4/8/2025 - Exploring the AI Frontier: Tools, Trends and Keeping Up in a Rapidly Evolving Landscape

Transcript

1. 0:00What are we going to sing? H What are we going to sing? Which glasses I want. I guess these are fine. [Music]
2. 0:25[Music] Round myself of my wrinkled [Music] sheets. Smoking. It was all a
3. 0:41dream in my head and a hole in my heart. Says it was more than just your be. [Music]
4. 1:04[Applause] There's been something baby I'm trying to say. For aging, it seems I don't know
5. 1:18how. Past and a future now surrounding me. Surrender to an empty f. There's been a little
6. 1:34trouble since you came to my [Music] rescue. And if you like all of the rest, I would have quit you long ago. But I couldn't do that.
7. 1:49[Music] Oh, tell me now. My mind never went too well. Make a man crazy, make him cold as
8. 2:03hell. I'm a woman that you wish me well, but I'm trying. Still going to have to find my own way through. [Music]
9. 2:21[Applause] Good evening good people. What is shaking? What's going [Music]
10. 2:37down in a westerly? [Music] Car is my
11. 2:51train and driving. I've been wondering what it is I'm from. Again, feel like an 80y old [Music] man holding on to
12. 3:0929. And up ahead on that [Music] horizon is a California [Music]
13. 3:24line. Up had a truck scaring a wide [Music] log. Cute little front door and two windows. My love is sure whether I should
14. 3:41laugh. You see, I broke a home of myself once. This I stumbled to that dog. I read a note by the dog said, "Don't you come around
15. 3:59here anymore. Well, I've had enough. There's freedom around. Never was good with
16. 4:16decisionions. At least that's what I've been [Music] told. Good singing, Champy. Real nice
17. 4:30tonight. Solid, good tone. [Music] [Applause] [Music]
18. 4:49Well, I heard there was a secret call and played in a pleasing love. You don't really care for music, do you? goes like this or
19. 5:07[Applause] minalujah hallelujah hallelujah hallelujah
20. 5:21hallelujah [Music] Yeah. Strong and you needed proof. Saw her bathing on a roof. Her beauty and
21. 5:37the moonlight through you. She tied you to the kitchen chair. She smashed your throat and cut your head from the lip. She drew. Hallelujah.
22. 5:53Hallelujah. Hallelujah. Hallelujah. [Music] Hallelujah. What do you think, champ?
23. 6:11[Music] Uh, I am still sharing. All right. Good
24. 6:25day. What's up, producer Brandon? Good to see you, sir. [Music]
25. 6:53I have 10% left on my watch battery. Ah, what's happening? Good people. I'm up to 46 published titles as
26. 7:08of tonight. Good. Keep it going. Nice. I am more convinced than ever that the era of the singular
27. 7:24author is waning. [Music] I think AI is going to transform just
28. 7:40[ __ ] everything. Like everything. I lost my subscription star. What the hell? What the hell's going on? I have no
29. 7:58idea. There's a new uh there's a new video tool I want to play with tonight. Well, it's not a new tool. It's an old
30. 8:05video making tool, but it looked like they had something that looked interesting. So, we'll see if it [Music]
31. 8:28is. There's something going on that's been going on for a while. The subscriptions expire. Where's that? On YouTube or here on
32. 8:42TikTok? I guess on [Music] YouTube. Oh, it's an US problem. Oh, that's right. I've got to No. Hello,
33. 8:59Kyle. Am I late? No, I've got to verify. I've got a I've got a what do they say in the crypto world? I've got to dox
34. 9:08myself. Um YouTube wants me to verify my identity. So I just I I didn't get to that today. I thought you YouTube had dropped
35. 9:19it. No, no, no. We got a I got an email that was like you can't monetize your channel anymore. And then I sent a note to
36. 9:27producer Brandon. I'm like what's up with this? And then I read the next email from them which said, "You've got to um prove that you're
37. 9:37[Music] real. I mean, we have a [ __ ] business [Music] account. That's why they don't have the
38. 9:53badge anymore." Okay. Oh, here. That's why they don't have the badge anymore. Oh, if it's on Tik Tok. Tik Tok's still fine with
39. 10:02subscriptions, but if you don't have your badge, it's cuz your subscription lapsed. Kyle, I'm doing my persona with chat
40. 10:10GPT and make. What do you think? I think awesome. Beautiful. You've entertained me nearly every night for a couple of
41. 10:18years now, Kyle. At least I could do a sub. Thank you very much, Mr. Architect. I appreciate that a lot.
42. 10:29Oh, thing I was going to talk about or a thing I was going to talk about is we've kind of talked about this a little bit, but it's like I
43. 10:45am, how do I put this? If anyone asks you or if anyone asks me what tool should I
44. 11:01use, we're at the point where it literally doesn't [ __ ] matter. It doesn't [ __ ] matter. Use co-pilot, use Claude, use
45. 11:12Perplexity, use Chat GPT, use Gemini. Like, which one do you like? Which one fits your personality? Which one is one that you
46. 11:23took the [ __ ] you know, four hours to learn how to sort of prompt it and you sort of get decent results
47. 11:33from, you know, which music generation tool? Doesn't matter. They're all fine. Which image generation tool? All all of
48. 11:42them. Midjourney's better. Okay. ID's okay. Chat GBT is better. Okay. I would actually argue argue chatbt image
49. 11:53generation is better than anything out there right now with the exception of midjourney for aesthetics. If you want it to look arty, use
50. 12:04midjourney. I plugged the AI learning lab today on an interview. Oh, awesome, Gareth. Thank you very much. I appreciate that a lot.
51. 12:11Um, yeah. You know, I've got to find a way to spread the word about this because since I've since I've split the stream to YouTube
52. 12:24and Tik Tok. Um, YouTube doesn't promote like Tik Tok used to and Tik Tok just isn't promoting the show anymore. Um, so
53. 12:35I got to find a way to get out there. So, anything you can do like that to promote this would be great. Um, and
54. 12:41then I gota I gotta put my put my noggin together with producer Brandon. We got to figure some [ __ ] out. Uh, because you
55. 12:49know it's better for all of us if there's more people here. And I would love to get I would love to find a way to get new people in
56. 12:56here. Not that I don't love y'all. Y'all are regulars. Y'all are regulars are what make this place special. And I would
57. 13:04love to see some new faces in here. They asked me how I stay on top of everything in AI. with how fast it's changing. So,
58. 13:12this is kind of my point, Gareth, about um I mean, yes, what you said, get your ass in community. Whether you come to
59. 13:19the AI learning lab every night, whether you go to the AI salon, whether you join she leads AI, doesn't matter. Get your
60. 13:24ass in community and participate with it because holy [ __ ] is stuff moving fast. I mean, it's it's funny. Well,
61. 13:34it's not funny for her. my co-founder in StoryVine. I'm the handwavy guy, right? I'm the inventor. She keeps the trains
62. 13:42running on time. So, she's the one paying attention to privacy policies and data policies and things like that. And
63. 13:51I keep changing tools because the target keeps moving for what's, you know, what what's going to produce a better product
64. 13:58for the for our clients, which is ultimately what we need to do. and and it's like from two weeks ago I'm using two
65. 14:07completely different tools than I was two weeks ago. And she's like, "Wait, I've have I heard of those tools before?" And I'm
66. 14:14like, "No, they didn't exist two weeks ago." One of them is Cartisia text to voice. Cartisia is amazing. I think it's better than 11
67. 14:25Labs. Um, and then Hedra or Hedra. H E D A for voice to animation is better than anything. But a new one came out tonight
68. 14:38called Vigle, which has got a new voice to um animation model. So, we're going to go look at [Music] that. Kyle isn't happy with us anymore.
