 forgive. There's all kinds of interesting deals with 1086 00:41:26,268 --> 00:41:28,978 these A fields for instance that for in 1087 00:41:28,978 --> 00:41:29,478 in 1088 00:41:30,028 --> 00:41:32,255 the comparisons in each shop that look different 1089 00:41:32,255 --> 00:41:34,721 or they have light, then then they have 1090 00:41:34,721 --> 00:41:36,733 a, of different set of clothes 1091 00:41:37,441 --> 00:41:38,873 and you are team at first that, you 1092 00:41:38,873 --> 00:41:40,147 know, that's slightly distracted. 1093 00:41:40,863 --> 00:41:43,967 But I really actually find that people don't 1094 00:41:43,967 --> 00:41:45,757 call. They don't comp 1095 00:41:46,134 --> 00:41:48,287 complain that in this shot, they had a 1096 00:41:48,287 --> 00:41:50,281 blow that in the next shot, they have 1097 00:41:50,281 --> 00:41:52,833 a white shirt. It's not that... It's not 1098 00:41:52,833 --> 00:41:54,841 something you see in the North, funeral know 1099 00:41:54,841 --> 00:41:56,675 to nobody would have the time to actually 1100 00:41:56,675 --> 00:41:59,866 change their clothes for a shot. But the 1101 00:41:59,866 --> 00:42:02,737 people are actually surprisingly or given that, you 1102 00:42:02,737 --> 00:42:02,976 know, 1103 00:42:04,267 --> 00:42:06,020 As slur as the holes are the same. 1104 00:42:06,896 --> 00:42:09,786 What Europa is talking about here is called 1105 00:42:09,844 --> 00:42:10,163 coding. 1106 00:42:10,721 --> 00:42:12,553 If you give the audience the code or 1107 00:42:12,553 --> 00:42:14,635 the grammar your film in the first few 1108 00:42:14,635 --> 00:42:17,095 minutes, they will most likely be able to 1109 00:42:17,095 --> 00:42:19,896 adjust and accept almost anything that comes afterward 1110 00:42:19,952 --> 00:42:21,618 as long as it sticks to that set 1111 00:42:21,618 --> 00:42:21,936 of rules. 1112 00:42:22,744 --> 00:42:23,460 Defined by the code. 1113 00:42:24,809 --> 00:42:27,987 It's just another example of how adaptive and 1114 00:42:27,987 --> 00:42:29,257 plastic our brains are. 1115 00:42:30,544 --> 00:42:32,880 And back to that issue of temporal 1116 00:42:33,338 --> 00:42:33,657 instability. 1117 00:42:34,535 --> 00:42:37,249 It lends itself to a dream like quality 1118 00:42:37,249 --> 00:42:39,405 where men is never solid. 1119 00:42:40,296 --> 00:42:41,332 Ideas are not fixed. 1120 00:42:42,447 --> 00:42:45,157 But the technical issue of temporal stability will 1121 00:42:45,157 --> 00:42:45,714 be solved, 1122 00:42:46,351 --> 00:42:48,822 but it is this period of development that 1123 00:42:48,822 --> 00:42:49,322 most. 1124 00:42:50,347 --> 00:42:53,309 It's not bad generative of art, it's simply 1125 00:42:53,603 --> 00:42:55,986 the current state of the art and video 1126 00:42:55,986 --> 00:42:58,782 generated during these years will be marked by 1127 00:42:58,782 --> 00:43:01,348 that sense of cutting edge possibility. 1128 00:43:02,039 --> 00:43:03,072 And then maybe 1 day, 1129 00:43:03,788 --> 00:43:06,188 we'll become nostalgic. In the same way that 1130 00:43:06,188 --> 00:43:08,656 we are now nostalgic for that first 3 1131 00:43:08,656 --> 00:43:10,805 d video for Dire straits money for nothing. 1132 00:43:11,362 --> 00:43:13,909 It looks silly now, but it marks a 1133 00:43:13,909 --> 00:43:16,309 time in that technology technology's development. 1134 00:43:19,010 --> 00:43:20,837 If you think about a dream sequence in 1135 00:43:20,837 --> 00:43:22,744 the field on me, that's like, they're super 1136 00:43:22,744 --> 00:43:24,356 hard to do actually. You know, people 1137 00:43:24,903 --> 00:43:26,810 do like a blur and they kind of, 1138 00:43:26,890 --> 00:43:29,273 like, put something glow there and they pretend, 1139 00:43:29,353 --> 00:43:31,339 you know, that slide of dream sequence. And 1140 00:43:31,339 --> 00:43:31,839 and 1141 00:43:32,293 --> 00:43:33,428 and they always suck 1142 00:43:33,738 --> 00:43:35,749 how do our minds still work in dreams 1143 00:43:35,885 --> 00:43:38,033 overall. Mean, it's hard to fold some something 1144 00:43:38,033 --> 00:43:40,442 and not anything is this eastern instant. 