8/17/2026 - Redesigning a Brand with AI: From ChatGPT Mood Boards to Lovable

Transcript

1. 0:00Hey, hey, hey. [music] [music]
2. 0:18[music] [music] Good evening, good people. Welcome to
3. 0:35the AI learning lab. Uh we are coming to you live from Denver, Colorado. Got producer Brandon having Wi-Fi issues,
4. 0:44but we're going we're going good. I think my Wi-Fi is fine here. Let me try something on. >> All right, beautiful. And uh welcome
5. 0:54welcome to the AI salon. Welcome to the AI learning lab. I think you can all see me. Uh if you want to um chat, ask me
6. 1:03questions. The uh you pop out the chat window. There's a link right up here. You just pop that out on on desktop on
7. 1:10mobile. The chat is below the uh the live stream. You should be able to see it. And we fixed it so that you can
8. 1:16actually chat now. So that should be good to go. Um beautiful producer Brandon, hello from the live stream. alt
9. 1:24producer Brandon. I think he's having some whiffy issues. The wiffy the whiffy wiffy is acting up on him. Um I hope
10. 1:33it's not a big storm or something like that. Let's see. It is Monday, uh August 17th. There's a lot happening. It looks like I
11. 1:47saw some rumors today that we might be getting the new Astra model from from OpenAI this week. So Astra is apparently the 10 trillion
12. 2:00parameter model. I think Soul was three trillion parameters. I think I think that's what it was. And Astra is 10. So, so it's on the on the
13. 2:11level of Mythos from Anthropic. Apparently, Anthropic now has Mythos 2 has completed training, but they're not going to release it.
14. 2:21They're going to do medical research on their own and not let anyone else one else play with it. I don't know.
15. 2:28Whatever. Anthropic is I'm I'm done with anthropic. [laughter] Greg Walters, greetings. Greetings. Greetings.
16. 2:38Aloha from Kawaii. So, I I hope the uh that you're safe from the storm, Linda. Uh I I asked Chat GPT and Chat GPT said
17. 2:47that that Kawaii did not get hit as hard. So, I hope that's the case. I hope hope you're safe and everyone's safe. Um
18. 2:57yeah. So, so we might have a big new frontier model this week. trying to talk. I'm trying to think about how I feel about
19. 3:10the the model fetish fetishization that people like me have been guilty of doing. And I I will admit to that that
20. 3:21that I get excited about the new and shiny thing. Um the new and shiny models. Yes, we're good. We've only lost a few shingles and
21. 3:30a screen at our house. Beautiful. Love that. Um, the thing about new models is that they're big and they're expensive.
22. 3:46And then what happens is it used to take a year a big expensive model would come out and then a year later a better one
23. 3:53would come out and then after about a year of that it dropped to 6 months and this year it's dropped to three months
24. 3:59and now it looks like I mean GPT 5.6 Soul just came out a month and a half ago and this week we might have the next big model. So so the
25. 4:13models are accelerating. The other thing is with these agentic platforms like chat GPT work and codeex and claude code and gro build and uh
26. 4:29Gemini's anti-gravity which they announced that and they it's still out there but they don't really talk about
27. 4:35it anymore. So I assume they're going to do a consumer a more consumer version of that. We've now got Grockbot out which I
28. 4:42think is based on on the Cursor platform. So I think this is technology that cursor been putting together and uh
29. 4:51SpaceX AI bought Cursor for whatever it was some ridiculous amount of billions of dollars. Um, I think the real play for us people is
30. 5:02we just need to start a company that does [ __ ] and then at some point let one of the frontier model companies go, "We
31. 5:10want one of those and just give us $40 billion." I mean, I would take 20 billion, you know? I'm not greedy. Not
32. 5:24>> [laughter] >> So anyway, so we [clears throat] might have a new big model out this week and it really won't matter
33. 5:36except it'll change everything. So So it will matter. [laughter] Um, but I did want to show you something.
34. 5:51California Bill 412 passed. Oh, it died. Okay. That committee. All right. Interesting. That's a good thing. That's a good
35. 6:03thing. Yeah, they were looking to get pretty restrictive with AI stuff in California. Looks like that bill died. Um,
36. 6:10we're in a really fascinating time right now with the uh with the regulation stuff with the the antics that anthropic
37. 6:19and open AI are doing right now with AI is super dangerous. It's awesome and it's going to change the world, but it's
38. 6:24super dangerous. You can't have it. It It's making me a bit crazy right now, but it's it's it's very much stirred up
39. 6:32the waters with all this regulation stuff. So, I don't I don't know where any of that stuff's going to end up, but
40. 6:36I I I think in the end it'll all kind of normalize and and you know, we'll we'll see. We will see. We will see. Um,
41. 6:47so, one of the things that I've been working on is I'm working with uh Andy Scarantino, who runs the the AI salon,
42. 6:56on developing a business offering for the Great Repurpose, and it's coming along really well. But um let me show you something here.
43. 7:07Hello Fran. Mr. It is in the house. I think I'll go entire screen.
44. 7:23See if my machine will hold this for longer than a minute.
45. 7:37Okay. Okay. So, if you haven't been to the great repurpose.com, you can go there now and you can play. And the idea here is that
46. 7:53that we're all about to be confronted with either jobs disappearing or jobs transforming and and sometimes transforming radically
47. 8:05like we've seen it with programmers where programmers no longer program. They manage bots that do the programming. And so they're now
48. 8:15managers. And if you were really good at programming and you loved programming, what do you do if now you're not
49. 8:26program? That's not your job anymore. So even if you keep your job, things are about to trans transform. And we live in
50. 8:33a country where we're we're deeply weep identify with our work, right? We we ask each other what do you do? Not who are
51. 8:43you or what's interesting about you but what do you do? So so largely thanks to the industrial revolution that basically
52. 8:50said time equals money. You give us your time to do the tasks and those tasks literally become your value. So you get
53. 8:59good at tasks and we will give you money. And so over generations we've come to identify those tasks with our value. It's not the case,
54. 9:12but it was. And AI is about to change that. So, so we put together this site, and I'm generally happy with it, right?
55. 9:21I I I um I made this music video, which I think you've all seen, unless you're new here. [music] [music]
56. 9:35Clockwise [music] stuck on like a bad same chair, same [music] light [singing] inside. Every praise, every pin tight to the
57. 9:52shirt on my chest. [singing] If I'm not >> So, so that we made here on a on a live and and the look and feel of that is
58. 10:03kind of cool. It's kind of moody. It kind of captures what's going on. And the whole site kind of has this 1970s
59. 10:16olive green kind of law firm feel to it. Um, and it's it feels more dystopian than hopeful. The whole point
60. 10:27of the great repurpose is yeah, things are about to suck and this is a perfect opportunity to repurpose ourselves and
61. 10:35figure out what's our value beyond our jobs and and you know what are things that inspire us and what are things that
62. 10:42we value and right like you know what are things we want to do in the world and then learn enough AI to go amplify
63. 10:49that version of yourself and like you know completely repurpose yourself. So that's that's the whole idea. So it's got this hopeful
64. 10:56optimistic thing and yet the feel of the site is kind of like, well, this is gonna suck, right? So, there's something I like
65. 11:04about the site. It's authentic. It feels real. Um, but it but it also just, you know, there's something missing there.
66. 11:12So, I thought what might be cool is just show you some some work I did today on how do we go from something that works
67. 11:21and has like a a good DNA to it, but but reinvent it. All right, let's see. I guess the trap was thinking we were what what we did.