69. 14:52He wants a new version. No, you all are fine. You all like Word 5.1 before before it turned into bloatware, right? You're a good solid word
70. 15:03[Laughter] processor, trading us irregulars for regulars. No, I want new irregulars. I want I want people You know what I want?
71. 15:16And I'm I'm giving a talk tomorrow night um at a just at a local like a whiskey club kind of thing. Um I'm doing an AI
72. 15:27talk tomorrow night. And I'm I'm really hoping that there's a couple of people in the audience that are like I was
73. 15:34thinking I should maybe try chat GPT sometime. So we're going to be late tomorrow. We are going to be late
74. 15:40tomorrow. My my talk is only from 5 to 6:30, but then I've got a dinner. Someone's in town from one of our
75. 15:46clients, so I've got a dinner tomorrow night. So, it might, depending on how late the dinner goes, it might even be a
76. 15:53uh scrap tomorrow night. I'll just come on for half an hour and do a little something. Viggle. Vi Gle. V is in
77. 16:00Victor. I have three or four different communities I stay involved in just so nothing slips through the cracks. That's
78. 16:07smart. I plug this channel to everyone I know. Beautiful. They're probably annoyed how much I talk about it. Yes.
79. 16:14Okay. This is the thing. If I if I teach you nothing else in this channel, be shameless about how much you talk about
80. 16:20AI and I don't if it's this channel, awesome. But like just [ __ ] put yourself out there. Do I hear Professor Brandon take
81. 16:31or producer Brandon takeover? Uh it Hey, it very well could be producer Brandon. You've got the keys to the car. So, so
82. 16:41you can probably uh teach them all how to make some Microsoft stuff tomorrow since I keep resisting it. Juporro imaging make it more
83. 16:51irregular. Exactly. Oh, we already bored you. No, not at all. No, you know what it is? I'm really pissed off at [ __ ]
84. 17:00Tik Tok that they just don't produce they they just don't promote this channel anymore. The lives. It's just it's it's really frustrating.
85. 17:10I just saw Hawk, some old dude that I just like watching, they just canled his channel. Yeah, because I'm not selling Madagascar
86. 17:20vanilla beans. Yeah. Yeah. If I were if I were twerking and uh and selling cardboard cutters, ice makers, and
87. 17:28Madagascar vanilla beans, we'd have all sorts of people. Or if I had if I had gone down the battling Dar the Star Tik Tok [Laughter]
88. 17:44battles, I would be a Tik Tok bajillionaire. But but no, no, I just wanted to go with integrity. [Music] [Music]
89. 18:04Yeah. So, I'm I'm kind of thinking of coming up with a new way to to recommend what people should learn because it's
90. 18:20like like it used to be it used to be pick a large language model, pick an image generation tool, you know, pick a video tool, pick
91. 18:33a right, it it was one and now there's like five or six in every category with the exception of music which has three three big
92. 18:48ones. And then and then within each tool, there's tools within the tools. This one's better at animation.
93. 18:56This one's better at natural movement. This one's better at, you know, consistency between frames. Tik Tok pin.
94. 19:03Techie Becky, my sister is isn't talking to me because she's upset I keep talking about AI. Keep doing it. Keep doing it. She's
95. 19:15afraid of it. I mean, so so quite seriously, Becky, here's here's what I would say. Have empathy for her, for that fear.
96. 19:27Um, I mean, to the extent that you can, try to understand it and and understand where she's coming from and see if you can find an on-ramp.
97. 19:38Um, because here's the reality for her. It's not going away. Like like it's the only thing the only thing I'm 100% convinced of 100%
98. 19:53positive like a thousand% AI isn't going away. It's not going to slow down. It's not going to go away. Did you hear that? They they
99. 20:04[ __ ] unextincted the direwolf. [Music] When was the last one? 7,000 years ago.
100. 20:1870,000 years ago. Please consider sharing the live. Yeah, do that. They found a tooth of a dire wolf somewhere in a riverbed and then they
101. 20:29found a skull somewhere else and between those two they had enough DNA and they basically took a wolf which is about 99% I think of the
102. 20:42DNA and they now have two seven-month-old unex unextinct unextincted deextincted direwolves I think they said a woolly mammoth is
103. 20:59[Music] next. So the [ __ ] ain't going
104. 21:17away. You know the other thing you could do, Becky? I mean, listen, she's she's got to be open to it. But I mean I like
105. 21:23I don't think force feeding anyone AI serves anyone. So I think it's I think understand where she's coming from is part of
106. 21:32it. And if there's if there's a particular area that she is resistant to. So like maybe she's an artist or an illustrator and she's like
107. 21:43totally freaked out by the images, but maybe she wouldn't be freaked out to write emails faster. Like one of the one of the
108. 21:53things you can do is you can segment the work that's personal to you where you feel existentially threatened from the
109. 22:01[ __ ] where AI can make your life a little easier. And so if you can help her identify like is is there an area where
110. 22:09you've got a hard boundary? I remember um Jane Endicott who's one of the early salon members [Music] u she was a copywriter and when I first
111. 22:21met her she was like should I be worried and I'm like yes I'm like you need to learn AI you know pronto and she and she
112. 22:29very very quickly did and like very quickly kind of 10xed her output for her client which the client basically coinciding with the client
113. 22:39going I need you to do more work, but I'm not going to pay you more. Um, she got her hands on AI and what she said
114. 22:46was there's a hard line for her between her creative writing and the the writing she does for work. And then over two
115. 22:54years that's gotten more sophisticated. She'll use AI to help her ideulate and outline things, but then she still does
116. 23:02writing writing to write to on the creative writing side. So, every person's got to find that personal boundary. Um, okay. Let's see.
117. 23:14Shopify CEO said in an internal memo that everyone should be using AI and if you want more headcount, you have to
118. 23:21prove AI can't do that job. Yeah. The there the the head of Shopify is calling AI a core competence. A core competence. [Music]
119. 23:35And yeah, I guess saying to managers, if you want to increase headcount, you've got to prove that AI can't do that job.
120. 23:45That's if the writing isn't on the wall, it is scribbled in pencil on the toilet paper. You better get with it. AI 2027.
121. 23:58As someone who's trying to find a new company, I'm like, [ __ ] we're all going down. the AI enthusiast in me, it's
122. 24:04like, well, I'm glad I'm fluent in AI and and we all saw this coming. Like, I I just at this point, I mean, it's
123. 24:14been what, two and a half [ __ ] years since Chat GPT came out. It's starting to get irresponsible for people to be on the sidelines with
124. 24:28this [ __ ] thinking about it. I mean, the the agents haven't even really started yet this year. We, you
125. 24:41know, we got a little bit with Manis, a little bit with [Music] Operator. Oh, the the Oh, Spotlight the Report.
126. 24:55AI2027.com. What's that? Is that [Music] the Is that the Spotify one or is that something different?
127. 25:17AI-2027.com. Oh, this is the thing. Is this is this from Daario Amade? Who's who is this? We predict that the impact of superhuman
128. 25:25AI over the next decade will be enormous, exceeding that of the industrial revolution. Yeah, I don't disagree with that. We wrote a scenario
129. 25:34that represents our best guess on what it might look like. It's informed by trend extrapolations, war games, expert
130. 25:44feedback, experience at OpenAI, and previous forecasting success. Mid 2025 stumbling agents. Yes, the world sees the first glimpse of AI
131. 25:56agents. They're jankier than [ __ ] This is AI-2027.com. I'm trying to use Manis to vibe code a game. You know what's funny?
132. 26:06I had I had Manis write me a game. I was really impressed with how it did it. It then wrote me a game and I had to like
133. 26:13install it on my Mac somehow and I couldn't quite figure it out and I was just like, "Oh, [ __ ] it. That's too much like work.