1145 00:43:40,817 --> 00:43:41,635 This temporal 1146 00:43:42,342 --> 00:43:44,402 stability is trying at at this moment in 1147 00:43:44,402 --> 00:43:47,255 20 23. To capture this moment in time 1148 00:43:47,255 --> 00:43:49,237 that is trying apply buried then fleeting. 1149 00:43:49,809 --> 00:43:51,170 In 5 years, you know, 1150 00:43:52,210 --> 00:43:53,890 the teams will be worked out. So I 1151 00:43:53,890 --> 00:43:55,250 kind of like, you know, it it it 1152 00:43:55,250 --> 00:43:58,210 will start to look look more than normal. 1153 00:43:58,543 --> 00:43:59,976 But I think we are now in a 1154 00:43:59,976 --> 00:44:02,205 in in a super specific moment of time 1155 00:44:02,205 --> 00:44:02,705 where 1156 00:44:03,240 --> 00:44:05,548 it's now at the threshold that is kind 1157 00:44:05,548 --> 00:44:07,954 of like, in... It's extreme main interesting interesting 1158 00:44:07,954 --> 00:44:10,346 in place. Like first music videos with 3 1159 00:44:10,346 --> 00:44:11,244 d, so 1160 00:44:11,701 --> 00:44:14,650 pioneer straight, you know, music video, just like. 1161 00:44:15,224 --> 00:44:17,382 Bad really. But at the same time, it's 1162 00:44:17,382 --> 00:44:19,619 not bad. It's actually fantastic and it and 1163 00:44:19,619 --> 00:44:23,076 it's really didn't capture something specific. You actually 1164 00:44:23,868 --> 00:44:25,800 create something with these tools that are 1165 00:44:26,334 --> 00:44:26,811 flawed. 1166 00:44:27,368 --> 00:44:29,516 They tried to create art with it in 1167 00:44:29,516 --> 00:44:31,843 not in a kind of are a sarcastic 1168 00:44:31,843 --> 00:44:34,000 vendor, but actually try to create something with. 1169 00:44:34,400 --> 00:44:36,897 I have this assumption that if I take 1170 00:44:37,356 --> 00:44:39,289 5 years to the if we go 5 1171 00:44:39,289 --> 00:44:41,127 years to the future, and then we look 1172 00:44:41,127 --> 00:44:41,627 back 1173 00:44:42,005 --> 00:44:44,723 that the art being created in 20 23, 1174 00:44:44,963 --> 00:44:46,001 will be actually quite 1175 00:44:46,721 --> 00:44:47,360 interested hopefully. 1176 00:44:48,014 --> 00:44:49,929 Because it will be kind of like the 1177 00:44:49,929 --> 00:44:52,324 first or glimpse sets of kind of light, 1178 00:44:53,043 --> 00:44:54,820 from, like this humbly noise 1179 00:44:55,277 --> 00:44:58,184 to actually something that is the meaningful and 1180 00:44:58,401 --> 00:45:00,413 and I suspect and hope that 1181 00:45:01,027 --> 00:45:03,177 this short moment in time that we are 1182 00:45:03,177 --> 00:45:04,314 in right now 1183 00:45:04,784 --> 00:45:06,369 you don't need to wait. So, I mean, 1184 00:45:06,449 --> 00:45:09,066 you can tell, you know, exciting stories today. 1185 00:45:13,926 --> 00:45:15,841 Thank you for watching part 1 of our 1186 00:45:15,841 --> 00:45:16,639 interview with a rope on. 1187 00:45:17,596 --> 00:45:19,772 Next time, we're gonna get more into the 1188 00:45:20,164 --> 00:45:23,186 ethical and financial matters associated with Ai. 1189 00:45:24,380 --> 00:45:25,016 Thanks for watching. 1190 00:45:27,818 --> 00:45:30,053 The well is written produced and edited by 1191 00:45:30,213 --> 00:45:31,251 Ans Mount and B, 1192 00:45:32,289 --> 00:45:35,163 theme music written and performed by Jonathan Me, 1193 00:45:35,562 --> 00:45:37,567 with additional music by Brandon agents. 1194 00:45:38,837 --> 00:45:40,980 Special thanks to Rope Rid stone for chatting 1195 00:45:40,980 --> 00:45:43,043 with us from Finland and for donating the 1196 00:45:43,043 --> 00:45:45,050 work that appears in this episode 1197 00:45:45,759 --> 00:45:46,478 until next time. 1198 00:45:47,118 --> 00:45:47,917 Have a good time.