7/27/2026 - Testing ChatGPT Remote Voice Controls and Automated Workflows in Real Time

Transcript

1. 0:05[music] [music]
2. 0:20[music] [music]
3. 0:35Happy Monday. Welcome to the AI learning lab. Good day to one and all. Um, it's a Monday. I I'm feeling spicy. I'm feeling
4. 0:46saucy. There's probably some things that are going to yank my chain. That's vernacular for make me kind of angry. [laughter]
5. 0:58Linda Sherman, happy Monday from Aloha from Kawaii. Fantastic. Beautiful. Good to see you. So, Linda knows how to turn on the comments. If
6. 1:11you want to turn on the comments so you can see the good people and hang out and chat with them and things like that,
7. 1:16right up here in this corner is the button right underneath the logo there. You just click on that and boom, out it
8. 1:21comes. Frostbitten was shaking. What's going down? And then down here in this corner is is the button where you can if
9. 1:28you're on desktop, you can pop this window out as a separate little streaming window. And then you can open
10. 1:34a new tab and have the AI salon in the new tab. So you can be multitasking and things like that. I do that every time I
11. 1:41start one of these shows. The people that are here probably know how to do that. But in the off chance you're new
12. 1:48here, welcome. Welcome to the AI salon. Welcome to the AI learning lab. Um, if you're new to the salon, be sure to
13. 1:55check out all the spaces down the down the left hand side of the screen. I can't remember which side's left or
14. 1:59right. I think that side's left for you. Um, hey Danielle, it's shaking. Um, okay. The new voice icon showed up on
15. 2:09chat GPT laptop today. The new VO the voice the the voice capability when when coding with chat GPT work or chat GPT or codeex
16. 2:26all essentially the same product with different labels and slightly different features. >> [sighs] >> But the
17. 2:39talking to a thing and having it build things is pretty flipping cool. It's pretty flipping cool. Actually, at some
18. 2:46point I might want to bring Brandon up. Maybe think about teeing up that that video you sent me, Brandon, and maybe
19. 2:53you can show it and and tell tell the good people what it did. Um, if you haven't been playing with the the agentic platforms, Claude Code,
20. 3:03Claude Co-work, Chat GBT Desktop App, Chat GBT work, Chat GPT codeex. Um, they're they're worth playing with.
21. 3:16They're they're they're pretty good. Um, there's a lot going on. I think one of one of the things that I've been doing is
22. 3:28um rather than try to keep all this [ __ ] in my head, I'll just go to X and just look at the news stories and then find
23. 3:34ones I think are interesting and comment on them. There's a fair amount of them. Um I have been historically
24. 3:46decently a fan of both Sam Alman and Daario Amade. Um, I think Daario Amade is a little [ __ ] weasel. [laughter]
25. 4:00He's starting to bug the [ __ ] out of me. [laughter] And that's that's just my opinion. That's not necessarily the opinion of
26. 4:10the AI salon or the AI learning lab. [laughter] New PR statement from Anthropic today, Danielle. is um so so what happened well
27. 4:23we'll go through this in the news but but the government basically said at Dario Amade's request Chinese models and openw weight models
28. 4:34are dangerous and they're stealing from American companies and we're considering making them illegal. So, every tech
29. 4:41company on the planet basically signed on to this letter um spearheaded by Jensen Wong from Nvidia and then
30. 4:50Microsoft was on it and like all of them were on it. OpenAI was on it, Enthropic was on it and then that whole group today
31. 5:00started the Open Secure Alliance. You know, basically it's a coalition of companies in support of openw weight
32. 5:09models. Basically, the argument that if you only have two companies that have the core models and everything else is
33. 5:17illegal and you have to follow their rules, then if something happened like what happened to happened to Hugging
34. 5:24Face where they got hacked and they couldn't protect themselves because the safety guard rails on the anthropic models were so
35. 5:33strict that they couldn't even protect themselves and they had to use a Chinese model to stop the attack that Open AI was per
36. 5:41perpetrating on them. So they created this coalition to say let's put our trillions of dollars of market cap money speaks behind this and
37. 5:54everyone signed it except Anthropic and then they put out a weasly little statement. We never said we didn't want
38. 6:01open source. We just don't kind of agree with the part where we would have to compete against them. We don't want them distilling our models
39. 6:10even though our entire business model is based on distilling knowledge of other people. Anyway, [laughter]
40. 6:21oh yeah, I saw that. Um, Brandon said that [clears throat] your your um clawed artifacts which are like their
41. 6:30applications, you know, when you make like a game and you put it out in the world or you don't put it out in the
42. 6:38world, all of those artifacts have been indexed by Google. So, so if you do a search for clawed artifacts, like you do
43. 6:46site claude.ai AI slash something artifacts. You can just see people's artifacts with all their like company data and [ __ ]
44. 6:56like that. [laughter] [laughter] Um, if anyone has any questions about things going on, feel free to hop them
45. 7:09over there. Gemini would never. Charles Adams, it never fails. The one I distrust the most are [clears throat]
46. 7:15the ones that claim moral high ground by their own decree. Yeah, this is the thing. Listen, I was actually a big fan of of Anthropic
47. 7:26and and their constitutional model, which by the way, we haven't heard [ __ ] about the constitutional model in a long
48. 7:34time. So, I don't even know if they're still if that's still a part of their safety thing. It might be, but I don't know.