134. 26:24Hey Kyle. Yeah, what's up? Ju just so you know it's um just giving you some context here. This report uh was
135. 26:32something that was put together by some of the leading minds of AI AI futurists and it's an interactive report. As you
136. 26:39scroll down the chart on the right hand side starts to show what impact AI is going to have on society. Oh, that is
137. 26:49it's a timeline over the next until 2027. What is going to happen to AI? Oh, I get it. Now, that is very clever. Thank you. Thank you for
138. 27:04that. All right. Got it. Okay. I see. Okay. So, over here we've got coding ability. coding ability is just starting. Basically,
139. 27:20hacking, boweapons, basically everything's just starting. And we and we have a little bit of compute and it's mostly in
140. 27:28America and a big chunk of it open AI. 2,000 unreliable agents thinking at 8x human speed, minus 25% approval.
141. 27:44[Music] Oh, for open AI. All right. Mid 2025. Stumbling agents. Late 2025. The world's most expensive AI.
142. 27:59Open brain is building the biggest data centers the world has ever seen. Wow. To avoid singling out singling out
143. 28:14anyone company, we're going to describe a fictional artificial intelligence company we'll call Open Brain. Oh, okay. Open
144. 28:26AI. So, basically, they're going to they threw they threw a dot worth of compute. They they did a dot worth of compute for GPT3.
145. 28:40um seven bricks and I guess a brick has got nine dots in it, something like that. And then a whole lot more bricks
146. 28:50for agent one, which is going to be the I don't know if that's going to be GPT5, but that's going to be their
147. 28:59agent. Those are my ink boxes. Yeah, exactly. There are lots of AI ident data centers across C's that aren't being used
148. 29:10because of R1 right now. Oh, that's interesting. Mind reading and genome encoding. Yeah, the mind readading stuff is going to be
149. 29:19crazy. All right, let's keep going down this this prediction chart or this prediction thing. early 2026 coding automation and that's
150. 29:31you know everyone's saying that Meta Open AI anthropic are all saying we're going to have fully automated encoding
151. 29:38this year so I guess it gets up to speed early 2026 mid 2026 China wakes
152. 29:55up Wait, isn't China Isn't China already awake and kind of kicking our ass with with models that are just as good and and cheaper? 10 times
153. 30:21cheaper. Late 2026, AI takes some jobs. This This feels like a weirdly conservative maybe I I'm just reading their headlines. So
154. 30:38So now we're starting to to get into monthly monthby-month predictions. Agent January 2027, agent 2 never finishes learning.
155. 30:53February, China steals agent 2. Nice. So, wait, this thing is we just passed this. We're just now Yeah, we're we're in the slow the
156. 31:06the flat part of this curve right now. Huh. Algorithmic breakthroughs in March, April 2027. Alignment for agent 3. Yeah, this we
157. 31:19start getting stuff every couple of months that are The next excuse me the next things May 2027 national security June 2027 self-improving
158. 31:34AI open brain now has a country of geniuses in a data center that's from Dario Amade research automation deployment
159. 31:50takeoff off a thousand tokens a
160. 32:07second July 2027 the cheap remote worker so cost plummets LLM inference prices will have fallen 9x X to 900X a
161. 32:23year. Basically, intelligence is free. Geopolitics, the superhuman AI researcher, agent 4 is in September of
162. 32:372027. So, basically, basically, we've got two [Laughter] years. Again, you know, I I I go back to I mean, this is a this is an interesting
163. 32:48report. I'm kind of reading it on the fly here, so I I don't know enough about it to to really comment on it. Um, but go to
164. 32:57ai-2027.com and it's that's a really slick uh interface to dynamically change that chart as you read the article. It's pretty smart.
165. 33:09Um, yeah, it's it it is irresponsible at this point for anyone that's not getting on this train. It just it just is. It
166. 33:18has a choose your own adventure ending. Oh, cool. Nice. Let me see. Choose your ending. Slow down or race. Well, I'm going to [ __ ] race.
167. 33:32We're here. The committee votes six to four to continue internal use of agent 4. Oh, and then it keeps going. OpenAI's approval has dropped
168. 33:45to minus 45%. Their revenue is 250 billion a year on a $9 trillion valuation. The Agent Five Collective.
169. 33:58Yeah. Now, now agents are in swarms. 2028 the the AI economy. 2029 the deal. 2030 takeover. By 2030, the robot
170. 34:14economy has filled up the old SEZs. I don't know what those are. Wow. Crazy. All right. Well, that's super
171. 34:28cool. I dig it. You know what's amazing? It's like We're at this place
172. 34:45where we're paying attention to this. We know enough about what AI is making possible to see the path to what this paper is talking
173. 34:55about. Maybe it happens by 2027 like they're saying here. Maybe it happens by 2030. Doesn't matter. it at some point in the next two to five
174. 35:08years. Everything will have changed in the next two to five years. Everything will have changed. And right now, we're still
175. 35:21having conversations about, well, was it trained on someone else's data? Does it make mistakes? Why do their faces look
176. 35:36weird? I think it's hype. I mean, I've just I've never I've never been I've been on the planet for a lot of years, relatively 59
177. 35:57years, and I've never felt like in the early days of the worldwide web, I knew that I was ahead of the curve, but when I look back at people
178. 36:09that were were when I was talking too much about the worldwide web in in 1994 and 1995 and all my theatrical friends
179. 36:18were telling me to [ __ ] off. We don't want to hear about the web anymore, including my current writing partner who
180. 36:23told me to [ __ ] off because he's like, I'm a designer. Why would I need to learn HTML, you big fat dumb dumb?
181. 36:33Um but but when I looked back to them, I'm like, "Oh, okay. They're just they're they're not that far behind
182. 36:41because the the acceleration of the worldwide web was was really slow, right? It was just like like we were getting a little more
183. 36:52knowledgeable, a little more knowledgeable. You can still kind of back look back and go, "Well, you could
184. 36:57now you could buy HTML for dummies and you could actually learn this stuff and catch up." Just look at I don't know you were
185. 37:07here, many of you were here the other night when I did that sort of joke rant on all of the different models within chat
186. 37:17GPT and they're adding two more now. Like it's it's it's [ __ ] insanity, right? That's just one tool. One tool has what? 10 models in
187. 37:31it plus Sora plus operator. I don't know how you catch up with this. Like we can't keep up with it
188. 37:46and we're [ __ ] trying to keep up with it. I just feel like the the the [ __ ] velocity is doing this and it's
189. 37:57like, you know, with the web thing, you know, was kind of on this slow trajectory and I could look back and
190. 38:03still see the people. I just feel like at this point the this [ __ ] rocket is taking off. I've never experienced anything
191. 38:12like this. Like I just feel like the gap between people that have some sense of what's happening and people that are just
192. 38:23utterly unprepared for what's coming is massive. We could pause the current frontier and still spend 20 years not
193. 38:32fully leveraging it. I totally agree with that. Honest to God, I feel like I could just take GPT40 right
194. 38:41now, just as it is right now, and spend the next decade trying to figure out what to do with it easily. Easily
195. 38:59crazy. Be like a rock skimming over the water. Yep. [Laughter]
196. 39:24Touch enough to get you by. If you stay long enough or you could fall beh If you stay long enough, you could fall behind.