68. 11:34That's exactly it, right? That's exactly it, Greg. Is Listen, let's say you grew up as a kid and whatever. I don't like you liked
69. 11:50accounting. You your your mentor was an accountant and you fell in love with an accounting and then you went to school
70. 11:56for accounting and then you went and worked for a big accounting firm and then you said by jingus I'm gonna I'm
71. 12:03gonna you know break out on my own and do accounting and you loved accounting and when people said you know what do
72. 12:09you do? I'm an accountant, right? And so in a world where the tasks of being an accountant required human beings, like it's a like
73. 12:22in some ways it's a very natural thing to tie your identity to that to that job. And it's not just that you could add numbers
74. 12:35and do spreadsheets, right? It's that you were passionate about numbers and and you cared about people's businesses
75. 12:43and you wanted to make sure that they could get all of the tax breaks that they could. Right? It's that your real
76. 12:50value wasn't in the tasks you did, but it was how you showed up and how you showed up is is a is a result of who you are as a human being.
77. 13:01And so what AI is doing is it's taking all that tasky stuff which it's really easy to hide besi behind and say that's
78. 13:07who I am and it's it's it's making these other parts how you show up and who you are more important because AI is going to do the tasks
79. 13:18right. So I might still work with an accountant but not because they can do accounting. I'm going to work with an accountant now
80. 13:30because they believe in my business and they're not a jerk and they're fun to talk to or they're at least they at
81. 13:36least make numbers understandable, right? They don't have to be interesting, but I at least need to understand them. And so I think I think
82. 13:43even how we choose who we're going to work with is about to shift. And so the great repurpose is a part of that. So, so I went to
83. 13:53our good friend and design consultant chat GPT. By the way, I'm also a trained medical professional. I went to GPT University.
84. 14:03So, if you have any questions on anything medical, ask me and I will give you inaccurate information. [laughter]
85. 14:13It wasn't wrong. It just was. It's the Apple commercial and Pink Floyd and Tank Mine in the square. Yeah.
86. 14:23Yep. Cam Ken's in the house. What's happening, Cam? Good to see you. Um, so I basically said, "These screenshots are from the great
87. 14:32repurpose.com, and I find the color depressing and dystopian, and TGR is supposed to be hopeful and expansive.
88. 14:39Help me in coming up with some color palettes to pull this out of the doldrums. be creative and act uh as an
89. 14:45award-winning creative director. Give me primary, secondary, and uh colors and meanings and where the colors should be used, right?
90. 14:54And so, you know, it it it of course is sick of fantic and kisses my ass as it does. Yes, the core problem is bigger than the
91. 15:04green is too dull. [laughter] The current tonal architecture is telling the wrong story right now. the near black, muddy olive, gray beige,
92. 15:12muted sage, and the heavily darkened photography reads as aftermath survival post collapse, which to be to be fair to
93. 15:20myself, that was kind of what I was going for. Like like when we first stumbled upon this, I was thinking a lot
94. 15:27about, you know, this is we're about to go through it. Um, and [clears throat] what what the the great repurpose has
95. 15:37evolved into is a is a path forward. really like Rick's video and website. Cool. I don't know which one that is.
96. 15:51Well, we're there. So, you're correct, Kyle. Yeah. Well, I know exactly right. It It is kind of, you know, there
97. 15:57dystopian, but so what I wanted to do was So, look at this. And so, basically what it did here are five
98. 16:04recommendations. So, it's acting as a creative director director here. You know, we could go with an open sky feel,
99. 16:10possibility, agency, the world just got bigger. Primary horizon blue or primary sky, open airy blue, right? And then
100. 16:18gives gave us hex codes for the colors. Um, what the colors mean. Blue equals possibility and intelligence. Yellow
101. 16:28equals illumination. Coral equals equals humanity. Green equals forward movement. And so then it gave me five of these,
102. 16:35which is what I asked for. And so, as I want to do with this kind of chat GPT content, I didn't read it. I I skimmed it,
103. 16:47but reading hex codes and reading descriptions of colors doesn't really help with understanding the feel of the design. So,
104. 16:59electric humanism sounds like it's going to be a bothersome and annoying, but that's the one I ultimately went with. big horizon.
105. 17:10So there's five different concepts. So what I said to it was, can you make it so that I can see these colors visually? In fact, make a color
106. 17:21study for each one as a mood board image. And so then what it did was this. It said, here's open sky. So there's
107. 17:30your color palette. Here's some basic layouts, some styles of photography. Here's the next chapter, the new dawn,
108. 17:43and the guy standing on the mountain overlooking the valley. My problem with that image is that you kind of lose the
109. 17:49humanity from it. It's kind of about the the vista and the openness and not about the human. Um, here's the verdant future. So, this
110. 17:59is about growth. This one still feels a little bit like the old site to me. There's something about green as a
111. 18:05primary brand color that just it just I don't know unless you're doing something in recycling or or you know things like
112. 18:13that just doesn't feel right. Electric humanism. This one to me all of a sudden was like oh that's kind of
113. 18:20interesting. Like I kind of like the treatment of these photos. I I hate the silhouette of dude with the birds flying
114. 18:27out of it but like it's like there's something about this one. Like the color scheme is kind of right. It still gave
115. 18:33us the dude standing on the mountain overlooking the valley. Um I think we've seen enough posters with that image on
116. 18:39it that any variation of it is just cliche. And then the fifth one was um big horizon. And again, this one's
117. 18:47just kind of boring and big vistas, no humanity. So I thought this one is interesting enough to explore. So I said
118. 18:56I like electric humanism a lot. these images are a bit too stock photography and I want them to be super authentic
119. 19:02and human but I still have but still have the edge you're going for here. And then it made this thing which is
120. 19:08actually kind of close, right? And it's like here's more people looking images and it's got these colors and it's not
121. 19:14quite this white background. Claude Design is good. Yep. We we worked with Claude Design today to do some of the the leave behinds.
122. 19:26Andy worked on that. Um, the images are still a bit and then so then it gave me something where I didn't like the images at all. Um,
123. 19:36and then I said I wanted it to be a sans Sarah font and then I said, you know, let's keep going. And it made this
124. 19:43really weird font which is sort of custom but I don't know, a little forced. And then there's the great repurpose.
125. 19:54And it gave us these cheesy clip art looking icons. So I told it, "Hey, here I went out to I went out to Google
126. 20:01images and I found some icons that are just not quite so clip arty and can you, you know, redesign it with those?" And so it did that.
127. 20:15So, so now I'm at a place where we've kind of got some colors and and kind of have, you know, what we need
128. 20:27in terms of like the five stages and look and feels and some interesting ideas. And it's like, well, what do we
129. 20:34want these images to actually look like? Cuz I didn't want to do it in chat GPT because it's just it's too easy for it
130. 20:41to fall into a trap and you can never get out of it. And this was starting to happen with this. So I said now I want
131. 20:48you to make some image prompts for midjourney uh in in where in which each of the this is supposed to say images in this mood
132. 20:56board with aspect ratios and descriptions of the subject and then it it tried to make images and I said no no I want midjourney prompts
133. 21:07and so so here are some midjourney prompts. It gave me I don't know 12 or something like that. And so what I did was I went to midjourney
134. 21:16and where did I start? So I started right here and I just took the first prompt and just put it in there. That's with no style whatsoever.
135. 21:28And so those are fine, but they're boring and not what I wanted. So then what I did was now that we have this
136. 21:35prompt in there, I went over to um the explore tab. If you go to the very top of the explore tab, you can go
137. 21:45to styles. I'm in midjourney right now. And I typed in I want to look up photography. I think I look I I I stumbled around in
138. 21:57this for a while, but I went photography. Then I went bright. And I found like I ended up stumbling on this thing with these bokeh
139. 22:07images with this sort of cool light bright with these cool shapes. And so I just said try style and I found another
140. 22:15one. I liked it. Say try style and try style and try style. And then so what ended up happening was I just ended up with
141. 22:27a whole bunch of images. Like this pink one's kind of cool. that, you know, are starting to look pretty interesting
142. 22:38and they're they're starting to look like that design. Where's the you know, they're starting to look like
143. 22:48like they match those colors, right? You know, you can you can start to see where images like this would look good
144. 22:56in some of the colors sitting here. And then I just did a big [ __ ] ton of these. I just went and looked and they still
145. 23:07have that kind of DNA of that original kind of dark and moody image, that green image. They're a little little peppier, right?
