3/28/2025 - Exploring ChatGPT's Image Generation & Narrative's Future

Transcript

1. 0:01friend. Go down to that horn place again. Do Do you know the sounds of closing doors? Have you heard that soundwhere
2. 0:19before? Do you want do they wonder if they know you [Music] any more more [Music]
3. 0:39[Music]
4. 0:55[Applause] [Music] Friday night date night. Good people. What's [Music]
5. 1:12happening? I'm on a little bit early. I was bored. I came in to get ready early and then I got ready early and then I got
6. 1:21bored. I got bored, people. I got ADD. I got the shivers. [Music]
7. 1:41Oh [Music]
8. 2:07yeah. Oh yeah. [Music]
9. 2:27Every time I see you now, get that look in mine. Every time I see your mouth, I hear that
10. 2:41smile. In the early misty morning light, I heard the engine turning in the old for [Music]
11. 2:58outside. You were leaving me again today. [Music] You will convince me again today. You're leaving this hard time
12. 3:15looking for someone else's golden ring. to say. So long to [Music]
13. 3:34say now don't you cry. So long, Susanna. Don't you cry for
14. 3:50[Music] me. Oh yeah, champ. That is some fine
15. 4:04fine fine singing. Like an aged wine or a fat dog. You're doing great. You're doing great. You sound good. Melodious. You got melodious
16. 4:17tones. Mel maleiferous. Melodious. I don't know. I don't know words. Some people have a way with words. Other
17. 4:26people not have way. I guess you'd say. All right. Something's different. Can you feel it? Did you
18. 4:41notice? Did you notice something has [Music] changed? There's more energy. People are excited again. [Music]
19. 5:24Let us be lovers. Well, marry our fortunes together. I've got some real estate here in my bag. So we bought a pack of
20. 5:38cigarettes and Mrs. Wagner's pass and we've all come to look for America. [Music]
21. 5:58[Applause] [Music] [Applause] [Music]
22. 6:14It's like I have precision with fingerpicking minus the precision. [Laughter]
23. 6:29Happy Friday night everybody. I hope you're [Music] well. We're going to continue playing with uh Chat GBT. Tonight
24. 6:45[Music]
25. 7:13Oh, Champ, you're such a good singing dog. [Music]
26. 7:31[Music] Woohoo! Woo! [Music]
27. 8:05Um, what's happening? What's happening? What's happening? What's happening? What's happening? I had a bunch of good meetings today. We
28. 8:17had a really nice AI office hours. Saw some old friends. That was really nice. It was small. It was a smaller group
29. 8:23today, which was kind of cool. It was It was a small intense group. Very well talked. It was a very well It was a very very smart group of
30. 8:34people that showed up today. And then I was there too. [Music] [Applause] Um, but yeah, we were talking about
31. 8:49this. Uh, something is shifting. I think we've crossed a threshold. I think we've crossed a Rubicon. I don't know what a Rubicon is,
32. 8:59but I think we've crossed it because I know you cross them in movies. Thank you, Spin B3. I missed it again. Oh, dang it. chemical
33. 9:10farmer. I always forget until like 2 PM. Yeah, it's at 11:00 a.m. Mountain, so it's a little early if you're on the
34. 9:16West Coast. Purple Rain. That was Purple [Music] Rain. Wait.
35. 9:31[Music]
36. 9:45Only want one time to see you laughing. Only want to see you laughing in that purple rain. Purple rain. Purple
37. 10:01rain. Purple rain. Purple rain. Purple rain. Purple [Music]
38. 10:15rain. I only want to see you laughing in that purple. Yeah. And then from that point forward, Martin Ston like sings really good.
39. 10:29That's where I stop. You got to know your limitations, people. Just go right up to the edge of them and then sneak over occasionally
40. 10:40and just go, "Yeah, that's not [Laughter] me." Although, you know, with this AI [ __ ] I think we should all be a bit
41. 10:53more bold in what we think we can do. [Music]
42. 11:12Um, let's let's go look at some stuff. Let's go look at some stuff. What's going on? How is everybody? I got no comments going on on
43. 11:23YouTube. What the hell's going on on YouTube? No thumbs up, no likes, no comments. the hell you people doing over
44. 11:29there? Probably just a bunch of people on X going, "Who? Why is there an old man on X? What's happening? What's
45. 11:38happening? I thought all the old men were on TikTok. I don't understand it. What's happening?" Um, sounds good, but Champy
46. 11:47always helps. He He howls good. Yeah, exactly. I mean, that's the thing. When you have when you have a lead singer as a
47. 11:55dog, you can fill in the gaps. Just let him take the lead. Uh, there are comments on YouTube. Really? I ain't seeing
48. 12:12them, huh? Oh, there they are. I had a I had a screen redraw issue. Well, let me see what we got here. I watched the Bob Dylan tonight. I
49. 12:27prefer watching Kyle. Thank you. Well, you know, it's so funny you say that, Vicki. A lot of times people are like,
50. 12:35you know, do you like Timothy Shalomé? And and people like, he's no Kyle Shannon. Like, I hear that all the time.
51. 12:44They're like, yeah, it's not the AI learning lab. I mean, sure, he took five years of his life and transformed
52. 12:49himself into Dylan, but you know, is is there a singing dog? I don't think so. Yeah, I hear it all the time. Shalom or
53. 12:58Shannon? Which way you going? And I listen, I can't help it if people lean my direction. Right. It's just it's a
54. 13:07magnetism I've got. It's a natural genetic gift that I have been sprinkled on by the gods. So anyway, thank you for the for the kind words.
55. 13:20Uh ah the chat servers must be working overtime. Oh, it took several tries until I got my message. Another thing
56. 13:29that could be happening, Sharon, um one of the things that they have started to do is they have started to censor the images.
57. 13:38What's fascinating is that they censor them about threequarters of the way through building an image and then they
58. 13:44realize, oh, this isn't going good. Let's just delete it. So, that's a possibility, too. I'm not the biggest Adam Lavine
59. 13:56fan. If you hear his purple rain live from the Stern party. Oh, cool. Nice. We're here. Okay, good. Everybody's
60. 14:03here. I couldn't see anyone. Chef's kiss. Very nice. Um, I did not enjoy the music. I agree. AI learning lab much better matter.
61. 14:10Shannon all the way. You guys are [ __ ] weird. We're not weird. We're irregular. I thought we established that yesterday. Good lord.
62. 14:25Okay, let's get going. Um, yes, that happened to me today. My chat art. I was very disappointed. Yeah. Where where it
63. 14:33deleted it after it generated it. Yeah, they're they're doing that now. They have an algorithm in there
64. 14:39that says, "Did this make an image that we might get sued by a big company for?" And if the answer is yes, they delete it. It
65. 14:52sucks. AI Learning Lab is the favorite hangout in Shellington. If you weren't here last night, we invented a
66. 14:58city. We've got the the Irregular Turtle Society. If you're not part of that, you weren't here last night. And I
67. 15:08got no sympathy for you. Here's the deal. You people, you people, you you're like, "Yeah, I want to learn AI." And
68. 15:16and then you're like, "Well, I'll come a night a week or once or twice a month." You miss out. And I mean, listen, I think there's
69. 15:27a legitimate argument to be said that a solid 90% of the hours of of entertainment that I procure here are
70. 15:38useless, wasted time, you'll never get back in your life. But it's that 10%, it's the [ __ ] magic. It happens. It's like a baseball game.