197. 39:32Um, where should I jump back in? Give me some ideas. Um, I think a place to jump back in, Becky, is um is is the new
198. 39:41image gen within within chat GPT. Just just keep playing in chat GPT. [Music]
199. 39:59Or the other thing you could do, Becky, is pick an area. This is something I've been thinking about. What's an area in
200. 40:07your life where you have convinced yourself that you don't have talent at something or you're not good at something and go find an AI that's
201. 40:15really good at that and start playing with it because I think that I think that we'll increasingly [Music]
202. 40:31um we'll increasingly start to discover that we can be good at anything we want to be good at, but you have to be
203. 40:41willing to like go explore outside of your comfort zone. And then and then I think if you have one or two of those
204. 40:47where you're like, I actually am really good at this and I'm really passionate about it and it was a thing you told
205. 40:52yourself you could never do. That feels really empowering to me. The new image gen is ready but on hold. Reliable
206. 41:00source. Yeah, that doesn't surprise me, Brian. They there and there there've been a couple of hints on Twitter where
207. 41:06they're like, you know, you're going to like what you see and see was capitalized. And um Sam Alman said the
208. 41:13the next version of Image Gen is really really good. He thinks we're going to like it. Um maybe the next version of Image Gen
209. 41:23can do clocks that have the actual time on them, not 10 after 10. [Music]
210. 41:42Um, a thing I'm getting kind of excited about is, and I mean I could see OpenAI being being the one to lead the way on
211. 41:51this, but there might be others, is is true multimodality where within the single model, you can do code, images, voices, sound effects,
212. 42:07music, animation, you know, video, 3D world building, 3D object building, and you know, world world
213. 42:23development. Like, we can't be that far from that, right? maybe a year, maybe two, because I think the first person
214. 42:32that cracks it, like it's [ __ ] exhausting right now to have to go to chat GBT for the the script and the
215. 42:44logic. Maybe chat GB for chat GBT for the image likely midjourney. Then you have to go to Lumalabs or Runway or some
216. 42:55other [ __ ] video tool for the video. Then you have to go to 11 Labs for sound effects. Then you have to go to Cartisia for
217. 43:06voice, right? Then you have to go edit it together in something. Like come on. I don't I don't want to have to work
218. 43:15anymore. I just want to have an idea and have it vomit out. You're describing NAN right now. No, I'm
219. 43:26not. Because NAN you still have to build your automations. I mean, yeah, I I'm describing the output of
220. 43:35NAND. And if you're a technologist, you can look at NAND and go, "Yeah, yeah, that's obvious." No, what I'm talking
221. 43:41about is you talk into the chat hole and out comes an image and then you're like, "Turn that into a movie.
222. 43:49Now give it sound effects. Now write a song about it." No, I don't want it to have lyrics. I just want it to be
223. 43:54background cinema. Cinemagraphic music. Cinematic. Cinemagraphic. I'm just making up [ __ ] words here. Oh man.
224. 44:10Christian Robbins. What's happening, dude? I didn't see that. It was you. I assume you're the same Christian Robbins from from $11
225. 44:22bill. If you have not looked up $11 bill and Christian Robbins, he is the [ __ ] man. He's been doing motion graphics in
226. 44:33um he's had a motion graphics company for what, 12 years, something like that. What's up, man? Runway's getting closer
227. 44:41these days with that. I know. So, is Luma Lumalabs right now is cool because they can do images, video, now they're
228. 44:48doing sound effects. It's like that's what I'm saying. There's and and the other thing Lumalabs has is the really
229. 44:54interesting interface where you can sort of go horizontally and deep in a in a visual story, but it's like I've got to
230. 45:02go write my story over in chat GPT, and then I've got to go over here and make my music. And it's just it's [ __ ] madness.
231. 45:10Chris, I want to talk to you about something, Christian. Um, do me a favor if if you're still on. I think you're
232. 45:15still here. Um, send me an email because I want to talk to you about a project. I've got an idea for something. Um, and I think you'd be
233. 45:23really good at at at least to bounce the idea off of. Maybe we could do something together. New Tik Tok from Source Camp [Music]
234. 45:34where I've built 100 plus GPTs and still learning how to build them. I know it's right. It's it's like the depth the
235. 45:43depth of what we have in our hands. It's it's so hard to get your head around. It was funny in the early days of the web when
236. 45:54when trying to think who did it first. There was a there was a search engine before Alta Vista but after Yahoo. So
237. 46:05Yak Yahoo came out and what Yahoo did is like we're going to categorize the internet. So they created a directory
238. 46:12and they were manually you know organizing the internet and then the first search engine came out. I forget what it
239. 46:19was and it was pretty cool. And then Alta Vista came out and Alta Vista was like, "Holy [ __ ] like they seem to like
240. 46:27be they seem to know everything on the internet, which was really quite mind-blowing." And then Google came out
241. 46:33and Google, you know, took over and wiped out everything because they they they they really got that indexing thing down cold, right?
242. 46:43And it was it was if you had ever thought about data structures and how much data it to crawl the internet every night or
243. 46:57every it was like every 24 hours they'd crawl it or every two hours. I don't know. It was some ridiculous amount of
244. 47:03data that they were ingesting every some amount of hours to keep the search engine active. And it was just
245. 47:10mind-boggling. Right. And I feel like we're in the same the same place with um with large language models right
246. 47:18there. It's it's almost impossible to comprehend how much knowledge is compressed into this this box the box.
247. 47:30You know, you've got this [ __ ] box and they just took like all of the output of humanity and they're like,
248. 47:36we're going to do this thing called embedding and we're going to create thousand dimensional latent space and
249. 47:42we're just going to jam all those numbers in here and then you can ask it [ __ ] and it knows everything we've ever done. Like I
250. 47:53I can't really comprehend what that means. But here's the thing I do know. Every time I talk to chat GPT in my little voice
251. 48:03mode, I like I have yet to stump her. She answers [ __ ] every question. Where's the fud box on my 2019
252. 48:14Ford Flex? Oh, the fuse box on your 2019 Ford Flex is under the hood on the left hand side just above the wheel. How the [ __ ] does it know
253. 48:22that? Because it does. Because that's what they did. I like ideas and projects. People sleep on Luma these days. They have some nice
254. 48:32camera tools, too. Yeah, they they added that new the I forget what they call it, like the rack zoom or the crunch zoom or
255. 48:39the poppy zoom. I don't know what the [ __ ] it's called, but it goes. It's pretty cool. Oh, man. All right. I'm bad at AI,
256. 48:51but AI is very good at AI. Yeah, exactly. That's true. That's that's the other thing, Pate, is right now and and in fact, this
257. 49:02channel is kind of dedicated to us as these severely limited cognitive beings trying to get our heads around how the [ __ ] do we use these
258. 49:17tools as the tools are going like this. And the answer is we probably won't have to for long. Like we'll just be able to
259. 49:27ask for what we want and get it. Bolt cam on Luma. Yeah, exactly. That's cool. My advanced voice knew I was mad about something and
260. 49:38lowered his voice about five octaves. That's pretty funny. Oh man. Have you tried asking harder
261. 49:53questions? Yeah. No, I haven't, Pete. And I mean, that's the that's But that's a really good point that the things I do in the
262. 50:06world. For the most part, I'm not asking hard questions. Now, I do a lot of creative stuff, which creative stuff is
263. 50:14interesting because you can ask really simple questions and and it's essentially an impossible answer because
264. 50:22it's not a quantitative answer. There's not a right or wrong, right? Creative work is all qualitative. Um, but I do a lot of that
265. 50:31sort of stuff and large language models are [ __ ] insanely good at that. Now, are there areas where large language
266. 50:38models still suck at creative? Yeah, like writing song lyrics in a sophisticated way, you've got to [ __ ]
267. 50:44work to get them to do it well. Um, and I would I would assume that there are similar things if you're doing, you
268. 50:52know, high-end math, high-end research, high-end science, that these things probably fall apart and you're like,
269. 50:58well, [ __ ] they're just not ready yet. But again, I don't like I don't think it's going to be that long. Dang, ran
270. 51:03out of credits before I could finish the game. Oh, that sucks. Um um let's let's jump over to I want to
271. 51:11jump over to Vigle for a second and just see if this thing works at all. All right. So, if you're still here,
272. 51:24Christian, I they just announced this today. It's another one that promises um good character movement with with voice.