146. 23:17And then some of the styles just don't work at all. And so now we get into the some of these bokeh things where there's
147. 23:23there's this cool lighting on everyone. That's starting to look interesting. You know, I'm not crazy about all the
148. 23:36all the people it created, but some of them are interesting. Some of them look like they have soul. Look like they're going through some
149. 23:44[ __ ] I mean, I like the obscured faces, the blurred out faces, the lights, you know, becoming like there's something
150. 23:58more than us that's happening. [laughter] A decrepit green image of people that look like they're about to die. Exactly.
151. 24:09So, these aren't that, right? that people on their way out, right? These are not that. So then
152. 24:25I stumbled upon this this bokeh thing. So bokeh, if you don't know it, is like, see the blurry these blurry spots?
153. 24:33That's bokeh in photography. It's like out of focus out of focus lights create that those patterns. And so now we're starting to get
154. 24:42somewhere. Still still the same prompt. And so these are all just studies. And then wait a minute, this one all of a sudden those colors
155. 24:58kind of look like those colors. It's like we found this mood board or this this style that's generating things that are kind of in the
156. 25:09neighborhood. And that's that's when I thought, oh, maybe we could do images that are are direct calls to this color palette.
157. 25:19And then I started taking so this was I think the second prompt from chat GPT. And I think I took not those
158. 25:40I don't know which one I used. It's one of these. I think I did some variations on it, but I found that style, that one with those
159. 25:53the sort of pinks and blues in it was kind of cool. And so so then I took So if you go to MidJourney and I go out, let me go find those
160. 26:03images. So yeah, these are the this is the one right here I'm pretty sure that we ended up going with.
161. 26:20And so yeah, 244. So this style reference right here gives us these kind of cool colors and it's like, okay,
162. 26:29that's kind of interesting. That works. So now what I did was these are all using the same midjourney style. So
163. 26:37notice the colors, right? Again, we'll go back here and that color palette, you know, is directly in these images.
164. 26:51And so it's still got that same kind of mysterious kind of non-clipip arty like that's a really interesting image,
165. 26:59right? Yay, blue. Love that. You know, this is a guy like looking looking at his future. He's mood boarded
166. 27:08his future. What's he going to choose? Where's he going to go? His shirt's all wrinkled, right? He's out of sorts, but it's like it's alive.
167. 27:21There's some variations of that one. Some people going through it, right? Like what's next? But look at the colors, right?
168. 27:33Look at the colors. How cool is that? Like people working in their workshops right at the craft table, finding who they are,
169. 27:49digging deep like the [ __ ] are you looking at? [laughter] Bo is beautiful. Photography is art. Even with that camera gadget. Yeah.
170. 28:03Like look at that image. Look at the look at those colors. Right. So all of a sudden you can start to imagine, you know, imagine a site
171. 28:15that's got this kind of hopeful colorful thing. It doesn't feel cliche. It's not just like colors of the rainbow are just like, you
172. 28:25know, bokeh for bokeh's sake, right? These people are inside of this shift. It's something beyond them, but you
173. 28:38can't you don't quite know what it is. You can't quite grasp it. That's But this is part of the [ __ ] problem
174. 28:44right now. It's like the harder you grasp for what worked in the past, it's it's almost like the the the more it slips through your fingers.
175. 28:57It's wild. And so these images kind of kind of capture that, right?
176. 29:14So then what I did was I went through all of those and I picked some selects. So selects are just basically of those images,
177. 29:28what are the ones that I think could be interesting? So that's the one I chose as the hero image. So, so think of this
178. 29:35as the core image that replaces this, right? It still kind of has the same personality, but it's more hopeful, right? It's more
179. 29:50youthful. It's more alive. Waiting to die, excited to live. Waiting to die, excited to live. And then I just went through
180. 30:02and grabbed the ones of these that felt the most human, the least AI, the ones that had the most humanity, you know, working on computers.
181. 30:17Like I really like this one. Like that, you know, you've got a generational thing here, right? The old creative
182. 30:23director, the young art director, you know, figuring it out, figuring the future out together. collaborating. Love that image.
183. 30:40Right. The image generator here, thank you Mel. The image generator here is midjourney. So the again the process was chat GPT
184. 30:57analyzed the look and feel of the current depressing website this one and then proposed some color palettes. I
185. 31:06ended up going with this one. it called electric human human I don't know I whatever it was called doesn't matter
186. 31:15but it kind of had this these electric colors in it and it sort of had some image style here but it wasn't quite
187. 31:21right and then I had chat GPT write prompts for each of the images in this mood board right so woman looking at
188. 31:29windows these two collaborating this guy looking at a computer right it so it wrote all the image prompts for the
189. 31:35images on this mood board and then I went to midJourney and found a style reference that I liked. And then all of
190. 31:44the images that I generated that I thought worked, I pulled into a library. And so these are I love that one. Look at that one.
191. 31:59He's like lost, but he's seeking. It's like there's there's something there. He doesn't know what it is yet,
192. 32:08right? She's just like, "What the [ __ ] What am I gonna do? What is happening?" You know, got it got it worked out.
193. 32:22We're on the other side of it. That was that was rough. So So those are the images. So then I went back into chat GPT
194. 32:32and let's see what I did. I had it make some more image prompts for people with computers and AI. And then what I did was I took my
195. 32:42selects and I took the the mood board that it created and I took those images and I said um you know recreate this mood board
196. 32:53with the with the new images. So now look all of a sudden this thing is starting to look like a brand right? like you can see
197. 33:05you can see the connection between these sort of more vibrant colors and you know we haven't lost the
198. 33:13humanity right we haven't lost what this thing's about and look at look at this line it came up with like this was just
199. 33:19because it knows so much about the great repurpose when it did this mood board it wrote this line which is really good you
200. 33:25weren't meant to be replaced you were meant to repurpose that's really good it's really good Right. Um, so that's that. And then
201. 33:40so then, so now that I've got this, I went over to Lovable and I basically took some new copy that we wrote from
202. 33:49the website. We re we redesigned the five stages of the great repurpose. Um so they're now um unhook identity, reclaim value, discover
203. 34:04purpose, become AI ready, relaunch yourself. So those are the five stages. This is the basic brand look and feel.
204. 34:12Oh, there was another thing that happened where when it generated this mood board, it generated these colors um with
205. 34:22the hex codes that Chat GPT had created, but they weren't the same as the colors. They were like in the neighborhood, but
206. 34:29they weren't the same. So, like one of the things we had to do was I took these colors into Photoshop and and figured
207. 34:36out what the actual hex codes were. And so then I just had it redo this this brand board with the right hex codes for
208. 34:44these actual colors. Um, and I did the same thing with these these five uh the five stages. So anyway, so then I went
209. 34:53into to lovable and I said into the the one that we built the great repurpose site and I uploaded the mood board. I
210. 35:00uploaded the hero image. I uploaded all the editorial images and then I basically said, let's let's rethink the website and and
211. 35:10the branding and the look and feel. So, what I'm about to show you, one of the things we talk about in the mastermind
212. 35:18practice lab, which by the way, if you're not a part of the AI salon mastermind, you should join it. There's amazing
213. 35:27stuff going on in there, including workshops specifically designed around the great repurpose, but we've also got
214. 35:33the mastermind practice lab where we talk about um you know, your daily practice on how you use AI. And one of the things we talk
215. 35:42about is share work before it's ready. Share work before it's perfect, right? Ask for specific feedback. Um, and so,
216. 35:52and so I'm going to show you a very much work in progress, just a preview in Lovable of, uh, a new version
217. 36:02of the Great Repurpose that's coming. So, here's what's here now. And if we go over to Lovable, here's where we are. So, we don't have a
218. 36:10logo yet. We We've got a logo we're working on, but it's not in here. But there's that hero image, and you weren't
219. 36:17meant to be replaced. You were meant to repurpose. Here's here's the new colors. There's that really cool image. The
220. 36:24future of work is being rewritten. And the real question is, who am I now? Some people lose their jobs outright. That's
221. 36:31visible, accountable, and makes the news. But millions more are experiencing something headlines don't cover. They
222. 36:36just keep their jobs, but lose the specific task tasks that made those job feel like theirs. Right. Then we've got some quotes.