71. 15:50Occasionally, something happens, you know, then you drink beer in between. Oh man. Um, Chef Kelly, I was embarrassed
72. 16:04after my botched virtual webinar today and then I remembered you do this every day. Yeah. Listen, botched. I assume it
73. 16:12was not massively botched. Maybe it didn't quite go as planned. But this is, think of this as inspiration for Well, it couldn't be that bad.
74. 16:25Now listen, to be fair, you don't have a $20 million production studio like I do, right? Like it's not going to look this
75. 16:35good, right? But just, you know, and sure, this is fully scripted, right? I spend most of my days learning
76. 16:45the lines for the performance of the night. This show is very much like the uh that Netflix series Adolescence, right? where I memorized
77. 16:53the script. It's a single camera take, you know, dramatic action happens. We produce this in such a way that it looks like I'm just
78. 17:09insane. [Music] See? All right, let's go look at some [ __ ] source camp in
79. 17:26house. Did I not share? What's going on here? Okay, there is just weird ass stuff going on with my StreamYard.
80. 17:37Um, okay. So, here I am. I'm at I'm at the chat GPT. I don't know if you all know chat GPT. If you go to chatgpt.com or even just chat.com,
81. 17:49uh you might have heard you can make money with chat. This is the place you go. You just go there and and when you
82. 17:57type in a prompt, you can make money from that. And you're like, "Well, how do I make money from that?" Well, you pay me the
83. 18:06$49.95 a week for the secret lesson. That's how you make money from it. Well, no. That's how I make money from it. It
84. 18:18And then but then what you do is you take what I did and then you do that. But it's not easy. Like you have
85. 18:27to be able to go you can make money with chat. You you can't just say you can make money on chat GBT. It's like it doesn't it just
86. 18:38doesn't work unless you're like at right. And that's what I'll teach you. I'll teach you the secrets to
87. 18:46artificial intelligence are making a Tik Tok channel to sell methylene blue and and Madagascar vanilla beans and
88. 18:54cardboard cutters. That's how you make money with chat GPT. The Tik Tok shop. You see it? Like is it all
89. 19:03starting to come together? Because it should it should be really clear. No. I I I I don't know what to say. Like if you're thick in the
90. 19:16head, it's that's not on me. You're like, "But Kyle, you told us we didn't have to be technical." You don't have to
91. 19:23be technical, but you have to learn how to say words. Words like cha, you see? Anyway, um, chat
92. 19:39GPT. Oh, good lord. All the regulars are like, "All right, I got 15 minutes here. I got to run to the [Laughter]
93. 19:53bathroom." Ah, yes. Words. I I am saying words, Tom. I words. Words, words, words, words, words. Um, I did start watching Sing
94. 20:09Sing tonight, the the uh the movie about the uh the prison theater troop. It's pretty good. It's pretty good. I'm like halfway through it.
95. 20:18Um, I wanted to like one of the things I wanted to explore tonight is you know how when you're using chat
96. 20:29GPT one of the secrets of it is is how you say it and the other secret is it's not so much that you get good at
97. 20:38prompting but it's more like you get good at conversation with chat GPT right you start to understand its personality
98. 20:46and you start to understand how you and give it feedback and if you do the conversations right you can maintain a thread of understanding
99. 20:55through a long conversation right and that's like I feel like when you get good with chat GPT it's kind of like
100. 21:02this ongoing conversation that you just keep improving it and improving it improving it and where it gets
101. 21:08frustrating is you improve it improve it and then it falls apart right and so now that image generation
102. 21:16is native within in the platform. I was like, can you maintain that same kind of um context and continuity across images?
103. 21:27And so, so I just put this prompt up. I said, "Make me a blueprint of a radical design for a new kind of house." And I
104. 21:36was expecting it to draw a picture. And and what it did, which I thought was kind of interesting, is it described the
105. 21:46house. So what did it say? A living, learning, breathing structure that adapts the spiral nest that adapts to
106. 21:53its occupants and environment. It's part bioarchchitecture, part digital habitat, part sensory cocoon. A house that feels
107. 22:03you thinking and responds in kind. Um and then they describe the different areas. Its shape is a spiral shell
108. 22:11inspired by the Fibonacci sequence. The level one is called the root zone. Level two is called the mind deck. Level three
109. 22:19is called the green loop. Level four is the dream nest bonus the morph room. So that was kind of cool. And then I said, well, you
110. 22:30know, make me a blueprint of the floor plan. But I said floor plan, but whatever. And then it wrote code and it wrote like a Fibonacci
111. 22:41spiral and then it and then it did like Python drawing which is not at all what I wanted but it's kind of interesting and it it's kind of an
112. 22:55interesting floor. It's like it's interesting that it that it it thought about the math of the house, right? So
113. 23:03then I said, "No, no, make me a like a blueprint as if an architect drew it." And it did it. And it essentially
114. 23:15maintained the same thing, right? It's got the spirals. It's got a It's got an area called the digestion zone. I don't know
115. 23:23if that's a bathroom or if that's if if you just walk in circles here after you eat. And then I said, "Now, let me see a
116. 23:34watercolor visualization of it." And it did that. It's pretty [ __ ] good. Like it's in the neighborhood of what an architect would
117. 23:45do. It's got a floating bridge and the entrance and the entry garden. Now, it's not complete. Like, this was all very
118. 23:54fast and it's got plants inside the house and there there's no actual kitchen where you can cook [ __ ] There's
119. 24:00just a table, but still pretty cool. Then I said, um, now make it a photograph as if it were taken for Architectural
120. 24:11Digest. And then there's there's the thing. And it maintains the foot bridge. It maintains the pond. It maintains the
121. 24:19table, the central spiral staircase, the spiral shape of the house. And then one of the things I heard is you can ask it to give you the
122. 24:32prompt of an image. And so I said, "Hey, can you give me the prompt of the image?" And what it wrote was, "This
123. 24:38this stunning highresolution photo captures a con contemporary spiral-shaped home with curved glass walls set against a vibrant green
124. 24:48landscape. Sleek concrete roof spirals upward, and the floating bridge leads to the entrance." But it like didn't feel
125. 24:54like it had enough detail to generate all that. And so then I asked it. I said, um, it seems like there would be more to
126. 25:07the prompt than that. Did it incorporate incorporate the previous watercolor image to get the shape of the house? And
127. 25:15it said, "Great question. You're totally on point." The photo style image did draw on the previous watercolor image
128. 25:24indirectly by carrying forward the visual memory of the design. Now, here's what I don't know. This is chat GPT. So
129. 25:32it's possible that it absolutely hallucinated this. But if it's true, what this is saying is it's it understands this code. It understands
130. 25:46this drawing. It understands this drawing. It took that drawing to make that image. It took that image to make that
131. 25:53image. And then it said, "Do you want to see the interior?" And I'm like, "Sure, show me the interior." here. And there's
132. 25:58like there's the interior, I guess, out the back of the house. I mean, that's pretty [ __ ] cool, right? It's a tur It has a turtle ramp.
133. 26:09It totally has a turtle ramp. But, you know, I thought that was pretty [ __ ] swell.
134. 26:24I saw Kelly Camp did a uh a magazine piece today here. Let me see if I can go find [Music]
135. 26:39it. So good. There's Kelly's little magazine. The new MVP, The Quiet Revolution.