273. 51:45Okay. So, so I don't know what this thing is. I think this is a background. So, I I put in an image here. Let's do someone at at night in
274. 52:00the streets of Paris. Bigle.ai. Okay, cool. And then I'm going to drop an image here. Assets. Upload. Upload a
275. 52:12character image of any size. For better results, use a full body photo with a clean background. All right. So, let's go to midjourney.
276. 52:26Oh, this was this was a very horrible experiment trying to get trying to get um midjourney to do a Chuck Close piece
277. 52:38of art. It failed miserably.
278. 52:55[Music] Oh, he might be fun. Let's do him.
279. 53:25There's a bunch of theatrical images I
280. 53:40did. Edit. Why did that not edit?
281. 53:54Wait. Editor. Ah, there we go.
282. 54:08Okay. All right, let's do creepy dude. [Music]
283. 54:29side book. Side note, Notebook LM is getting an app. That is not surprising to me at all. It's pretty
284. 54:35cool. All right, so we have a crazy ass character here. Which one do we
285. 54:51like? I think that one. Maybe. Yeah, I think that one. All right. That's [ __ ] weird. All right. I like it. So, I'll do a creative a
286. 55:01subtle upscale and a creative upscale. All right. So I may start using Gemini more because I have direct access
287. 55:09to Docs, Drive, etc. Yeah, Gemini is I would say at this point Gemini is right up there with Chat GPT, Claude,
288. 55:19um and uh like Perplexity, just any of the ones you want to use. Um it's just that with all of them, you
289. 55:30got to figure out their personality. Like some of them some of them technically execute things that you ask them to execute, but
290. 55:40you're like that's just not good or I don't like the formatting. Like I feel like that with Claude. Like I just don't like how how
291. 55:47it formats [ __ ] And like the logical part of my brain is like, "Well, that shouldn't [ __ ] matter. It's just data. You're
292. 55:55just making words. What the [ __ ] does it matter?" It matters. It matters. So anyway, so just you know, but yeah, there's it does all
293. 56:05sorts of stuff. It can generate images. It can edit images. One of the advantages of Gemini's native image
294. 56:12editing thing where you just say like you create this character and you say, "Give him a different hat." Gemini will
295. 56:18just give him a different hat. Chat GPT's new image gen will redraw the whole image. So it can be really [ __ ]
296. 56:25frustrating. But like those little subtleties you only learn by jumping in there and doing it. He looks a little
297. 56:33plastic there. He looks a little [ __ ] up there. Uh let's see. And then what else? How are these? I think these look more
298. 56:47realistic. So we're going to go with we're going to go with this one. All right. So I'm going to download that
299. 56:53image. Then we're going to go back to Vigle. We're going to upload an image. Should we edit that? I think we
300. 57:04should. Let's lighten it up a bit so we can see a bit [Music] more. Tools. Adjust the
301. 57:18color. Let's go. Exposure. There we go. Fantastic. Bob, tell him what he's won. Oh, he hasn't won anything. This isn't a game
302. 57:28show. Why do I work here? What is my [Music]
303. 57:42job? Look at all the granular control I've got right there in the preview tool. Save. Okay, so we can at least see that
304. 57:52a little better. That was an interesting thing. I I read on uh on on the Twitter not that long ago. Someone was bitching
305. 58:02about I'm gonna I'm gonna get rid of this. How do I Do I get rid of this? Upload image or video? I don't get it.
306. 58:24Um, okay. It's still the right one. Why was that not showing up here? There it is. Oh, and then we've got to upload
307. 58:40audio. So, we got to go make audio. So, let's go to um we'll go to 11 labs. All right, hang on a
308. 58:57sec. Um, there's my new Embrace the Jank cover. Yay. I know I'm not talking much uh and I'm not telling you what I'm doing, but
309. 59:13sorry. I I will at some point. I've got I've got three. So, here's what I'm trying to do. I'm trying to get an
310. 59:20image, get a voice, and marry those two together. I could also do a song. And maybe we should do a song.
311. 59:29Let's do a song. Okay, hang on. Hold, please. Let me stop presenting there. I'm going to present. I'm going to
312. 59:36present. Oh, wait. I can't present yet. This is chat add. They don't call it chat add for nothing. Well, actually, well, I guess
313. 59:50this is very add- like, but um I know what I want to do. And it's just in my head it's a simple thing. And
314. 1:00:00then as I'm doing it, I'm realizing, oh, there's 16 [ __ ] tools here we need to use. Christian Robbins knows this game
315. 1:00:07well. He's been he's been doing this for years now. He understands the nightmare that that bringing all these tools together is.
316. 1:00:16Um, okay. So, we're going to create a song. Oh, we're going to go to chat GPT. Hang on a second. I want to look up Griswald.
317. 1:00:34Griswald. Griswald the Griner. Backstory. Okay. I'm going to say, "Write me some twisted lyrics about
318. 1:00:49Grizzwald scaring kids with stories." Not sharing. Oh, I know. That's all right. Hang on. I know. I know. I just just calm down,
319. 1:01:05people. No, I didn't. I didn't share. Come close. Come near. Don't look behind. Griswald knows what's on your
320. 1:01:12mind. He's seen the things beneath your bed. Okay. I'm just going to say you want a full length version. Um, I'm going to tell Chat GPT.
321. 1:01:22Um, this feels very surface and cliche and like you didn't think about it. I want to win a Grammy. You
322. 1:01:41ninny tick tock pin. Wouldn't surprise me if you turned this song into a musical. Shut up. I've already done that
323. 1:01:48once. 16 tools is what makes it fun. Embrace the jank. Yeah, this is totally embrace the jank time. They gather like
324. 1:01:56moth at the edge of his boots with licorice breath and balloon animal roots. This is better. His voice it's
325. 1:02:02match strike inside of a jar. Okay. Oh, twisted lullabi slow theatrical with the weight of a funeral carousel. Um, love this angle
326. 1:02:17and song description, but I want it from his point of
327. 1:02:32view. Yes. Now we're rolling into something real. A haunted show tune with soul. Let's step fully into Griswald's
328. 1:02:41shoes from G Grisswald's point of view. Dark cabaret, minor key, 68 swing with toy piano, bode saw, and an accordion
329. 1:02:49that wheezes like it remembers things it shouldn't. I should never shout boo. That's for am amateur. I never shout
330. 1:02:57boo. That's for amateur sport. I prefer tales with longer retort. A whisper, a pause, a welltimed aside, and poof. The
331. 1:03:06brave runa brave run brave ones go running to hide. They say on the cloud, but okay, this is great. Okay, so we're
332. 1:03:15going to say we're going to go grab this and an accordion that wheezes. Okay, that's that's probably just about
333. 1:03:22enough. So, we're going to go to suno and we're going to put this in song style style tags. more options. All right, there we
334. 1:03:39go. Okay, and then we're going to go grab the lyrics. Copy. We're going to come in here. We're going to plop the lyrics in.
335. 1:03:55We're going to call this song Where's the title? Exclude options. I can't. Oh, up
336. 1:04:12here. Grisswald's world. Griswald's kids. No. Griswald's.
337. 1:04:28Griswald's Griswald's tail tail. I don't know. I don't [ __ ] know, people. Okay, so I'm going to create these. Let
338. 1:04:39me go change my sharing so you can hear this. [Music]
339. 1:05:00Right. [Music] All right. We got a dark cabaret with toy piano.