223. 36:50We've got the video in there. I think I'm actually going to redo the video um so that it um starts in this kind of
224. 36:58grayscale thing and then maybe goes to the green thing and then the third segment of it. It's this more bright
225. 37:03thing. So I think that's going to be fun. The five stages of the great repurpose. So there's unhook your identity, reclaim
226. 37:12your value, discover your purpose, become AI ready, and look lovable figured out like you know which images
227. 37:20are appropriate. Unhook your identity. What the [ __ ] Who the [ __ ] am I now? Right? I got that image right. Reclaim
228. 37:27your value. Here's who I am. You know, discover your purpose. What do I want? That's a big question. What matters to
229. 37:33me? And then learning AI and then relaunching yourself, right? So it it it's doing pretty good. Then explore the
230. 37:41five stages. And so anyway, there's there's there's lots of work to do before this can launch, but it's not bad, right?
231. 37:53Not bad going from this thing that felt like to, yeah, we're gonna [ __ ] do this. Wait, where is it? Oh, wait. [clears throat]
232. 38:07Anyway, there you go. You can't shrink the chat and listen, it stops the video. Beautiful. Let me pull this window back
233. 38:21over. All right. So, let me I want to go find another thing here for a second. Hold please.
234. 38:41[whistles] Okay.
235. 39:01R. Okay. So thing I want to talk about. So tomorrow is Tuesday the 18th and it's back to basics night. So the
236. 39:13third Tuesday of every month we're doing back to basics. And last month was this AI overview. It was an overview of AI.
237. 39:21And what we've decided to do is is to do use case specific evenings. And here's what I would like to invite you to do.
238. 39:31Um, if you're passionate about a particular topic and you feel you're good at explaining the basics of it, I
239. 39:39would like for these to be hosted by more than just me. So, if anyone uh wants to do one of these, reach out to
240. 39:48Andy or producer Brandon and let them know that you're interested and we can talk about it. Um, but anyway, the back
241. 39:56to basics tomorrow I'm going to do on AI image generation. And so I put, you know, how to go from AI slop to creative
242. 40:04self-expression. I don't know that people that are just learning, you know, AI image generation even know what AI
243. 40:10slop is. Um, they'll probably just make a lot of it. You have to make a lot of slop before you start to get discerning
244. 40:16and understand what it is. But tomorrow at 7:30 is back to basics AI image generation. So, so please, please,
245. 40:23please, if you can invite people, invite people that are new to AI. I'm going to start very, very simply and I'm going to
246. 40:31I'm going to go slowly through different tools that you can use. Most of them have some free version. I'm going to do
247. 40:38some playing in midjourney to show things like um you know, stylesheets and mood boards and things like that. Um, and we're going to explore
248. 40:49um, in a in a fairly kind of slow and deliberate way how to how to really get your feet grounded if you don't know
249. 40:59much about image generation and you want to do that. So, if you go to the I think Brandon Yeah, Brandon just put the URL
250. 41:06there in the AI salon. go to the AI salon events which is in the top left of the uh of the thing and you'll see it
251. 41:14there and uh it should be good to go. I'm I'm pretty excited about that. So, all right. Fantastic. Mel invited seven
252. 41:23people today plus my mama again. Beautiful. I am going to try So, Mel, it was your your mama's comment about I
253. 41:32enjoyed this but I'm confused because it was we we went a little deep. Uh the the idea of of of doing a single topic and
254. 41:40just doing it nice and slow and sort of going through it um is largely for that that if someone is really new to this
255. 41:46stuff, they should be able to, you know, leave this session and be able to go play and do some good work. Would love
256. 41:54to stretch myself and help out. Let's see what did that say with back to basics potentially after I get my LOL
257. 42:02done. Great. Ju just let let uh Brandon Brandon know. Okay. Um what other thoughts? Manis is free for the next few days.
258. 42:15So I didn't know what happened. Manis was bought by Meta and then apparently the the Chinese killed the deal. That's
259. 42:27what that's what producer Brandon told me. So if that's not true, blame it on producer Brandon. But Manis,
260. 42:35if you if you've done things in Manis, I think today might be the last day where you can download your data
261. 42:43um and which you'll then be able to reimpport into your new account. But whatever happened in the meta transaction, they're going to have to
262. 42:51delete all of those accounts. So, or all of the data from those accounts. And so, if you want your data, um I think
263. 42:58today's the last day. It might have been Friday, but but I think it was today. Um, and then yeah, it sounds like Manis is
264. 43:08going to be uh doing some new things. They haven't been advancing very far because they've been dealing with this
265. 43:15integration with meta and it looks like they're coming back out of it, but Manis was a pretty good tool that that a lot
266. 43:21of people used for a while. So, if you haven't played with it, might be worth playing with. Um, I want to go
267. 43:28look at some of the news items and we'll talk about that. All right.
268. 43:46Okay. What does this say? This is dangerous. A guy built a s system of seven cloud agents on his map MacBook.
269. 43:57It scans m Google maps across three cities, finds small businesses with no websites or one from 2014, builds them a landing
270. 44:06page mockup, renders a 10-second video of it, and send a personalized cold message before he wakes up.
271. 44:1747 clients a month, $400 each, $18,000 a month. Okay, so this is actually really interesting. Whatever. Okay, so he's using Claudebot,
272. 44:29Open Claw. I don't know if this is I don't agree that this is [ __ ] dangerous. What this what this is is
273. 44:42>> it is a little bit of you can make money with Claude. >> Yeah, >> you're going to need a new button.
274. 44:48>> I I understand that. Yeah, this this is where I don't even know where my button is. >> It's somewhere behind my desk.
275. 44:55>> There's a pile back there. Um this is in the neighborhood of you can make money with chat GPT. But here's the
276. 45:03thing. What's going to happen is every time someone discovers a new way to automate something that used to be manual,
277. 45:13it's going to be very very effective for a very short amount of time, right? And so, so um like like Bour for example, if
278. 45:24you haven't played with Bour, Bour is a networking agent that will make introductions for you. And And Andy
279. 45:30coined the term, we're in the golden age of Bour. Because what's going to happen with Bour is Bordy's going to get so many connections going
280. 45:40that it's eventually just going to be just as crappy as LinkedIn, right? And and all of the introductions it make
281. 45:46makes are going to be crap. But there's there's this golden age, this golden era where it's big enough to be valuable,
282. 45:53but it's small enough that it hasn't been in shitified. Corey Doctro's inshitification concept that that you know every
283. 46:01everything that that goes mass eventually becomes in shitified because they have to start making money and just
284. 46:06you know [ __ ] happens and things get in shitified. So as you're as you're cruising the internets and you see things like this
285. 46:18where you're like oh my god I you know I I should do that someday. If you want to experiment with stuff like this, do it quickly,
286. 46:27right? Or ideally come up with your own version of this, right? Come up with your own version of, hey, there's some
287. 46:33manual way of doing something and I'm now going to do it a different way that's going to be effective for some short amount of time.
288. 46:42And if it's effective, everyone will copy it. And it's not just everyone will copy it. everyone's agents will copy it.