136. 26:53Beyond the buzzwords or buzz whirls as chat GBT made it. AI isn't coming. It's here. That's good. The human interface to AI architect. That
137. 27:05cool. So if you don't know how she made that, you can just go to chat GPT right now. Uh you can do three of these on the free
138. 27:15plan. I don't know if they've rolled it out to the free plan yet, but for 20 bucks a month, you can upload a picture
139. 27:21of yourself and just tell it to put you on the cover of a magazine and then describe the magazine and let it write
140. 27:26the headlines or give it the headlines and it does everything. It does the
141. 27:41typography. So, that's pretty [ __ ] swell. So, we'll we'll come back to chat GBT because I I want to keep I want to keep digging
142. 27:51into what's now possible. I also want to see if we can discover more about the upgrade of 4.0 and why it's good and how it's good and
143. 28:02what's different about it. Um, if anyone has any thoughts, feel free to put them in the comments. Careful with thought. You'll invoke the
144. 28:10wrath of Pate. I'm waiting for the day when it says, "Great point, but you're totally wrong, human." I, you know, I'm actually
145. 28:23waiting for that day, too, because I appreciate the fact that they made Chat GPT a helpful assistant, but sometimes it would be
146. 28:36nice to have it be a helpful critic, right? It'd be nice if it could be intelligent enough to navigate those two roles. What's this
147. 28:46say? It carries things forward. Just try creating a normal person and then creating a Gibbli person. That's true.
148. 28:53Yeah, it's doing character consistency that way. Um what I wanted to um watch together. I only watched a little bit of
149. 29:01this. I'm going to I'm going to switch my my tabs for a second here. Um, Anthropic put out a big paper today,
150. 29:12um, and a blog post about how large language models work, how they think. Um, and I saw a couple of people
151. 29:22mention it as being a really profound piece with a lot of jaw jaw-dropping moments. Um, and I haven't I haven't
152. 29:29read it, so like I just saw it late in the afternoon, but it's got this video with it that that explains some of how
153. 29:36large language models work and how they think. Um, and I thought it'd be fun for us to watch it together. So, we'll do a
154. 29:43little movie night tonight. So, get your popcorn. Go make your Jiffy Pop. Well, Jiffy Pop, boy, that's that's old and
155. 29:52timey. What What What is the popcorn? Is it all Is it Newman's own or is that still old? What's the What's the popcorn
156. 30:00you're making in the microwave these days? Or did we give up on microwave popcorn and we're back to the stove? Are we all bespoke artisal
157. 30:10Oregon-based handpicked every kernel massaged before it's placed in the bag? It Where are Where are we with popcorn? Or do you
158. 30:20just buy the bags of [ __ ] now? Done. pre-done. Orville Red and Bucker. Okay. Waffles. Waffles. Waffles are good, too.
159. 30:34I say we go waffles or [Laughter] popcorn. Oh, man. All right. Let me change my let me change my tab
160. 30:49sharing and then we will watch a we'll do a little movie [Music]
161. 31:07watching. You often hear that AI is like a black box. Words go in and words come out, but we don't know why it's said
162. 31:16what it said. That's because AIs aren't programmed, but trained. And during training, they learn their own
163. 31:24strategies to solve problems. If we want AIs to be as useful, reliable, and secure as possible, we want to open up
164. 31:32the black box and understand why they do things. But even opening the black box isn't very helpful because we don't know
165. 31:41how to interpret what we see. Think of it like a neuroscientist investigating the brain. We need tools to work out
166. 31:48what's going on inside. We want to know how the model connects all the concepts in its mind and uses them to answer our
167. 31:55questions. Now, we've developed ways to observe some of an AI model's internal thought processes. We can actually see
168. 32:04how these concepts are connected to form logical circuits. Let's take a simple example where we ask Claude to write the second
169. 32:12line of a poem. The poem starts, he saw a carrot and had to grab it. In our study, we found that Claude is planning
170. 32:21a rhyme even before writing the beginning of the line. Claude sees a carrot and grab it and thinks of rabbit
171. 32:30as a word that would make sense with carrot and rhyme with rabbit. Then it writes the rest of the line. His hunger
172. 32:36was like a starving rabbit. That's kind of fascinating. I mean, it's like the in in that first line, the valuable things
173. 32:45are carrot and grab it. He saw and had to are not as important. So, I picked the more important stuff. Kind of
174. 32:54cool. It's also when you write poems, this actually makes sense. Now, it's very hard to get it to write an
175. 33:03abab rhyming couplet. It always goes a bb. So, it does them two lines at a time. Be interesting to experiment and
176. 33:11see if you could prompt it out of behaving like this. We look at the place that the model was thinking about the word rabbit and we
177. 33:21see other ideas it had for places to take the poem. We also see the word habit is present there. Our new methods
178. 33:30allow us to go in and intervene on this circuit. In this case, we dampen down rabbit as the model is planning the
179. 33:37second line of the poem and then ask Claude to complete the line again. His hunger was a powerful habit. We see that
180. 33:45the model is capable of taking the beginning of a new poem and thinking of different ways it could complete it and
181. 33:51then writing it towards those completions. The fact we can cause these changes to occur well before the final
182. 33:58line is written is strong evidence that the model is planning ahead of time. This poetry planning result along with
183. 34:04the many other examples in our paper only makes sense in a world where the models are really thinking in their own way about what they say.
184. 34:13Just as neuroscience helps us treat diseases and make people healthier, our longerterm plan is to use this deeper
185. 34:20understanding of AI to help make the model safer and more reliable. If we can learn to read the model's mind, we can
186. 34:28be much more confident it is doing what we intended. You can find many more examples of Claude's internal thoughts in our new paper at
187. 34:36anthropic.com/ressearch. Um, yeah, that's fascinating. I mean, this kind of blows the doors off of, you know, it's just predicting the
188. 34:49next token. It's just a stochastic parrot, right? You know, if it's thinking ahead and like I don't know how
189. 34:58long it's thinking ahead, but that's pretty interesting. And and don't forget anthropic, their core imperative. So,
190. 35:08OpenAI's core imperative is to create AGI, AGI, artificial general intelligence, right? They want to make
191. 35:15it accessible to the world. Enthropic was a spin-off of Open AI from some of the main scientists from Open AI that
192. 35:24said, "We don't like the way you're thinking about safety of these models. We want to create an enthropic where our
193. 35:29prime imperative is safety, AI safety, and and AI alignment." Um, and so that's that's why this research makes really
194. 35:39good sense that they're doing it. Um, but I find it fascinating that basically what they're discovering is these things
195. 35:47actually think. Now, I don't think you can go as far as, you know, they're conscious, but, you know, this is this
196. 35:56is a different thing. And I'm sure if Pate's in here, he's probably screaming at the top of his lungs right
197. 36:01now. Um, let's see. Where's the paper here? [Music] Um, probably at the bottom. Is there any Yeah, there's lots of charts and graphs
198. 36:12that I I that won't make good won't make good for tick tocking. Um, but the the blog post is pretty comprehensive and then I I assume
199. 36:23the paper's really good. Let's see. revealing the computational graphs in language models. Many more details of the above
200. 36:38use cases in our second paper on the biology of large language model. Wow, that's a pretty bold that's pretty bold statement.
201. 36:50We investigate the internal mechanism used by claude 3.5 haiku enthropics lightweight production model in various
202. 36:58contexts using our circuit tracing technology. That's amazing, huh? So anyway, go check that out. It's at uh it's at uh
203. 37:13anthropic.comressearch. I would imagine you can probably just get there by that. Anthropic.comressearch. Hang on.
204. 37:22Yeah. Uh maybe not. Uh let's see. Our mission research principles tracing the thoughts. So here it is. So it's you
205. 37:38scroll down. Go to enthropic.comressearch and scroll down till you see the rustcoled box. Tracing the thoughts of a large language model.
206. 37:47That'll take you back to that where I just was. Crazy. It's crazy, I tell you. Crazy. Crazy. Crazy. All right. Um,
207. 37:58let's see. Did I do anything in in chat GBT today? That was interesting. Oh, that was an interesting one.