340. 1:05:14Um, let's see. Um, I'm going to say, what's another word for grally? Scratchy.
341. 1:05:32Scratchy male with smokers rasp. Dark Cabaret. All right, let's create
342. 1:05:47those. Oh, Griswald's Grim. Yes. Yes, correct. That is the correct title. By the way, if you're new here, I do these creative exercises and
343. 1:06:00I fancy myself a creative. I'm like, I can I'm good with words. Some people have a way with words, other people not have
344. 1:06:11way. Old Steve Martin joke. Thank you. Um, shout out if you if if if there's an obvious good answer to the thing where
345. 1:06:20my brain is just [ __ ] shut off. Grizz Griswald's Grim is absolutely what this song's about. Okay. So, did we make any of these yet?
346. 1:06:31No. Griswald's Grim. Boom. All right. We'll go listen to a couple of these other ones. It seemed to kind of get what we were looking for
347. 1:06:41there. Griswald's black bar. There you go. That's for amateur sport. Yeah, these all sound too nice. Like we need this guy to sound [ __ ]
348. 1:07:04terrifying. I never shout boo. That's for No. [Music] No. All right, let's do Griswald's [Music] [Applause]
349. 1:07:19[Music] Grim. I never shout boo. That's for amateur sport. I prefer tales with a longer retort. A whisper, a pause, a
350. 1:07:35welltimed aside, and poof, the brave ones go running to hide. They say I'm a clown, but I just [Music]
351. 1:07:49That's pretty close. [Music] Nope. Let's see. Uh, let's do older
352. 1:08:06male. Older male with smoker's rasp. Okay, let's see. This This one here is probably usable. It still sounds a little clean. Like he
353. 1:08:21sounds like he's about to drop [ __ ] dead is what we want. I never shout. No, I never shout.
354. 1:08:38[Music] Boo for amateur sport. I prefer. These all sound like young anenus like you know like some SNL character is going
355. 1:08:55to whisper. This is not a Disney film. [Music] I never shout. That's for This is honest
356. 1:09:09to God. This is a Disney sport. All right, [ __ ] it. We're going to go back to this one. What's that, Vicki? A
357. 1:09:18grally, menacing vocal tone. Okay, we'll try that. A grally, menacing male
358. 1:09:34vocal with a smoker's rasp. They were not they never thought to train it to sound like crap. but also like minor key um 68
359. 1:09:51swing and an accordion that wheezes. Um, I'm going to say, um, the beat is
360. 1:10:09disjointed, twisted, and darken. Gank gank gank.
361. 1:10:25Okay, let's create a few more of those. We'll do three more. All right, Vicki, you're on the hook here. If this isn't
362. 1:10:38good, Perry Kik Jr. try grainy rasp. Well, you know what I mean. Okay, let's see. So, one, two, three, four, five, six. Here we go.
363. 1:10:55I never amateur sport. Okay, that's not bad. [Applause] [Music]
364. 1:11:09Jesus Lord, these are bad. Well, by the way, if you're new here, you might be amazed that this thing's making songs like this. It
365. 1:11:19is. Here's one of the things that people who are not actively playing with AI miss this. What we're doing right now is creative development,
366. 1:11:33right? I'm serving as a creative director right now. I have in my mind, we've got we've got a picture of this crazy ass [ __ ] dude,
367. 1:11:43which I can't switch screens and show you. We have this crazy ass carney, right? And I've got some idea of
368. 1:11:51what I want, but I don't quite know what it is yet. And and a big part of working with AI, even though it makes things fast, it rarely makes
369. 1:12:01things that match what you want. And so the art of AI is just like swinging back and forth like, you know, tacking, you
370. 1:12:10know, in a sailboat back and forth and back and forth. You kind of know where you're going, but n it's too much that
371. 1:12:16way. That's too much that way. So anyway, um this is kind of how it goes. That's for amateur. I never shout
372. 1:12:32boo. You know what I'm going to do? Final chorus. I never shout quotes capital B O exclamation
373. 1:12:50point and then I'm going to put that a capital T. All right. And I'll create two more and just see if we can get that to hit the word
374. 1:12:59[Music] better. for amateur sport. [Music] No, I never shout. Boo. That's for
375. 1:13:17[Music] amateur. I never shout boo. That's for amateur sport.
376. 1:13:33Not bad. [Music] That's for amateur sport. I I never shout. That's for
377. 1:13:50amateur sport. I prefer tails with a [Music] longer. I never shout boo. That's for amateur you
378. 1:14:06sport. I prefer tales with a longer close. We got to change the music. A grally menacing. Oh, I spelled that
379. 1:14:17wrong. Menacing male vocal. The the rasque. I didn't put in the whole thing. Tik Tok pin. I wish I could audition for this
380. 1:14:27one. Let's see. Like a clown dragging its laugh through broken glass. Let's see. Vocal like a
381. 1:14:43clown dragging it laugh. Oh, I'm out of I'm out of characters.
382. 1:14:58Grally menacing male vocal clown dragging its laugh through broken
383. 1:15:13glass. Okay. Dark cabaret minor key toy piano bode saw accordion that wheezes disjointed twisted and dark.
384. 1:15:23Create. Create. Okay, these last This should be it. This should be it, people. You can just type in Tom Waits
385. 1:15:34and see what happens. Yeah, that's not bad. Let's see. Um, a grally menacing vocal. Let me just say um like Tom Waits through broken glass.
386. 1:15:49We'll see if it it might reject that one. Did it? Yeah, I think it did. It rejected it. Couldn't generate that. You
387. 1:16:03can't put in an artist's name. I never shout boo. That's for amateur
388. 1:16:18sport. I never shout boo. That's for amateur sport. I never shout boo. That's for amateur
389. 1:16:32sport. [Music] I think we had it way back when. [Music]
390. 1:16:48It was one of these early ones. Hang on. Is it this one? A [Music] whisper. That's part of the
391. 1:17:06game. The brave goveout. [Music] All right. This is the one we're going to do. This is the one we're going to
392. 1:17:22do. It ain't perfect, people. Just perfect enough. That's one of the songs from Sydney, my new Broadway musical. Not yet on [Laughter]
393. 1:17:34Broadway. Okay. Um, let me change my screens again so y'all can see. I want y'all to see. Gank. Yeah. Oh, yeah. Tomorrow at 400 PM
394. 1:17:52Mountain time is the uh AI readiness project podcast with myself and Anne Murphy. Uh you can find it at the AI
395. 1:17:59salon. You can find it on my LinkedIn. You can find it at sheleadsai.ai. Um we downloaded that. We're going to
396. 1:18:09vigle. Now we're going to upload audio. Griswald's Grim. Here we go. [Applause] [Music]
397. 1:18:28Wait, who had archetypal note? Never thought to train it to sound like crap. Good advice for Sunno. Where's the good
398. 1:18:36advice for? Oh, down here. Click the like on it. Make it easier to find later. Oh, yeah. I know. That's
399. 1:18:43That's one of those add oversightes that every time I I'm like, "Oh, click like and then you can find it later." And
400. 1:18:50then I just forget to click like. It's But it's a good call. Audio must be under 60 seconds. Fine. We can We can fix
401. 1:19:05that. Oh, no. That wasn't what I wanted to do. God damn it. Open with QuickTime
402. 1:19:20player and let's see. [Music]
403. 1:19:39[Music] kid. Okay. [Music]
404. 1:19:59Okay. Clip clip that [ __ ] Wait. could not be split because some clips could not be loaded. What do you mean it can't be loaded?
405. 1:20:09We're [ __ ] playing it. Oh, come on. God damn it. I just every [ __ ] time I want to
406. 1:20:24do something on this channel and it's just like just [ __ ] doesn't [ __ ] work. My stories don't end, kid. They make you pretend.