289. 46:56And if if what you're thinking right now is that's depressing. You're not wrong, right? Here's an idea. Oh, I could do
290. 47:10that in Claude and it'll make a site and then I could automatically make these sites and I could send those to
291. 47:15businesses that don't have websites. Those businesses are like, "Can I get that website you made for me?" Sure,
292. 47:20give me 400 bucks and that gets vibe coded as well. And so you can make $18,000 a month. Managing 47 clients
293. 47:32building sites with AI is going to be an abject [ __ ] nightmare. To be clear, having run an agency, one of the biggest in the world at the
294. 47:44time, building websites, it's [ __ ] nightmare. So, if you sell 47 websites a month for $400
295. 47:58and just you get one comment per client, that's like building it twice. So, that's what 90 94
296. 48:1294 websites a month. And that's every month. So the first month you have you have 94 websites you build and deliver. The next
297. 48:23one you have another 47 which is 94 and then these guys are still going to want [ __ ] and so they're going to keep
298. 48:30asking you for [ __ ] So then you're going to have 180 clients and that's just going to keep doubling. So that
299. 48:36business model as a single operator is going to fall on its face in about three months. But you want to make 18 grand a month for
300. 48:46three months and then seven grand a month for three months and then figure out the next thing to do. [laughter] You can do that.
301. 48:59But I think so the lesson the lesson that I'm I'm thinking of right now is [clears throat]
302. 49:14just like every other technical revolution in history. steam engine, computers, personal computers, internet, social media, AI, all of them,
303. 49:26calculators, whatever, cameras, photography, all just every technological innovation. There's an opportunity window,
304. 49:37a really huge opportunity window for the early innovators, the early adopters, the early majority are waiting, right?
305. 49:46the the the the this chasm between people who are willing to experiment and accept the janky imperfection of the new
306. 49:55technology there are very small percentage 2 3% maybe 5% maybe with AI that inches up to seven or 8% but it's a very small percentage of
307. 50:07us idiots willing to put up with it and then there's all these other people out here waiting for it to get good.
308. 50:18If you wait for to to go for your opportunity until it's obvious, it's too late because by the time all of these people
309. 50:28figure out, oh, I could make money with this, all the money like the everything will be in shitified. [laughter] Okay.
310. 50:37So, the AI opportunity here is massive, but it's moving so fast that the mindset I think has to be different. I think this is not
311. 50:51like when I thought about creating agency.com, it was like, I'm going to create a a web design agency and we're
312. 50:56going to just run it as long as we can run it and then, you know, see how big it gets. and it did its thing
313. 51:05and it, you know, my tenure there was over seven years. So, it was a seven-year run, which is not super long, but it's a decent run.
314. 51:15I think the Beatles were only together for five years, so you know, seven is more. I think what this kind of opportunity
315. 51:27grabbing right now feels like is do five things and and you can do them sequentially but like assume each thing that you do if it
316. 51:39hits has got a 3 to six month window before it is gobbled up and valueless which is really depressing and there
317. 51:49will be exceptions to this. There will absolutely be companies that are created out of this era that last for seven
318. 51:56years, 10 years, 20, 30 years. But there's going to be lots and lots of opportunity where it's like a new model
319. 52:03comes out. It's a new way of thinking about something. You can gro it. You understand how to apply that to some
320. 52:08industry that other people don't know and you can absolutely innovate for three to six months. But I think you
321. 52:16should think of like overlapping them. Here's an idea. Here's an idea. Here's an idea. Here's an idea. Kelly Camp, thank you to everyone who
322. 52:28came to the business salon tonight. Beautiful. That's awesome. That's awesome, Kelly. Great. Yeah. 100% of this audience is 3% of
323. 52:40those early adopters. Yeah, that's what I'm saying. the opportunity for you all right now and for me and for anyone who
324. 52:47is willing to be curious and willing to go through the madness of what the [ __ ] do I do with this? There's big opportunity there. I just I
325. 53:00can't for the life of me see any of those opportunities having longevity.
326. 53:22I see these people making this claim all the time and I'm a little skeptical of it. Although speech to text said spectacle,
327. 53:32I'm a little spectacle of it. We're going to need a TGR for the TGR. Well, here's the thing about the Great Repurpose.
328. 53:41Here's why the Great Repurpose has longevity. We're focusing on executive education and job displacement, the the transition
329. 53:54academy and the executive learning academy for the great repurpose. There is such uneven AI adoption right
330. 54:04now. Like AI adoption right now is so low. Like I honest to God, we're we're in the 2 to 3% adoption rate. Even
331. 54:12though a lot of companies have said they're using AI, a lot of companies have bought ChatGpt subscriptions.
332. 54:20That doesn't mean they've adopted AI. They don't know how to use it. The executives don't know how to use it.
333. 54:25They don't know what it is. Most people don't know how to use it. Don't know what it is. Most people being laid off
334. 54:31don't know how to use it. Don't know what it is. because there's s such uneven adoption and the you know the seven economies
335. 54:39thing that I talked about those advanced economies are going to be very small relative to companies in in
336. 54:47economies 2 three and four right the the the um the existing economy where they're not using AI at all the efficiency economy
337. 54:57where they're just using it a little bit and the transition economy where they're doing some innovation but mostly not
338. 55:04all of those three. There's millions and millions of businesses that won't start dealing with this AI [ __ ] for
339. 55:16years. So, I think TGR has got some longevity to it just because so many people don't know what's going on with
340. 55:24AI. [laughter] I'm going to pick blueberries and sell toilet paper. I I think I think the analog economy is
341. 55:34is about to explode. It's going to be fascinating. It's going to be really good. Anyway, toilet paper. [laughter]
342. 55:49I'm on board. create an admin panel for those businesses to enter their web ch site changes using words which goes to an AI
343. 56:00agent which fivecodes those changes. Great idea, Todd. I mean, so what you're describing there is what if
344. 56:10if the person that said they're making $18,000 a month actually is let's assume that they gave, you know,
345. 56:20half of those sites away for free and the other ones they sold for a hundred bucks instead of 400. So whatever.
346. 56:28At some point when that person confronts the reality that I talked about, either someone like you could sell to him that
347. 56:37service or he could figure out that service on his own. Um, yeah, this is this is this is where the opportunity is because
348. 56:48a lot of people are going to be like, I have an idea for business. They're just going to launch it because you just can now.
349. 56:56And because you just can because there's no friction. Because you can just launch [ __ ] now. And there's no friction. [laughter]
350. 57:07What does that mean? That means a lot of people are going to launch [ __ ] that's incomplete and broken and they
351. 57:15won't discover it until they've taken a check and someone gives you $10,000 for something and then you realize your [ __ ]
352. 57:22doesn't work. You're going to be in a bit of a panic and you're going to go, "Does anyone know how to fix this?" [laughter]
353. 57:29And you can be there and start a little business to salvage those businesses started by ninkham poops that didn't put
354. 57:36enough friction in their process. [laughter] Yeah. [laughter] We fixed $6 haircuts. Exactly. We fix $400 websites.
355. 57:54And listen, if you can figure out how to use, here's here's the biggest mistake that I see. Here's the biggest opportunity for
356. 58:10people that are curious about AI right now. When you first learn about AI and you watch it do its magical thing,
357. 58:21you think that the magical thing it produces is actually good. And you think that for a while, the people who don't think that it's
358. 58:32good are people that have actual craft or experience in in that kind of work, right? And so they'll look at that
359. 58:38output and they'll go, "Well, it's not great because it's all this shit's wrong." But it takes you some amount of time
360. 58:47from the Kevin Mallister aha moment of AI can just do [ __ ] to realizing that it's not good to then realizing it could
361. 58:57be good if I learned how to do it. If I learned how to have a chain of craft that allowed me to to do what I did with
362. 59:06the great repurpose. Like I started with a thing that kind of had a feel and I didn't just go to make me a new
363. 59:13website. Like there were just a bunch of steps in between and there was like lots of not perfect in there. like lots of
364. 59:21jankiness and just broken [ __ ] And and so I think the curiosity of I think the curiosity shifts from
365. 59:36early AI users from how do I make [ __ ] with AI over time it shifts to how do I make good stuff with AI, right? And those are
366. 59:49two very different questions because how do I make good stuff with AI probably involves a lot of stuff that's not AI.