208. 38:13Yeah. Um, this was kind of fun. So, I took a screen sh Oh, wait. I got to I got to change my sharing. Hang [Music]
209. 38:32on. So, today I was on a uh conference call. Can you all see that? Hey, Andy, what's happening? Um, today I was on a
210. 38:44conference call. I took a screenshot of some of the crazy people there and then uploaded those pictures and said, "Make
211. 38:51me a cartoon." And one of the cool things that it did was it took the names off the screenshots and turned it into
212. 39:00like the names of the people. So there's Cindy [ __ ] there's Rick Lacroy, who's not really that angry. There's me with a
213. 39:07guitar. It picked the guitar up out of my background, which is interesting. Um, then I tried to add someone to it
214. 39:22and it sort of screwed that up. And then I did something and it sort of screwed that up. But then I I said, you
215. 39:30know, here's more people. Have the people with their cameras on, make fun of the people with their cameras off. So
216. 39:37that sort of did it well. I said, make it like a Pixar thing. So, it kind of did a funny little cartoon. Let's do image gen projects.
217. 39:46Yeah, we're going to sitting here with my computer machine making stuff. Yeah. So, so yeah, let's go play with this.
218. 39:53So, so upload a picture of yourself or your friends. Oh, then I said this was kind of cool. I said take Eric who I put
219. 40:02here based on a screenshot here. Let's see. Where's Eric? Right here in the top. Let me see. Can I zoom in on
220. 40:11that? There's Eric. So that's that's the only picture it had of Eric, right? And then from that it
221. 40:26made this version of Eric right here. And then I said,
222. 40:44um, create a dramatic art photo of Eric where the photographer used lighting and a low millimeter lens and open aperture
223. 40:52to capture a personality that is thoughtful and generous. Make it a 2x3 portrait. And so it made this photograph. It sort of invented this
224. 41:03guy. And then I said, make it black and white. And it did. And then I said, "Make it grittier." Made it really gritty. So, it's pretty
225. 41:16crazy. Um, what else do we do today? The studio Gibbly thing. Oh, yeah. This was pretty cool. So, this was
226. 41:28I wanted to do I wanted to do take my my stupid uh profile photo and turn that into a character. It made me too fat. So, I
227. 41:39said, "Make me skinnier because I'm vain." And then I said, "Make me a Macintosh computer, the original
228. 41:47Macintosh, and convert my portrait into Mac Paint, the original version of Mac Paint." And it did it quite well. So, that's a good
229. 41:56one. Um, I did some I did a New Yorker cartoon today. I had it write New Yorker cartoons. It came up with 10 different cartoon
230. 42:06concepts and then it created this this one that is a bunch of goats sitting around a boardroom and the the quote is,
231. 42:15"Let's circle back after we've eaten the agenda." That's pretty good. That's pretty good New Yorker humor. Come on. [Laughter]
232. 42:28Oh man. Um, one thing that I did notice, let's go to Sin City Sipper. I don't think this has changed, but let's say
233. 42:37make me a margarita with unexpected sass. And now it's making an image, but it is
234. 42:51not using the new image gen tool. So your um custom GPTs have shitty old dolly in them. You know, it's still pretty good. Not for nothing,
235. 43:07but but it's not the new image model. So that I learned. Um, last night we were playing
236. 43:22with turtle meme character ideas. Was this it? Yeah. So, last night we created the the Irregular Turtle
237. 43:39Society and we were taking pictures of people and making versions of them. And then people went to the AI salon,
238. 43:47started making their own pictures of themselves. So, there's Silver Fox with a chrome shell, which I love. I think
239. 43:54it's genius. S Fox, the founder of Shellington, Silverf Fox, congratulation. I didn't realize you were the founder of the town. That's
240. 44:07funny. Oh, you got that same drink pick about three months ago. Oh, that's funny. Um, here's Sharon Crawford. Very nice. So, everyone's
241. 44:17making their own little Oh, that's nice. Who's that from? Steo. I can't make it tonight lunch date with this one. Oh,
242. 44:24that's the one he uses to tap all the hearts on the Tik Tok. Very nice. Marlene Paul, love it. Oh, that one's cool from W. Very
243. 44:38formal. Ocean. Nice. Corin Mueller. Do we all represent ourselves with turtles? Apparently, you are in Shellington, but
244. 44:52in Shellington. Now, Dr. Jay decided to go with the the Blue Jay, and I think Cindy [ __ ] would obviously go with the
245. 45:00raccoon. Um, there's Vicki. Vicki went ahead and made a collectible that you can you can buy at Toys R Us.
246. 45:15Oh my god, there's Greg. Oh, good lord. Oh, this is a good one from Jackie. Still working. Still work shopping my
247. 45:31Shellington peron persona, but I had a bit of issue in the [Laughter] lab. So anyway, so that's kind of fun.
248. 45:43Um, I think what might be kind of cool tonight, we'll we'll come back to the Shellington stuff, but I wanna I want to try something new.
249. 45:57Um, let's turn on web search and let's say um research um
250. 46:14papers and books on modern narrative theory. and what resonates with today's
251. 46:33um audiences. So, I'm going to have it go do some research. Now, you could do this with some deep research, but one of the
252. 46:42groups that I'm working with were playing around with deep research, and Cindy [ __ ] kind of inspired this
253. 46:48today where Cindy went and did a bunch of research and then she's she's digging deeper and and doing research from different points
254. 46:57of view and using personas and things like that. And one of the things that we talked about was um could you take research and bring it
255. 47:09to life uh using the image generation tool? We might be able to do that. So there's some books and things. So now I'm going to say um what
256. 47:23are the main narrative ideas across these
257. 47:37works. Searching the web. Oh, she's here. Cindy [ __ ] is here. So, I figured I'm going to try to do some some crazy ass image generation
258. 47:50like like like bringing research to life, but I I don't want to do deep research. I'm just going to do like, you
259. 47:56know, sort of faker deep research. All right. So, there's that. So, I'm going to say um now look at these ideas from four [Music]
260. 48:10different points of view. um each from a different decade. I'm just I'm just kind of randomly trying to put enough context in
261. 48:28here that we've got something to work with. So, I don't really care what the answer is here. If you were doing actual research, you
262. 48:39might would you like me to explore how these perspectives might apply to a particular format? Um, I want strong POV pieces from each of
263. 48:58them about the research you have found. All
264. 49:13right. Uh, now I want a new
265. 49:31novel hypothesis that incorporates everything we've talked
266. 49:54about that could reolutionize narrative in the future. All
267. 50:09right. Absolutely. Here's a novel hypothesis. the narrative as emerging emotional architecture and multi- aent
268. 50:16space. You know, Cindy, it's funny. I think you and I the other day, we've we've been talking about narrative as an
269. 50:23emerging emotional architectural in a multi- aent space. We talk about that a lot. Like it's a fairly obvious I don't
270. 50:30know how novel that is. Seems pretty obvious to me. Narrative as a collab collaborative ecosystem, emotion as an organizing
271. 50:38principle. Okay, so it's got some points example. All right. Whatever. So, we got that there. So, now we're gonna go um uh now I want you to
272. 50:55um play out the transer mason to this new narrative reality. and write a
273. 51:11report on the impact of the current media ecosystem. All right, that's kind of
274. 51:27cool. So, like let's pretend this was actual all real real research, right? This is the kind of thing where you
275. 51:34might go look at some [ __ ] then you might look at it through some different lenses, then you might come up with new
276. 51:40ideas. Then if you're a future caster like Cindy [ __ ] you might say, "Well, what's this going to do to the
277. 51:50world?" So, let's see what did it do. Executive summary. Between 2025 and 2032, the global media ecosystem
278. 52:01underwent a profound shift from static author ccentric content to participatory emotionally responsive narrative
279. 52:10systems. This transition sparked by advancements in AI, biometric feedback and immers and immersive environments
280. 52:17has catalyzed a new narrative paradigm from content to experience. New formats and fluid story worlds. The death of the singular
281. 52:30author. The author is a conductor, not a god. The 2032 Emmywinning story story world ashfall has over two million
282. 52:44acknowledged co-authors. This is kind of cool. Okay, great. Now, I want [Music] you to give me three concepts for how we
283. 53:01might visualize The new world of
284. 53:16narrative that includes the impact on current media into a
285. 53:31single infographic. that will make it instantly
286. 53:45understandable and bring the research to life. Okay, so let's see. Okay, the story wheel from broadcast to biometrics. I
287. 53:59don't like it as a wheel because I don't think this is going to be linear. A radial infographic showing the
288. 54:05evolution of narrative as a timeline, but instead of progressing linearly, it radiates outward from a central point.