407. 1:20:40Fine. Split clip. Please work. Yay. Okay. Bitter. You betcha. Export as audio only.
408. 1:20:56Um, Grizzwald's Grim. Uh uh uh short. Oh, and um you can check with Christian Robbins on this. I've talked
409. 1:21:11to him about this. My best practice for file management. Save everything to your desktop. Just loose with a completely
410. 1:21:21nondescript name so you can't find it later. That's the really the best way if you're going to do especially if you're
411. 1:21:26doing big complicated projects and you got to make a lot of different assets from a lot of different places. Just
412. 1:21:32save them all the desktop because they'll all be there. It's really easy for you to find. If you think I'm being serious,
413. 1:21:41then this is not the channel for you. Confirm.
414. 1:21:57I don't want this customized thing here. How do I remove this from image? Okay, there we go. All right,
415. 1:22:14generate. Generation created successfully. Just blame the blame the producer when [ __ ] breaks. Well, if if it weren't if
416. 1:22:26you worked for Apple, I'd [ __ ] blame you. It's like, come on, people. Let's get your [ __ ] together with these softwares. All
417. 1:22:35right, I'm going to have this first line stuck in my head for the next week. You're welcome. All right, let's go.
418. 1:22:41Let's go. Let's go listen to our little song here while that thing's rendering. [Applause] [Music] [Applause]
419. 1:22:54I never shout boo. That's for amateur sport. I prefer tales with a longer retort. A whisper, a pause, a welltimed
420. 1:23:05aside, and poof, the brave ones go running to hide. They say I'm a clown, but I ain't in disguise. is I just dress
421. 1:23:15the way I remember your cries. One button's for secrets, the other for shame and the smile. Well, darling,
422. 1:23:24that's part of the game. I don't chase. Wait, where else? Which which other one is good at? Hedra is Let's go do Let's
423. 1:23:37go do the same thing in Hedra. Um. Oh, did it do it already? Come on.
424. 1:24:07Guessing you know I'm missing a kind. Hang on. Sorry about that. Tab, tab, tab. Oh, and I got to change my audio, right? God damn [Laughter]
425. 1:24:20it. It's so [ __ ] exhausting. It's not bad. It's a little janky. I want to try this in Hedra, too, because I think
426. 1:24:26Hedra is better. But we'll see. This is not bad. [Applause] I never shout boo. That's for amateur
427. 1:24:41sport. I prefer tails with a longer retort. A whisper, a pause, a welltimed aside, and poof, the brave ones go
428. 1:24:52running to hide. They say I'm a clown, but I ain't in disguise. I just dress the way I remember your cries. One
429. 1:25:02button's for secrets, the other for shame and the smile. Well, darling, that's part of the game. I don't chase,
430. 1:25:11no need to begin. I curl up beside you and crawl right. All right, we're going to do some AB testing here. We're going to do this in
431. 1:25:20Hedra. And Christian Robbins, do you know any if you're still here, is are there That's creepy as [ __ ] Yeah, it's
432. 1:25:26totally creepy. Let's do it in Hedra. Are there anything other than Hedra and uh Vigle now that do this uh voice
433. 1:25:35stuff? Good. Where's Hedra? I guess I can just do a new tab. Why not do a new tab? Because you're not sharing the new tab.
434. 1:25:46Yeah, I know. I'm gonna go do that now. I really like it. Okay. Present. We're gonna do this. We're gonna do
435. 1:25:55this. We're gonna do this. Okay. Yeah, Kyle, just looking at that one that we just watched, it it reminds me a lot of
436. 1:26:02Will Smith eating spaghetti a year ago. Yeah. Yeah. You you said it's 2025 good. Yeah, it totally it totally is that. The
437. 1:26:10thing that I've been impressed with Hedra is it's got the the least amount of jankiness of anything so far. It's
438. 1:26:17still a little janky. You know what these things all do, which is weird, is when there are pauses, the characters
439. 1:26:24like look up and to the left. They're like they they just look up that way and it's just freaky and weird. Um Oops, I just lost my
440. 1:26:35light. Can't lose my [Music] light. That's kind of a catchy song.
441. 1:26:50It's It's a lot. Uh, it's way too Disney, I will say. Okay, we're going to upload an image, which is going to be found here in
442. 1:27:02downloads. We're going to upload our audio script, which is going to be found on the desktop, which is where we save
443. 1:27:08everything because we have good file management skills. Oh, this is nice. God damn it. Okay, so now we get to go to um free M4A to MP3
444. 1:27:27converter web convertio. I want free convert. Con
445. 1:27:43convertio. Which one of these doesn't run me down a stupid [ __ ] sales funnel? Oh, wait. I wonder if I can do a
446. 1:27:55threem minute song in Hedra. That would be a good thing, wouldn't it? Let's try to add the three minute song.
447. 1:28:11All right, it says yes. Um, clown looks.
448. 1:28:30Um, pervertedly happy as he sings. Shrank dank. You could also try
449. 1:28:46downloading the stems in Suno so it doesn't react to the music. Right click. Oh yeah, that's not a bad idea.
450. 1:28:54Oh [ __ ] I just dropped my deck of cards all over the floor. it. You know, when you use a deck of cards as your add
451. 1:29:05pacifier, it's just not good. But let's go to Sunno. We'll go to this song and we will grab stems. That's in create, right?
452. 1:29:18Create. Yeah, get stems. If you haven't done this before, you can get stems. So what what Sununo will do is it will split your vocals
453. 1:29:29from your music. So we can just do the the thing to uh just to the vocals. But but Hedra seems to be pretty good at that
454. 1:29:41actually. They've sort of figured it out. $20 million studio, no black bar. Okay. [Music] I wish the song weren't quite so Disney,
455. 1:29:55but you know, it's like it's like Disney menacing and I want it to be like really [ __ ] menacing. 52 card
456. 1:30:05pickup chat GPD converts MP4 to MP3. Oh, it does. That's cool. Real estate in Columbus, Ohio. Isn't it funny that Elon Musk
457. 1:30:21Oh, he does the same pose. He looks up and to the left. Maybe he's a All right, so let's let me change my tabs here.
458. 1:30:29Maybe refusion. Yeah, actually, you know what? Refusion's not a bad idea. Let me go let me grab the these.
459. 1:30:41Copy. Do I have Refusion up here? I thought I did. Nope. New tab. refusion. Uh
460. 1:30:56shank. Am I logged in? Lyrics. We don't want an
461. 1:31:11instrumental. Okay, there's that. prompt. Let's go back to Sunno. The nice thing about this is you
462. 1:31:26can put in longer prompts in Refusion. All right.
463. 1:31:45Get shink. Oh, AI
464. 1:32:00Bob. Bob, tell him what he's [Laughter] won. Um, the uh, refusion I think is a little bit better than Suno. Certainly a
465. 1:32:12prompt prompt adherence coherence. All right, let me go change my tab so you can hear.
466. 1:32:29[Music] Jank [Music]
467. 1:32:50dank. Sit here close. [Music]
468. 1:33:11These are too peppy. Prompt. Let's say a slow,
469. 1:33:35menacing, dark cabaret, minor keys with both saw and a slow, menacing, disjointed twisted and dark
470. 1:33:52cabaret. A slow and put male voice. Whoops. Let's see.
471. 1:34:08Um like a 50year smoker who never sleeps.
472. 1:34:27There is a strength in his voice despite its frailty. Look at that fantastic [ __ ] writing. Could have written a [ __ ]
473. 1:34:43screenplay. All right. [Music]
474. 1:34:57Let's go check back on Hedra. What's it doing? Processing. Clown looks pervertedly happy as he sings. All right, we'll see.