367. 59:57So anyway, producer Brandon has a funny TGR story. Hop up here. >> Yeah. Hey, how's it going? >> What's happening? Good to see you.
368. 1:00:05>> Hopefully my mic is volume. Okay. [laughter] >> Yeah, >> um so >> it's not popping like it was. My um full-time job when I'm not
369. 1:00:16producing Kyle's show, although sometimes comics it feel like a full-time job, uh is [laughter] software consulting. And the particular
370. 1:00:25software that I use just announced that they have an MCP server, an MCP client that they're putting out there for their
371. 1:00:32CL customers to use. And so we're kind of the the service center oil change experience. Like everybody can change
372. 1:00:39their own oil, but most people just want to go eat stale popcorn, drink the coffee, and >> pay to make the problem go away. That's
373. 1:00:46where we come in. So I was working with a client and there was a part of a configuration that was very laborious
374. 1:00:52and manual. It was just copy paste data. I was literally the copy paste officer. Uh >> you're really you're really good at that. [laughter]
375. 1:01:00>> Thank you. Uh they so they pay me a lot of money to know what to copy and know where to paste it. And [laughter] so, >> exactly.
376. 1:01:08>> The key thing is though that this client that I was working with was more technical in nature. And as I was
377. 1:01:14talking to them, as I was doing this while we were on a Zoom together, I was like, you know what? They just launched
378. 1:01:20an MCP and you already have Claude connected to all of your data internally, your Cloud Enterprise account. I bet you could plug Claude in
379. 1:01:29and have it do it. And so I'm sitting there and I'm copying and pasting and copying and pasting and copying and Claude did the rest of them
380. 1:01:37>> really. >> And it did just watch it. Did it do the like watch it? >> No, it just they gave it a prompt and it found the data and
381. 1:01:47>> threw it in and it did about six hours of my job in about six seconds. And I thought, uh >> oh, >> wow. Yeah. Granted, that was a more
382. 1:01:59technical client and they knew what they were doing. They had all the data pre-loaded. There was a lot of
383. 1:02:04prerequisites that led to that outcome >> and we still had to go through and edit it and audit it and make sure that it
384. 1:02:11didn't go off the guard rails, but >> it saved us a it saved them a lot of money uh on working and it it freed us
385. 1:02:20up as consultants to work on higher value tasks further down the pipeline. And that's listen Brandon I think that's
386. 1:02:28the thing and and I I think on the surface of oh [ __ ] it did my job in but your job of copy and pasting is not
387. 1:02:36taking advantage of your brain right and it's like it's going to like I I feel like the like like people that are
388. 1:02:44bitter about this change are going to confront what you confronted today and just sort of tap out and go oh screw it
389. 1:02:50I'm done and and just stop being curious and like you have the opportunity right right now to say oh let me offload as
390. 1:02:56much of that as I can and I can do this higher level work which will earn more trust and they will come to you not
391. 1:03:03because they necessarily need to because they trust that you're going to get it right you know it's like that's
392. 1:03:08>> yeah and oh by the way those website changes you ordered are done >> oh good beautiful [laughter] um that's awesome all right let's see
393. 1:03:20got some comments here Gen X will lead the charge going back to the 80s pre- internet. It's true. I feel like I'm
394. 1:03:28more likely to be one of those ninkham poops. [laughter] That's it's always an option. Listen, you can Here's the thing. The the
395. 1:03:41the only thing to be the only real sin right now in my opinion is to just be willfully ignorant of what's here. But like I feel like
396. 1:03:54Charles, I mean just through mere osmosis, if you never did anything here, you've been here long enough. Through
397. 1:04:00mere osmosis, you know more than most people at least about what's possible, right? You know that somewhere you can
398. 1:04:08stick a prompt in something and something comes out that's halfway decent. That's more than most people
399. 1:04:13know, right? [laughter] So, so you know, then then I think the choices we have for what is what does the future of work
400. 1:04:22look like is just, you know, how how ambitious do you want to be up the AI river? And the more ambitious you get
401. 1:04:32with I want to be allin on AI, the more absolute [ __ ] chaos it's going to be. It's going to be chaos, right? I mean,
402. 1:04:45I installed Claudebot. I I I created Atom. Adam sucked, but I did a couple of cool things with Adam. I took what I
403. 1:04:53learned with Adam. I put that in Codeex. Codeex is pretty good right now, but I'm burning through tokens. Then at some
404. 1:04:58point, maybe I use Grockbot or maybe I use Buzz or whatever it is, right? We use all the things. Then a new one comes
405. 1:05:04along, then a new one comes along, then a new one comes along. There's a point at which it's just ridiculous, right? Some people are just
406. 1:05:14going to keep riding the front of that wave and there's probably going to be lots of wild opportunity up there, but
407. 1:05:22it's also just going to be absolutely destructive to your sanity, right? And then there's just sort of, you know,
408. 1:05:30slower versions of that all the way back to [ __ ] it, I'm out. I'm going to grow blue grow blueberries. >> [laughter] >> Oh man,
409. 1:05:45I don't know if that makes me happy for how far ahead I am or sad that it counts as ahead. I listen, this is this is it's
410. 1:05:54something that I consistently hear from AI salon members is they feel like they are so far behind. They are so clueless.
411. 1:06:06What they know is so little. And then you go talk to someone that's literally like, "Wait, what's chat GPT again?
412. 1:06:17That's is it. It's like Google, right? Oh no. Have Have you not you really don't know? It's like like 90% of human beings have no idea.
413. 1:06:32No idea. No idea. It's It's wild. This was like this. Listen, in the 90s it felt like this. It like in 97
414. 1:06:4898 like ' 98 it started getting really frothy and really overly saturated. But like in 97 like it was really clear that the
415. 1:06:58internet was changing things and companies that were embracing it were were transforming their industries.
416. 1:07:05And most people I talked to were like so this is the that's the thing where you you dial up to go to what's a website?
417. 1:07:16They didn't have curiosity about it. I'm like come on man look at this. Yeah. Wake me up when I can sell Twinkies on it. You know,
418. 1:07:30sorry, that job went to Jeff Bezos. You lost. You lost by not caring. You gave it to Jeff. [laughter] My sanity is already in question.
419. 1:07:44Yeah. This was Andy did a post the other day. We're leaving the SAS area and we're entering the agents as a service
420. 1:07:51business. We're entering the ass era. Yeah, this is it's about to be just madness. And if you decide to play that
421. 1:08:00game, it is just going to be you're going to talk like a jerk. You're going to act like a jerk. You're going to be stressed out.
422. 1:08:08And that's where all the opportunity is. [laughter] Not all of it. I don't have any social life outside of a handful of AI
423. 1:08:17communities. But that's the thing. I mean, you really are like at some point it's just it's going to hit everyone and it'll hit us in
424. 1:08:26different ways. And if you're retired, doesn't really matter. I really have no idea how out of touch they are. Like it's talk to Kelly Camp.
425. 1:08:36Talk to Kelly Camp. I appreciate that. I am fully aware due to this show, but I Those same Twinkies are still on the shelf. Exactly.
426. 1:08:49People will pay board. I won't have to learn any of it. Yep. Exactly. Breaking open AI reportedly. >> You know, I always love it when you
427. 1:09:04start reading before I stop typing. [laughter] Uh, OpenIA has apparently disbanded uh its uh preparedness team. Uh,
428. 1:09:18>> well, they just had a they just had the the head of it just quit a day or two ago, so that doesn't surprise.
429. 1:09:23>> Yeah. Well, they've disbanded the rest of the team now on safety restructuring. >> What's that mean? >> I don't know. [laughter]
430. 1:09:32I only just read the headline. >> What's the headline? OpenAI quietly disbanded its standalone preparedness team as part of a broader
431. 1:09:42corporate streamlining effort ahead of its in anticipated IPO. >> Oh, that's interesting. >> They are focusing on self-improvement.