289. 54:13Narrative 1.0 linear broadcast toward the new emergent narrative 4.0. Okay, that's not bad. So, it's not in a cycle. It's radiating
290. 54:23out visual structure. The center is the 20th century linear narrative. Maybe we do all of these. We'll we'll we'll see which one we like.
291. 54:34The living narrative web, an artwork visualization like a neural map that shows how different media nodes are now
292. 54:43interconnected. And then the narrative machine, how a story is made. Now, this is kind of cool. All right. Um why
293. 54:58don't Oops. Why don't [Music] you take a stab at number one?
294. 55:25[Music] Is this making sense? Am I Am I totally off my rocker? Is this making
295. 55:40sense? I'm here slow to find the chat. day. So, one of the things Cindy was saying today, which I thought was really
296. 55:55or no, it was actually M Turner, but then Cindy commented on it. M. Turner was talking about um you do all this research and a lot of
297. 56:08times it doesn't have an impact because people can't instantly get it. And like and like could this
298. 56:23new visualization tool be a way to take research and and turn it into things that people get easier. Co-authorship spatial
299. 56:40worlds narrative 4.0 know linear co-creation, co-authorship, spatial worlds, dynamic cannon, AI agents,
300. 56:48emotional intelligence. So that I don't quite get. So I'm going to say I think it is flat and boring and lacks clarity.
301. 57:03Wait. Flat and boring and boring and lacks clarity of the
302. 57:18evolution. Make it have a dark background and a much more sophisticated color.
303. 57:36pallet and info design. Don't be lazy like me, you poopy head.
304. 57:54And again, it's always good to swear at your large language model and put them down and judge [Laughter]
305. 58:11them. Oh, maybe convert it to a kids book. Oh, that's pretty cool. Like that idea. Nice thought process, chat. Yeah. Well, that's
306. 58:30like like I kind of feel like the the transition if if you think about what Manis does where it goes off
307. 58:41and does these incredibly complicated multi-step research things and then writes a report and generates code and
308. 58:48does whatever it does and then comes back and reports to you. I feel like we at the center of these workflows are going to increasingly have
309. 59:02to think about the chain of events rather than just the here's a prompt, here's an answer, and then repeat, rinse and repeat that. Or we're
310. 59:13going to have to think like, oh, okay, I'm going to go do research in these three tools, then I'm going to take that
311. 59:18and I'm going to visualize stuff over here. Then I'm going to take that and I'm going to go over here and I'm going to do
312. 59:24that. Actually, the the uh the idea that this thing's putting out that the idea of central authorship goes away
313. 59:32completely is actually kind of fascinating. Emotional co-creation. Yeah, th this is it's really not good at this. So, we're say
314. 59:44um I'm going to go back up and one of the things I noticed is it doesn't seem to remember the [ __ ] it wrote once it makes an
315. 59:52image. So, let's go copy this and say I don't like it. Let's try this one.
316. 1:00:13See how it does. Now it's a dark background and foreground. I know it it it didn't I was not happy with that aesthetic at
317. 1:00:27all. Can you tell me a super cool Manis use case? I'm obsessed, but I'm stuck in business use cases. Um I Oh, we can go
318. 1:00:38look at it. And I had it um I I told Manis I wanted it to um go analyze the plot of Shaw Shank Redemption and Snakes on a
319. 1:00:51Plane and come up with a new um screenplay outline that combined the best plot elements of both. And it came
320. 1:01:01up with some prisoners on a on a flight that has plane that has snakes on it kind of concept. So I think you could do
321. 1:01:12creative stuff with it. Um I don't know you Oh, you know what could be fun, Ann? is is have it go find
322. 1:01:26um like what are stories that are being told in other parts of the world that aren't told in the US and like have it be a
323. 1:01:35translator. The living narrative web. Let's look at this. This looks kind of interesting actually. So there's an emotional
324. 1:01:45architecture in the middle. There's books, films, collective authorship, biometrics, AI, dynamic story thread, Tik Tok. This actually is
325. 1:02:01is not this actually kind of a cool idea. like if if you're so one of the things that's been
326. 1:02:15striking me is that when I have when I have a story idea that I shouldn't just be thinking of a single output. I should be
327. 1:02:27thinking of multiple outputs. Right? the book, the screenplay, the stage play, the poem, the song, the kids book, the the album,
328. 1:02:40right? Thi think like that. But I'm still very much thinking in single authorship mentality. What this is saying is you take that same
329. 1:02:57basic story intention and you throw it into a framework, an emotional architecture that you design and then other people can
330. 1:03:11participate. Let me show you an example of it. Actually, this is kind of cool. Okay, I'm gonna hop over to midjourney for a
331. 1:03:26second and I'm going to go down. Well, you can kind of see it here. You see all those images have a sameness to them, but I'm going to go
332. 1:03:38find. So, we can actually play with this right now. So see all these images. These are all
333. 1:03:56images that I was just playing around with creating that use the mood board. That is the official brand mood board for the AI salon.
334. 1:04:09So, what I can do is I can copy this mood board and I'm going to pop it into um I'm going to pop it into YouTube and
335. 1:04:23then I'm going to put it up on screen here. Wait, why did that There we go. I'm going to pop it up on screen here. So you see that
336. 1:04:33profile where it's S3 P as in Paul YM 2F as in Frank. I'll also put this profile over at the AI salon in the irregulars channel. All
337. 1:04:50right. So I'm going to say um go make some midjourney images about Uh, let's see.
338. 1:05:05Phrases that are in your head right now about generative AI. The state about the state of generative AI about the
339. 1:05:20state of generative AI. This is really cool actually. Um, here is the official AI salon mood
340. 1:05:40board. And then I'm going to put note use style weight of 50 to 1,000.
341. 1:05:56If you use a style weight of zero, it won't give you anything. And I'm going to show you how we do this in a second
342. 1:06:01here. But oh, let me let me put an example in here. Let me go grab one of these midjourney images. That's a cool
343. 1:06:08one. Curiosity. So, I'm going to download that and I'm going to throw it up in here. Wait, there was a problem. Oh, of
344. 1:06:20course it was too big. Of course it was. Of course it was. Of course it was. You're [ __ ] kidding me, right?
345. 1:06:37Seriously, I Why is he doing that? That's very disturbing. Why does he make that sound?
346. 1:06:52Oh, you see Kyle, he's got he's got he's a Gen Xer. See, they're a little angry. Why? Why is it not letting me upload a
347. 1:07:06thing? You can't get there from here. What the [ __ ] is going on here?