475. 1:35:06We'll see how ugly that is. All right. Deathbed vocals. That's not bad. Let's see. male voice like a 50-year-old smoker who never
476. 1:35:21sleeps. Deathbed vocals. Generate two
477. 1:35:37more. Hi AI fam. Just been conversing with my BFF chat GPT. I know pretty amazing. Played through a phonograph is
478. 1:35:44a good idea, too. All right, hang on. Um, played through a phonog graph. That's good. Like I said, if I'm doing it
479. 1:35:59wrong, don't just secretly make fun of me. Make fun of me in public. And even better, have a good idea like Christian just did. And I know
480. 1:36:11you said that before. Part of the joy here is I'll ignore most of the good ideas. Okay, here we go. Let's
481. 1:36:29see. Okay, this is starting out promising. [Music] It's a woman. [Music] [Applause] [Music]
482. 1:36:50Sawustle. God damn it. It's so [Music]
483. 1:37:05frustrating. [Music] Uh they why can Can they not [ __ ] train these things to say like a pull down menu? Man, woman, I know we're not
484. 1:37:18allowed to gender anymore, but honest to God, like the male voice like a 50-year-old smoker who never sleeps. Deathbed vocals.
485. 1:37:29There's a strength in his voice despite his frail its frailty played through a photograph. Why are we getting an women? God damn
486. 1:37:40it. Maybe cabaret's throwing it off. Oh, that's a good point. What's another word for cabaret? Um um carnival. A slow, menacing, disjointed,
487. 1:37:54and dark carnival. Diddy. Diddy. TTY. Right. All right. We'll try that
488. 1:38:09one. About Jack and Diane. A little diddy about Jack and [Laughter] D. Okay. Let's
489. 1:38:24[Music] see. No.
490. 1:38:41Sit down. [Music] What? That's okay. It just [ __ ] up. Okay. I
491. 1:38:55was like, did I put in the wrong lyrics?
492. 1:39:10[Music] alone. I don't know. I don't know what the [ __ ] going on. It also rewrote all my lyrics. I'm sure there's a place to
493. 1:39:26[ __ ] fix that. Add to instrumental. Roll the dice. H. Do I need to be in this mode? Let me generate
494. 1:39:41it. Let's go back to Hedra. Oh, you can't see this. Still processing. Vaudeville's not bad, but I can't get it. Let's see if I
495. 1:39:53can get it to do the the right lyrics. [Music] I never shout boo. That's for amateur sport. That's not bad.
496. 1:40:09I never shout boo. That's for amateur sport. I prefer tails. All right. Whatever. All right. I'm growing I am growing weary of
497. 1:40:23this. Let's go back to tabs, tabs, tabs. Yep. Let's go back to uh let's go watch the um the vigle one again.
498. 1:40:43[Music] Shank
499. 1:40:59shank shanking shank. Yes, I am that annoying. I know
500. 1:41:13I'm annoying. It's just it just comes with the [ __ ] deal. It's just part of the deal. All [Music] [Applause]
501. 1:41:25right. It's actually interesting. What these guys are doing is they've masked the main character and then it's just in front of this
502. 1:41:37background. So, they basically just make the background the background. Not to throw you down another rabbit hole, but
503. 1:41:54Uh, yeah, I could do that, but I I don't Yeah. Yeah. I don't want to I don't want to sample my voice. That's too much of a rabbit hole.
504. 1:42:05[Applause] I never shout boo. That's for amateur sport. I prefer tales with a longer retort. A whisper, a pause, a welltimed
505. 1:42:19aside, and poof. The brave ones go running to hide. They say I'm a clown, but I ain't in disguise. I just dress
506. 1:42:29the way I remember your cries. One button's for secrets, the other for shame, and the smile. Well, darling,
507. 1:42:38that's part of the game. I don't chase. No need to begin. I curl up beside you and crawl right in a lullaby
508. 1:42:50twisted. My stories don't end, kid. They make you [Music] pretend. All right. It's not bad. It's not bad, but it's I'm with Brandon. It's
509. 1:43:042025 good. Oh man. All right, Hedra, please be there. All right, this is taking for [ __ ] ever. Um, what time is it? It's
510. 1:43:17getting to be late. All right, so here's what we're going to do. Well, this is what I'm dreaming about
511. 1:43:23tonight. Um, so Christian, good to have you here. I look forward to catching up. So, let's definitely talk. I want to
512. 1:43:30tell you about this project. I'm just going to say I've never had Sunno sing entirely different lyrics to what I gave it. No, I know
513. 1:43:38that was Refusion. I've never had Refusion do that to me before. I think I must have had some weird cash thing
514. 1:43:44going on. Probably because I've got 80 tabs open. Um and uh so so anyway, so so let me let me get out of here. I look for whichever
515. 1:43:55I played the most and like combo and Suna. Oh, that's not a bad idea. Silver Fox, that's really good. Um okay, let
516. 1:44:05me let me entertain you. Let me entertain you. I never say boo. That's for amateur sport. It's pretty good. It's pretty
517. 1:44:19good. All right, one more. One more play. One more play. Come on. It's going to be good. [Music]
518. 1:44:37I never shout. That's for amateur sport. I prefer tales with a longer retort. A whisper, a pause, a welltimed aside, and
519. 1:44:50poof. The brave ones go running to hide. They say I'm a clown, but I ain't in disguise. I just dress the way I
520. 1:44:59remember your cries. One button for you know who this this looks like is that um Puddles the Clown. Remember when he was
521. 1:45:10on Idol or America's Got Talent or whatever the [ __ ] show? I don't think it was Idol. It was one of those one of
522. 1:45:16those talent shows. That's that's who this looks like. It's got a claimation feel. Juggle image in Irregulars. Okay,
523. 1:45:24let's go look at irregulars quick. Look at that new AI readiness cycle. Isn't that cool? We did that last night.
524. 1:45:38That used to be linear and we were like, "Hey, it's not really linear. It's a cycle." Well, no. I said that. I'm like,
525. 1:45:46"Here's this thing, but it's not really linear." And I was like, "Wait a minute. Couldn't we have one of these things fix
526. 1:45:52that for us?" And the answer is yes. This was a midjourney fix. Um, going down to AI learning lab, juggling all the different tools.
527. 1:46:08That is totally what it [ __ ] feels like right now. What you have to have is one one like coming at him, smacking him in the
528. 1:46:22forehead. Oh man. Oh, that's pretty. Oh, it's a sununo song. Nice. All right, one more check. No, still
529. 1:46:34processing. All right, I'm getting the [ __ ] out of here. It's getting late. I'm getting crispy. Um, okay. So,
530. 1:46:42tomorrow I'm speaking in the afternoon, then I got a dinner at night. So, it's possible I won't be here tomorrow or it
531. 1:46:49might be like 9 or 9:30. If it is, I'll probably just do it late because I'll probably be exhausted. Um, and that's Wednesday. Thursday,
532. 1:46:59we'll be back to normal 8:00 PM time. All right. Trying to think what else. Um, at work, I'm doing a lot of uh
533. 1:47:09persona. Like the stuff we're doing here with this thing with music, I'm doing it with scripts with like doctors and
534. 1:47:15patients telling their stories. Um, and it's going pretty well. So, I'll I'm gonna one of the things I'm going to do is
535. 1:47:25um I'm going to capture my workflow for that and I'll share that with you guys. But, um all right, cool. Um peace out
536. 1:47:34YouTube X and LinkedIn. Hope you're doing well. Christian, good to see you. Talk soon. And everyone else, great to
537. 1:47:41see you. Thanks for hanging out and I'll see you tomorrow or