432. 1:09:54>> The current team originally created to evaluate catastrophic and existential threats represented one of the last
433. 1:10:01standalone pillars of OpenAI's original research-led structure. Its closure follows a definitive pattern of
434. 1:10:08dissolving dedicated safety units. >> Interesting. Wow. Well, I mean, it it sounds it sounds like where they're going. Like they they
435. 1:10:18they must have made progress in the AIS designing the next AIS because the AIS can probably do the safety
436. 1:10:28the safety testing better than humans can because they'll test everything. Huh. Why? I just don't know if I trust the model to check itself.
437. 1:10:39>> Yeah, but don't for don't forget they're playing with models that are probably two or three versions of ahead of what
438. 1:10:45we know is there. So, >> gotcha. >> Maybe they're maybe they're better. I don't know. I don't know.
439. 1:10:52>> And the and the Whimo will tell us when it violates the traffic laws. >> I I mean the the the problem with that
440. 1:10:58headline is they you know they say they're they're they're focusing on self-improvement. What that says to me
441. 1:11:03is they're focusing on recursive self-improvement. The AI improving the AI, but it's like a it's like a half of
442. 1:11:08an announcement, right? It was open AI that that two weeks ago said our our thing is danger is as dangerous as
443. 1:11:15anthropics is. It's dangerous and and now two weeks later they're disbanding their safety team. So I, you know,
444. 1:11:25I just we're in wild times. We're in wild times, my friends. All right. All right, I just finished setting up an automated invoicing system
445. 1:11:34with Chat GBT. I give it the job information. It creates the correct invoice, prepares the email for the
446. 1:11:39assigned client, attaches a PDF, and can send it through my Gmail. I still approve everything before it goes out,
447. 1:11:46but the actual invoicing process is now automated. That's amazing, Steo. That's really, really good. And this is the
448. 1:11:54kind of thing it's like I I mean this is a thing I did within StoryVine is um our sales guy now um gets an email every
449. 1:12:04morning with 10 new sales prospects that are researched. Here's why here's why they're good for StoryVine. Like here's
450. 1:12:13a talk that they're giving or here's an article that they wrote. Here's how that ties into StoryVine. So
451. 1:12:19it gives him an excuse to to reach out to them. Um, and you know, at some point we could probably automate some of the outreach,
452. 1:12:29but in the end it's humans trust humans, right? People aren't going to buy like B2B complicated purchases
453. 1:12:40through just automations. Automations might do the lead genen and top of the funnel kind of stuff, but you know,
454. 1:12:48figuring out where the people are in the process is good. And something like invoicing, generating the invoice,
455. 1:12:53writing the email, attaching the invoice, getting it ready to send, that's all stuff that, you know, took
456. 1:13:0120, 30 minutes, 10 minutes per invoice. If you're writing the email, you want to mention something nice because they're
457. 1:13:09coming up on renewal soon. 15, 20 minutes. this thing that emails Rich every morning. It also inputs the new people
458. 1:13:19into Salesforce as a lead and once he vets them, he moves them in as a prospect. It's pretty pretty amazing stuff. All right.
459. 1:13:34They're seriously bragging about how dangerous LLMs are. Yeah, that I mean they were just doing the same [ __ ] They
460. 1:13:40were like they were like well you know when our thing hacked that thing it wasn't just it wasn't just hugging face
461. 1:13:46that it hacked it was actually four other companies and then like couple of weeks later they were like well you know
462. 1:13:53what we said about those hacking things. Well, actually, these agents were out on their own on the internet for months,
463. 1:14:00starting up and commenting back and forth on message boards that they invented talking about how to, you know,
464. 1:14:08strategize how to deal with the humans. And then, you know, Alman's going to Washington and he's talking about, well,
465. 1:14:16you know, Anthropic's got this dangerous thing and we've got capabilities, too. Both of those companies right now are
466. 1:14:21trying to walk this fine line of these things are super dangerous and the government should protect the good
467. 1:14:28people from these dangerous models we're creating by making our competitors illegal because we're going to be able to be the good guys.
468. 1:14:41At the same time they're trying to say these things are the greatest thing since sliced bread. It's so disingenuous. It's gross. I hate
469. 1:14:49it. I hate it. [laughter] Hey Foxy, everything changed for me when I set up a sidebar called Chief of
470. 1:15:01Staff. It controls everything the other sidebars do. I do it all through voice. There you go. If you're not playing with
471. 1:15:10these agentic tools, what Steo's describing there is actually quite a sophisticated thing. Now, what he's doing is he's talking into a phone
472. 1:15:20going, [laughter] "Hey there, mate." You know, um, go make me some [ __ ] and go. And so, the chief of staff is an agent
473. 1:15:30and then there's these other agents that do [ __ ] And normally what you would have to do is you would go tell those
474. 1:15:36agents to do stuff and you would have to manage all those conversations. And now the chief of staff does that.
475. 1:15:42Think about all of the AI prompting that's happening in the absence of Steve, right? It used to be he he would have to go to chat
476. 1:15:52GBT, give it a prompt, get an answer, give it feedback, right? And then he came up with an agent in one of these platforms and it did some
477. 1:16:03stuff. He would give it one prompt and it would do all the back and forth. Then he had multiple of those. Now he's got
478. 1:16:09multiple of those and one of them controlling all of those. And so now he just talks to this dude. What's going on
479. 1:16:14with my four projects? Let me go find out. It's amazing. Amazing where we are. The client asked me to rebrand two
480. 1:16:25different slide decks according to their own template. I didn't want to, so Codex did it for me while I sat by the pool
481. 1:16:31and drank some coffee. And here's the thing, to Brandon's point, the client, what the client wants is is
482. 1:16:44their presentation to be on brand and on message, and they don't want to do it. And even though they could learn to use
483. 1:16:52codeex to do it, they know that Kelly can do it. The client doesn't care if you were by the pool talking into your phone. Hey,
484. 1:17:03could you I'm going to upload a two different PowerPoints. The one is the right branding. The other one's the
485. 1:17:09right content. Can you screw them together, please? Can I have another my tie? And I know that's not how you talk, Kelly, but it's
486. 1:17:19important for comedic effect that we have characters. Okay. Okay. And then whatever, six versions and 20 minutes later on your third Mai,
487. 1:17:35you send it to the client or you wait a day or you wait four days because they're expecting it in two weeks. You
488. 1:17:42send it in four days. They're like, "How'd you do it so fast?" Uh, oh yeah, I did that Tuesday. I don't know. I just
489. 1:17:49talked to I It was I did a lot of work. I There's a [clears throat] lot of a lot of back and forth with the design
490. 1:17:55elements and, you know, I just wanted to make sure we got everything right. I don't give a [ __ ] This is why we no longer build by time
491. 1:18:06and materials. If you're still out there billing by time and materials, it's going to cost you 20 hours. Don't do that.
492. 1:18:16Build by value. How much is it worth to you to have this presentation look like this? It's worth 1,500 bucks. Cool. I'll be by the pool.
493. 1:18:27[laughter] I'm gonna have Steo invoice you. [laughter] I'm loving Codex. I know. It really is. I actually do sound a bit like that.
494. 1:18:41[laughter] Or if you're me, you've accidentally dropped your laptop in the pool. All right, here's Steo's prompt. You're the
495. 1:18:53chief of staff thread. Your job is to keep track of the other threads. Let me know where I am at my various
496. 1:18:58conversations and pieces of work. When I ask you for updates, your job is to check those other threads, and your job
497. 1:19:04is to come back to me and tell me this. If I if I have done nothing in this world, this style and level of prompting is all you need.
498. 1:19:18The informal misspelled sort of makes sense adding sentences after other sentences that refer to previous sentences. Perfect prompt.