348. 1:07:21I have a feeling I've got some weird [ __ ] with my Wi-Fi. All right, there it is. Fine. Whatever. Okay, so that post is there.
349. 1:07:33Okay, so you can go grab that profile if and here's how you use it. So, see up in um in the little prompt box there in the
350. 1:07:45prompt hole. Isn't that what we call it? The prompt hole. You put in a prompt like if if uh like like I feel like um
351. 1:07:58Rubicon. We've crossed a Rubicon. I don't know what that means, but we've crossed a Rubicon. So, I'm going to just put the word Rubicon. And
352. 1:08:06then I'm going to go pick my style waiting. It's at 50 right now. And I'm at 16 by9 aspect ratio. That's fine. And then I'm going to do
353. 1:08:16d-r space 4. And that's going to make four sets of four. So I'm just going to say go. And so this is now going to go
354. 1:08:24generate a bunch of images. So here's here's why I'm showing you this. Going back to this idea of this
355. 1:08:33um the future of narrative, I I know we were just trying to visualize kind of a a silly idea, but like this actually
356. 1:08:41came up with an idea that I'm really excited about right now that as authors, our job is essentially going to
357. 1:08:54be to create the world, the emotional arch architecture that we can cocreate stories within with our audience. In a lot of ways, that's what
358. 1:09:10this channel is, right? Cindy [ __ ] was the one that pointed it out to me. She goes, "Kyle, this channel isn't about
359. 1:09:16you. It's about the collaboration that happens while the show is going on. Like, it it really doesn't matter what
360. 1:09:23you say." Now, I know she didn't mean to hurt me, but but I cry myself to sleep every night. Nobody left
361. 1:09:31me. I don't I but I think it's a I think it's a really good a really good point. And and the the shared branded mood
362. 1:09:44board for AI Salon means that all of us can collaborate. You all can go make your own images, but you can, if you
363. 1:09:54want to, you can come into the AI salon and go, "Here, I made some really cool images and here's the words and here's
364. 1:10:00the ones I think are cool and why, right?
365. 1:10:17See, these all have kind of a brandness about them. And you can all be making your own your own fully realized
366. 1:10:28self-expressions that are curated through the lens of this brand look and feel. And so imagine not just having a
367. 1:10:36midjourney mood board, but you've got an animation style over in this animation program. And you've got a a kind of
368. 1:10:43music over here. And you've got kinds of storylines and kinds of comedic situations and kinds of dramatic
369. 1:10:53situations. And I, as the author, am creating that world that we all get to play in. That's what this thing is saying. I
370. 1:11:03didn't get that. I because I wasn't talking to you, Siri. I think that was clear. I think we've established that about two years
371. 1:11:10ago. And by the way, Siri, you haven't gotten any better. Like, everyone else is changing the world. And Apple's like, you know,
372. 1:11:20you can talk to it, right? You just what you do is you go, "Hey, Siri." And then Siri will pop up and you talk to it. And
373. 1:11:29then what she'll do is give you a really disappointing response or just say she doesn't understand. And that's the
374. 1:11:36intelligent part that sometimes she can answer it and many times she's smart enough to know that she isn't smart
375. 1:11:43enough. That's kind of the genius. That's we call it Apple intelligence because it's smarter than what your
376. 1:11:51expectations are. You see you see how we work in a circle. We circle it back around to the starting point, right? You
377. 1:12:00thought there was a destination and you're like, "Wait, how am I back here?" That's the genius of Siri. I mean, I used to be an Apple
378. 1:12:13fanboy. Like, like I I would go to battle for them. It's so [ __ ] sad. All right, let's let's have ChatGpt make
379. 1:12:22us a new a new drawing here. [Music] All right. Uh, let's go grab the third
380. 1:12:36concept. The narrative machine. How stories are made. This was the one that seemed most interesting to me. Um, I like that. Let's try this.
381. 1:12:52You know, I'm almost feeling Cindy, if you're still here, like the the the value of these isn't even so much is it going to generate
382. 1:13:01an a visualization that is usable wholesale, although at some point it might, but it's almost more like, oh
383. 1:13:10yeah, that's kind of the right, you know, instead of bar charts, it's a spider graph, and instead of, you know,
384. 1:13:15linear, it's radial. And that those kind of ideas are actually kind of interesting. Did Kyle's screen blur? Oh,
385. 1:13:21it blurred just because I moved the camera too close. Sorry about that. Lol. So, we're out of focus now. Shut
386. 1:13:36up. You're so funny. I could loop that and laugh for a while. I don't know what I said. Where can I find the emot
387. 1:13:42emotional architecture image? I will go um put it into Let me do save image as. I'll go throw it into uh emotional
388. 1:13:55architecture. I mean, it's it's a very incomplete uh infographic, but conceptually, if it like reminds you of
389. 1:14:01something, I'll go throw it in a regulars right now. Um here's the emotional architecture and then I'll say future of storytelling.
390. 1:14:20um motion architecture future of storytelling diagram from the live tonight and then I will go bang and then I will throw
391. 1:14:33this where did I put that here no where did I see screen where where did I save it where I save
392. 1:14:51Hang on. Was it in downloads? No. No. What? What? Why is it not Why is my life so
393. 1:15:06hard? Oh, file wasn't available on the site. Okay, fine. That's why it wasn't my fault. Shut up. It's It's still not there. The [ __ ] All
394. 1:15:23right, we'll go screenshot it. Oh, it's not there. Well, it's here. So, we're going to go grab it.
395. 1:15:46Dtop. Bang. Screenshot. There we go. We'll come back here. We'll look up screenshot because I have two different
396. 1:15:57desktop folders on this machine and I don't know how to get I don't know how to consolidate them because I'm a loser, baby.
397. 1:16:11Okay. So, there you go. The emotional intelligence or the emotional architecture or narrative architecture,
398. 1:16:18but I think it is. It's emotional architecture. Well, that's going to make people think it's [ __ ] psychology. Maybe it's the emotional
399. 1:16:28world. It's the I don't know. This is something worth the living narrative web. As an author, you design a world that
400. 1:16:43your fans and readers and watchers and listeners generate within. And then, you know what's so funny? It's just it is
401. 1:16:55it's just like this channel. We come here five nights a week. We're doing [ __ ] on our own. I'll occasionally say,
402. 1:17:03"Go do a thing over here." We kind of travel over. we collaborate on something. Some people go off and take
403. 1:17:09that and really run with it. Other people like, "Yeah, that was boring. Whatever." But this is like that on steroids. This is like
404. 1:17:19saying we're going to have all these different outputs and I as the aur going to create the framework that
405. 1:17:28all of us can play inside and anything that's generated in there is going to have a AI learning lab an irregularness about it in this case
406. 1:17:41right but imagine Wes Anderson making a world like this is so [ __ ] Cool. Because this is so if you think of
407. 1:17:49Wes Anderson films right now, Wes Anderson as the single author makes his films like he makes his films. But we can right
408. 1:18:00now just go into chat GPT and say I want um let's see, let's upload something here. um
409. 1:18:14desktop. Uh uh uh uh uh. All right, that's that's good. So, let's open that up. So, we got this crazy ass car. say, um, I want you to
410. 1:18:30re-imagine this car in a Wes Anderson film where it is parked near the
411. 1:18:45yachts of Monaco during the 1965 Grand
412. 1:19:03Prix. All right, this is this is where we're going to get Okay. Okay. For all the AI haters out there,
413. 1:19:15no, sorry, I'm not chatting with anyone right now. Um, for all the AI haters out there that are going, "Ah, this is
414. 1:19:21intellectual property theft, all that [ __ ] Not not that it's [ __ ] Intellectual property shift stuff is not
415. 1:19:30bullshit." But all of the tropes about people saying AI is evil because of that. I wasn't able to generate that.
416. 1:19:36Oh, because it resists our Oh. Uh, so can you describe what it might look like?
417. 1:19:55Uh, make that [Music] image. It's now a visual assistant, too,
418. 1:20:11that we just have to tweak. I'm here. Okay. So, we're going to do this thing. Okay. So, so this is going to make an
419. 1:20:20image. Oh, come on. Um, let's see. Don't use his name. Just make the image based on the style you
420. 1:20:35described. You're smart. make the image and don't get
421. 1:20:54sued. This is why I have a Wes Anderson mood board in midJourney. Oh yeah, this exact nightmare, right? Um, okay.
422. 1:21:02So, so right now Twitter is nothing but Studio Gibbli images, right? And Studio Gibbli is notoriously copyright frenzied. I still
423. 1:21:15wasn't able. Come on. Yes. Adjust the prompt. Yeah, good idea. Not that I
424. 1:21:30described it three prompts ago, but but go ahead. Yes. Yes. And don't forget to include the car I
425. 1:21:48uploaded. So, right now, however it gets done, someone wants something that looks like Wes Anderson. They use his name. They generate an
426. 1:21:58image. He looks at that. He's like, "God damn it, they're stealing my shit." Studio Gibbli looks at all of Twitter
427. 1:22:04and it's like, "God damn it, they're stealing my shit." All of the art haters look at this stuff. Oh my god, this is
428. 1:22:11this is [ __ ] painful. All right, we had two good days. We had two good days with um um let's just make it a light color
429. 1:22:28palette with pastel colors, [Laughter] please. I'm trying to make my point
430. 1:22:42here. And Sam Alman put his little [ __ ] safety guard rails back up so he can't [ __ ] use the [ __ ] thing anymore.
431. 1:22:52Um, imagine now if Wes Anderson makes it makes this transition. Wes Anderson realizes Okay, I'm I'm [ __ ] done with this. Um, upload
432. 1:23:06[Music]
433. 1:23:21I don't know where that picture was anymore. That's our [Music] bug. I kind of want that picture of the the thing. God damn it.
434. 1:23:37I'm I'm cranky. This is [ __ ] shitty. [Music] Um, save image as Let me try a new thing.
435. 1:23:51Um, upload. Reimagine this car sitting next to the
436. 1:24:08yachts with uh sitting next to the yachts in Monaco. I'll get back to the point in a second. in
437. 1:24:231965 at the Grand Prix. Um, the colors should be muted with pinks and orange
438. 1:24:39hues, making making it all a bit surreal. The fashion should feel
439. 1:24:56um familiar yet off. Okay, that should make us an image. Let's see. We didn't say we're in
440. 1:25:12a new chat. We didn't say the word Wes Anderson. Yeah, this really seems like a pro workflow. You wouldn't get your head
441. 1:25:20bitten, chewed off, and swallowed in for trying in a studio. Yeah. Okay. So, here's the thing. So, now imagine Wes
442. 1:25:28Anderson gets religion around this AI thing and realizes the solo is going to be a thing of the past
443. 1:25:43because everyone can now make anything. But people people are lazy. People like to be directed. They
444. 1:25:58like to be told what to do. They like to be part of a tribe. So now imagine Wes Anderson saying, "I'm going to create a
445. 1:26:08co-creative world that within it you can buy things, you can sell things, you can make things. I'm going to create a universe that you
446. 1:26:21can create video games inside and you can have your own creative expression just like the the AI salon mood
447. 1:26:32board. You're going to make anything you want within this, but it's going to follow these certain constraints, which isn't a bad
448. 1:26:41thing. And then it creates this infinitely complex world where there might be millions of co-creators. And within that, some of those
449. 1:26:54stories are going to take off and become hits. Wes Anderson is going to get compensated for anything created in this
450. 1:27:02world and is going to get a cut of any of the transactions that happen in this world. But also within this world, there
451. 1:27:10are going to be filmmakers and game producers and image generators and 3D printers that print little Wes Anderson world kind of
452. 1:27:21things that are going to get compensated for that. That you're going to be able to make a living within a world that someone like
453. 1:27:31Wes Anderson creates. There's going to be entire economies existing inside what are they called? Emotional architectures. This is [ __ ]
454. 1:27:44cool, right? So that's not quite Wes Anderson, but still like imagine, you know, that was the aesthetic. So anytime
455. 1:27:53you made a car, it was kind of in this neighborhood. Whenever you created characters, they kind of were dressed like this.
456. 1:28:06Right. Isn't that cool? I think that's [ __ ] cool. My ear [Laughter] itches. That was really annoying that it
457. 1:28:18would just would not make that image. Really annoying. Good lord. What made your studio? Hang on. Let's see. What
458. 1:28:32major corporation is Wes Anderson associated [Laughter] with product? No, no, no.
459. 1:28:49Um, let's see. No. who distributes his movies like is it Disney or Paramount
460. 1:29:10etc. Bottle Rocket Rushmore Disney Fox Search Light. It's noteworthy that with Disney's acquisition of 20th Century
461. 1:29:24Fox, films previously defi dist distributed under Fox labels are now part of Disney. Okay, so here's the
462. 1:29:33logic. Here's the logic that that Open AI put into their image generation tool. If image is tied to Disney, do not make
463. 1:29:48it because did Disney is a notorious lawsuit machine. Oh, [ __ ] All right. Wasn't aimed at you, but
464. 1:30:02more the obstacles that are the default safeguards. Oh, yeah. No, I get that. um still going to be demand for someone
465. 1:30:10who shows up with a functionally functional proprietary workflow. Oh, I agree. I listen what what I'm talking about is right now there's not
466. 1:30:24infrastructure for Wes Anderson to do that. if he wanted to do that or if I wanted to do that. Right. The the only thing that I've done
467. 1:30:33that is remotely in this neighborhood is this new um mood board for the AI salon that you can create images for the
468. 1:30:44AI salon. But like now I'm thinking like what is the equivalent of that for for music and for I don't know other [ __ ] like like I
469. 1:30:58right now it's all ad hoc. This is a this is a major major opportunity. It's a major opportunity to create an
470. 1:31:13infrastructure where an author can go in and set the emotional boundaries of a world of a creative of a co-creation
471. 1:31:25space that everyone gets to participate in. the person that generates the world gets compensated for that activity and then
472. 1:31:39the people within that world can benefit from it either through you know access to content for free like like it could
473. 1:31:47be a kind of thing you come into that Wes Anderson world or let's just call it the AI salon world you come into the AI
474. 1:31:53salon world you can make [ __ ] you can sell [ __ ] um if you just want to come in and consume stuff you pay for that if
475. 1:32:02you want to come in and contribute ute stuff then you don't have to pay for that. You get access to content because now you're a
476. 1:32:09contributor. And then if you contribute something that starts to make money over a certain threshold, you get that or you
477. 1:32:17know some piece of that and then the the entity gets the whole and this is the kind of thing you know all anyone's talking about right
478. 1:32:31now is the AI is going to take our jobs. The AI is going to take our jobs. The movie industry is dead. The music
479. 1:32:38industry is dead. Well, what if it's just changing? What if our jobs are participating in these
480. 1:32:54ecosystems and you could be someone that just lives in Wes Anderson world or you could be someone that bounces around
481. 1:33:00between 20 of them and your specialty is rave music in virtual stages and you go to all these different worlds and you
482. 1:33:10you create raves within the different worlds and you get really good at that and you get known for that or someone
483. 1:33:18else is just in this one world and they're kind of the master of all of this world. That's [ __ ] cool. [Music]
484. 1:33:37[Music] All right, that was a rant, but that's pretty cool. All right. Yeah, the Wes Anderson
485. 1:33:52metaverse massively disruptive to design of all types. Well, but it's, you know, it's funny is it's a metaverse, but but
486. 1:34:01it's not just a metaverse, right? It's a it's a cocreation universe because if everything's going to be
487. 1:34:17hyperpersonalized, then no one get gets compensated anywhere, right? If if I can just say I want a movie that's like
488. 1:34:24snakes on a plane, but instead of snakes, it's rabid bunnies, and instead of Samuel L. Jackson, it's it's William
489. 1:34:32Defoe. And that starts playing for me. Yeah, you could, you know, find individual copyright holders that are
490. 1:34:43drawing and figure out a percentage. Or you could just say, "No, let's put that movie creation thing inside of a world that is an economy.