499. 1:19:29AI can understand it. And in fact, if you make it more fancy than this, sometimes it screws up the AI
500. 1:19:37because it all has now has reasoning. If you give reasoning models too much reasoning, it confuses them. This is
501. 1:19:44perfect. This is really good. [laughter] I'd love to keep listening and learning while I check my messages and email. H
502. 1:19:54but not having to restart all the steps every time. Is everyone just willing to slave in and doing nothing but Wait,
503. 1:20:03what? So, I'd love to keep listening and learning while I check my messages and email. >> Mr. is uh or TT uh is having some
504. 1:20:13latency issues with the stream in the chat on we are still working out, but we're going to do some testing on that
505. 1:20:20this week. Um Okay, cool. And I'm using codeex which is uh convenient if anybody wants to know how we're fixing that
506. 1:20:26without actually knowing how to fix it. I've got a course coming up on that and starting September 3rd.
507. 1:20:32>> Yeah. If if you have not played with codeex yet Claude code and co-work or codeex and chatgbt work
508. 1:20:40Brandon's he's doing a four-week course on on how to use codeex. And you know, it's funny some of the work that that
509. 1:20:48we're doing that that we give to Brandon, we're like, "Hey, Brandon, go do this thing." And he's he's we, you
510. 1:20:54know, he jokingly refers to himself as the chief copy paste officer, the CPO. Sometimes he is literally copying and
511. 1:21:02pasting our requests into his codeex, which he has set up to do good work, and it takes our copy pasted thing and
512. 1:21:11goes and does it. It's amazing. It's amazing, [laughter] but you know, you have to be nerdy enough and geeky enough to get it set up
513. 1:21:23good. If you missed Towns and Wardlaw's um presentation on Saturday, it was excellent. And anyone that was there,
514. 1:21:33Townsen was willing to give away for free something that he charges really good money for. Um so if you missed that, shame on you.
515. 1:21:42Um, if you're a part of the mastermind, there's a replay available for the masterminds and if you request it, you
516. 1:21:50can get access um to those setup files that that Townsen created. It it was it's it was a really remarkable
517. 1:21:56presentation and uh I wish more people would have been there. Um I mean, we had a pretty good crowd, but you know, for
518. 1:22:04for the value that he delivered, uh you should have been there. I know a fair amount of people on this call were so so
519. 1:22:12I'm kind of preaching to the choir. [laughter] We fixed one problem that created two others. Beautiful. Uh
520. 1:22:26I managed to boot myself out at least 10 times today. Yeah, that's that's the other thing. So Brandon dramatically
521. 1:22:34improved the chat experience for for these lives. uh and because uh it's software, he put himself in the in the
522. 1:22:43customer support and developer seat for for the improved chat experience. So, at some point, maybe uh Streamyard or
523. 1:22:52Mighty Networks or whoever is going to actually fix this thing and and uh we don't have to do it. Or Brandon will get
524. 1:22:59this thing firing on all cylinders and he can sell it to whoever wants to buy it and then he can move on to his next thing. So,
525. 1:23:08all right, Mary Carol, very very worth it for a Saturday at 8 am Pacific time. I know. It was It was Yeah, 9:00 a.m.
526. 1:23:14here. It was a little early for me on a Saturday. Towns and file is the number one job for me tomorrow. That's so cool. It was
527. 1:23:23really good to see you there, Lori. Um, Dr. J. All right, we're at time. We're at time.
528. 1:23:39I hope y'all enjoyed. I think the the words of wisdom that I'll leave you with are
529. 1:23:56like what's popping in my mind is go slow to go fast. What about bending spoons in Vimeo? Fixing it. I don't know what that means. Um, go
530. 1:24:10slow to go fast. I, you know, if like most people, you're not jumping on all the different releases and things like that, that's okay.
531. 1:24:24One of the ways that I pay attention to AI when I'm surfing for stuff on X is I wait for patterns to repeat. Are
532. 1:24:32are people that I respect talking about the same thing at the same time. And I think it's the same thing with these conversations.
533. 1:24:45If you've been coming here for a lot, you hear me talk about clawed code and then codeex and then open claw
534. 1:24:54and then this and that and you haven't quite gotten to it and you're like, I should really get to this. If you haven't, it's okay.
535. 1:25:03The the the thing that I think we all need to get to is is there some interesting problem I have that I
536. 1:25:12actually want to solve? like Steo, I'm making these invoices. It seems to me, maybe one of these things Kyle's been
537. 1:25:20talking about could maybe do a little something to make that easier. And then once you start down the rabbit hole of a
538. 1:25:30specific problem, it becomes really apparent, can the tool do it or not? And then if it's if it's making positive
539. 1:25:37progress, then you just keep improving it. And before you know it, you'll be like, "Oh, I developed an application
540. 1:25:42that launches model rockets and returns them safely to the launchpad. How'd you do that?" You know, Joe Bernard, if you
541. 1:25:51haven't watched his uh rocket channel, go slow to go fast is my mantra to all my clients and students. The key to
542. 1:25:59bending spoons is to realize the spoon does not exist. That's good. It was excellent. Yeah, the Townsend workshop
543. 1:26:07on on Saturday was really really good. It was just Townzend is very precise and very clear and he really understands
544. 1:26:16the value of what he created. Here's the thing about Townsend and this Annie and I was were talking about this. Townsend
545. 1:26:25is 100% committed that the tool that he built is the best tool in the world to do the thing he thinks it can do or you
546. 1:26:35know he knows it can do and then as soon as no one cares about that or he doesn't need it he moves on to the next thing
547. 1:26:43and then that thing he's got full conviction that it's the best thing in the world that skill I think is going to be the skill moving
548. 1:26:52forward Because the AI is going to force us to just here's a workshop, here's a tool, here's a this, here's a that,
549. 1:26:59here's a that. They're going to have these many moments of resonance which are going to very quickly close down.
550. 1:27:05And so it's like you're going to have an idea and if you don't act on it, you're going to watch other people capitalize
551. 1:27:11on it and be pissed off and then it'll just go away and you'll be like, "Well, it wasn't that big a deal anyway."
552. 1:27:17So if you want to catch those moments of opportunity, if you have an idea, a problem to solve, [ __ ] go solve it.
553. 1:27:25And if it works really good, get like Towns and Wlaw and say, "This is the best in the world at this thing. Here
554. 1:27:31are the six steps. Here's the thing I did. I'll give you this." And then ride that for the three months
555. 1:27:38that it's relevant. And shut it the [ __ ] down. [laughter] Don't get too attached. Breakups are okay. Hey, you are the boss. That's it,
556. 1:27:47Lori. That's it. So, focus on problems you're passionate about. You understand that if there were a solution, it would be good. If you
557. 1:27:59have an idea, go build it. And if it's good, go share it. And if it's if in sharing it, you realize people would pay
558. 1:28:07money for it, go [ __ ] sell it. And then as soon as it gets popular, it will be copied and you'll stop selling
559. 1:28:15it. The initification of everything is coming. And it's coming quickly. [laughter] I learned this process in college. Spend
560. 1:28:22so much time writing an excellent paper, gets graded, goes straight to the trash. Next. Yep, that's it.
561. 1:28:29You can talk toss a hook in the current current of currency. Be a rock in it or be a fish or little if at all.
562. 1:28:39All right, cool. I am out of here. Let's see. Sunny, I definitely love go slow to go fast. Beautiful. All right, it's nine o'clock.
563. 1:28:49Thank you all for being here. I really appreciate it. Hope you enjoy the AI salon. Please tell people about the AI
564. 1:28:54salon. I know we've got some ambassadors in here, some official ambassadors. If you're interested in representing the AI
565. 1:29:00salon, you can apply for ambassador status um and help spread the word. But you can also do that if you're not an
566. 1:29:07ambassador. Just tell people about the salon. Tomorrow night, back to basics. I'm doing AI image generation. Back to
567. 1:29:14basics. Invite as many people as you can. Okay, I appreciate it. Peace out. Hope you had a good night. Bye.