491. 1:34:55It's an ecosystem of storytelling and of entertainment and and business. I think there's something there. I think there's something now
492. 1:35:10people. I'll tell you what. Thoughts, questions? Am I full of [ __ ] Is this Is this making any sense?
493. 1:35:33Everyone's asleep. All right, I took him out. I took him out.
494. 1:35:49Let's go see if anyone [Music] uh if anyone made anything over here in AI. Look at that future of AI from
495. 1:36:05Valerie Cox. That's beautiful. And look, it's it's it's like it's got an AI saloness about it, right? This is the idea.
496. 1:36:16a developing ecosystems. We're making pics. Oh, that's that's true. I told everyone to go make pics. You're off making
497. 1:36:22pics. Just listening to me ramble in the background. Ah, Cindy [ __ ] A salon.
498. 1:36:38Nice. Oh, man. Very cool. I like that. That's nice. It's
499. 1:36:53nice. Oh, lordy lordy lordy lordy lordy. Who wants to talk about anything? Bedtime story with Kyle. I
500. 1:37:09know. people making pics. That's very cool. So, it's like a write a novel, invite people into that world. Yeah,
501. 1:37:25Kak, it's it like it's it's exactly that. It's like so right right now in today's media world, if you go back to the
502. 1:37:37before time, right? If you go back to 2022, the before time when to write a novel, you actually had to write a
503. 1:37:48novel. And it took you anywhere from 3 months to 30 years to write that novel. And then you would get it written
504. 1:37:58and then you would try to get it published and they would print it on paper and they would stick it in a book.
505. 1:38:03And in the 2022s, they would stick it on a Kindle. And then in the after time, it now becomes really easy to make a book.
506. 1:38:16And so, not only can you make a book, you could make it a kids book with illustrations and you could make it an
507. 1:38:21album and you could make it a this and you could make it a that, but it's still this idea of you're the author of that
508. 1:38:29book and all those individual pieces. And the hypothesis that this thing came up with tonight is you think of that story and that's
509. 1:38:46just your story in that world in that emotional architecture. So you then we as authors get to elevate ourselves up above the
510. 1:38:58the the tactical outputs and just say what would this world look like? What would it sound like? Who would be the characters in it?
511. 1:39:07What would they be dressed like? What era is it? What are the technologies that are there? Right? So, just like you
512. 1:39:15would if you were writing a book, you're going to design the world of that book or the framework of the thinking in a business
513. 1:39:25book. And now we elevate ourselves above that tactical output and just say, I'm going to create this dome that anyone that comes inside this
514. 1:39:36dome is free to create anything, you know, that we have the capacity to make within this domed world. But it's all going to be within that
515. 1:39:50world. And then the creators and managers of the world get paid. There's an economy that allows creators to get
516. 1:39:59paid. There's a world in which you just go in there and consume content, right? You don't have to co-create.
517. 1:40:07You can just go listen to Beyonce's music or anyone else of Beyonce's fans that made music that sound like this world.
518. 1:40:20Imagine John Mayer taking all of his music and creating a sonic emotional architecture where you could enter that
519. 1:40:28world and you could listen to all of his music, but you could also take a custom-designed a a customtrained version of
520. 1:40:38Suno that was trained on John Mayer's sound and you could go in there and you could write your own songs and when you
521. 1:40:46use the tools to make the lyrics, they'll have a John Marness about them. And when you make songs, they'll sound John
522. 1:40:54Marike. And some of those songs might be so good that they become hits. And some of them might be so good that they
523. 1:41:00become hits that John Mayer goes and plays on tour. And because this thing will be all blockchain tracked, we'll know who made what this
524. 1:41:12world, and they'll get compensated, and so will John. That's [ __ ] cool. That's a way to end a [ __ ] weekend or to end the
525. 1:41:28week. You know what I'm saying? Jeff Flanigan, I had 4.0 remake a very old snapshot of me and my mom as
526. 1:41:42a professional painting. It turned out really good. That's very cool. That's very cool. You should get some research from
527. 1:41:55Gettysburg to feed GPT and have it create photos of the battle. I created a whole music video of uh Gettysburg battle scenes that I brought
528. 1:42:06to life in video that all look like dger types. It's pretty cool. First and last frame interpolation will be huge. Yep, I agree with
529. 1:42:20that. It's just crazy. We're just such in such [ __ ] wild times, man. Anyway, I dig it. Chat GPT in this
530. 1:42:32modern narrative theory. Here's the other thing that it came up with. sensory environments, personalization, emotional story, archetype,
531. 1:42:43dialogue. So, you have a narrative generator at the center. Yeah, I could see I could see creating a whole architecture where
532. 1:42:52you've got a narrative generator that that develops voice and characters. You could have look and feel. You could have
533. 1:42:59audio. Like, what do the sounds sound like? Is it is it is there like birds chirping and it's this really peaceful
534. 1:43:07place or is it like [ __ ] terrifying and everything's like like imagine a GMO del Toro world. You know what this is? This is theme
535. 1:43:22parks that are dynamic. So it's not just a metaverse. There will be metaverse areas within it. But this is a theme park
536. 1:43:36where the crowd at the theme park gets to make new rides, gets to make new stories, gets to make new storylines. [ __ ]
537. 1:43:46insane. It's cool. All right, that's where we're going. There's your five-year vision of of AI dynamic Roblox. Yeah, exactly. Yeah.
538. 1:43:57This is to a to a degree this exists in Minecraft and Roblox, right? Those are kind of worlds like that, but but I think that they're very
539. 1:44:06limited to it's got a specific kind of gameplay and it's got a specific kind of mechanic and a specific kind of
540. 1:44:14aesthetic. And this would be a much more broad open kind of thing where some of it might just be making music.
541. 1:44:26It's not making music inside a virtual reality world. It's just making music in this intellectual property dome. That's
542. 1:44:39cool. That is cool. All right, I'm going to get on out of here. So, um yeah, your homework for the weekend is explore the
543. 1:44:48[ __ ] out of Chat GPT image generator. Um, if you want to geek out on that anthropic paper, I'm going to probably
544. 1:44:56read that this weekend. Um, and now start thinking about the kind of world you would create that would allow other
545. 1:45:05people to create within your world. What would you want that to be? That's that's a fun thought exercise for the
546. 1:45:14weekend. All right, everybody. All right, everyone. Listen, you have yourselves a fantastic weekend. Next
547. 1:45:21week there's a AI salon uh presents on um on Tuesday night with Brian Moahan and Adnan talking about their book
548. 1:45:32Culture Shift about AI and how it's shifting culture. We talked a lot about that tonight. Um and uh yeah, have yourselves
549. 1:45:42a fantastic weekend. Hope you had fun tonight. All right, peace out everybody. Salom. So long.