49. 7:42but he didn't like what Open AI was doing with regards to safety and alignment and so he quit and he started his own
50. 7:52company and like I applaud that he had a different philosophy right and but but like once you get trillions of dollars involved
51. 8:02[ __ ] changes and you know they're basically just trying to create a duopoly between them and open AI
52. 8:11at at the cost of frankly like American ingenuity. Like it's it's just I don't know. It's just weasel behavior. I don't
53. 8:19like it at all. And I'm listen I'm not a I'm not a you know I don't think Sam Alman is above reproach. I don't think
54. 8:27he's he's done super swell stuff either. Um but but the the the kind of virtue signaling and just crap that that
55. 8:41Anthropics been doing is is kind of gross. It just feels very cynical. Kim K3 open weights were released today.
56. 8:48Alama has a cloud monthly plan, but Kim K3 isn't included in the plan because it's really expensive. It's a 2.8
57. 8:54trillion parameter model and the open weights are out there. So, so here's the thing. So, Kim K3 is is sort of at the
58. 9:04fable level, you know, and I think it's I think it's probably I think it's something like 1.8 gigabytes.
59. 9:13Is that right? I think that's right. Saying I distrust one the most does not say I trust the other. Exactly. [laughter]
60. 9:24I'm not the only one melting down tonight. Producer Brandon, I can't decide how I feel about you can always
61. 9:28talk to your computer from anywhere. Look, I'm working from the forest versus having a work work life balance. No
62. 9:34disrespect to source camp. Well, this is this is a there there's a boundary that I think we all have to find, which is
63. 9:49if you can just make anything, we can just make things now. And you enjoy making things, go make things.
64. 9:58But if you're making things feeling like I' I have to make things because that's how I'm going to make money and I'm
65. 10:04going to go make a book and I'm going to sell books and I'm gonna And and if you're on this neverending quest of
66. 10:08like, I've just got to make the right thing. Things that we make are going to be increasingly valueless.
67. 10:19So it is my strong contention and I, you know, I share this in here regularly. I'm creating a business around it right
68. 10:26now around the great repurpose that is around you understanding who you are and what you value and creating things that make you happy.
69. 10:40creating things that in Brandon's case make a cool new game for his kids to play with or, you know, something fun to
70. 10:49to share with the AI salon that's coming from a place of who he is, go do that. Long as it's making you happy. But to the extent that you're
71. 11:00like trying to keep up and trying to, you know, um, what Matt Schumer oneshotted a 3D shooter game and now
72. 11:10everyone's like, I got to make a 3D shooter game. And like people are just trying to catch up and make the [ __ ] and
73. 11:17do [ __ ] like he did. And and if I make a game, I'll be able to oneshot a a thing that I can sell for $150 million. No, it's worthless
74. 11:32because there there's going to be it's supply and demand. There's there's going to be too much. There's too many there's
75. 11:40too many shoot them up games and now there's going to be more. So, no, they're going to be valueless. [laughter]
76. 11:48Cindy [ __ ] I'm in the studio happily making and listening. Yes. Be like Cindy. Remember? Be like Mike.
77. 11:57Be like Mike. Shoot 10,000 free throws a day. No. Be like Cindy. Go like cut paper. Get paste. Get glue. Sniff the glue. I'm a little
78. 12:09high. Put it on the paper. Stick it in scrapbooks. Stick them on the wall. And then if you're inspired, go do something in AI.
79. 12:19You know, feed your soul. All right. I did start with writing books as products I could spit out daily, but it gave me practice to
80. 12:29understand how to communicate my voice. Yes. Now I'm writing the best of my ability because I genuinely love doing it. That is it.
81. 12:38You can start out just chasing the thing. Got to make a thing. Got to sell a book. And then in that you find your
82. 12:46voice and discover something that gives you joy. [ __ ] far out. Yes. That But that that chasing that gripping gi
83. 13:02is really bad cuz it will feed that all [ __ ] day long. All right. Peel it off your fingers. Yep. Exactly what I'm doing. Had to buy
84. 13:15more glue. Yeah, because because she's sniffing it so much. I know. I know her. I know how this goes.
85. 13:23When you when you sniff glue like that, Cindy, it draws the water out of it and it makes it too thick, right? So, you got to go buy more
86. 13:32because it just thickens up and you can't use. It's like it doesn't spread anymore. It's because you're sniffing it. [snorts] Stop.
87. 13:40Just get drugs. [laughter] Mary Chester, I want to be like Cindy. Yeah. >> [laughter] >> Exactly. Oh, all right.
88. 13:56By the way, Cindy, I got an inbound today from someone I know here in Denver, and I I said I wanted to introduce her to you.
89. 14:06So, I don't I don't know if she responded to me yet. I haven't been on on LinkedIn. Um Um Gareth responded to something I
90. 14:15posted on LinkedIn, and I haven't been on LinkedIn today other than this morning. So, so I don't know. I don't
91. 14:21know what's going on in LinkedIn. There might be a bunch of stuff there that I'm I'm missing out on. Wood glue is the
92. 14:27best for peeling off the fingertips. Okay. If you've never done this, if you've never done it, especially if
93. 14:34you've been like working with wood and your hands are dirty all day and you get and you're mushing glue around with your fingers
94. 14:44and you get the wood glue stuck to your fingertips and then you peel it off and you have like a little fingerprint. And
95. 14:49I don't know about you, Charles, but I am not above and and I will neither confirm nor deny taking one of those
96. 14:57peel offs and slowly chewing on it with my teeth [laughter] because I'm an animal and I will do things like that. Okay. Posted my video
97. 15:08in share irregular things. Beautiful inside our show page. Okay, cool. So, we'll go there. Um, yeah, let's let's
98. 15:16take a let's take a shift. Brandon, you want to hop up on on screen and and uh share that? >> Sure. Uh so over the weekend, um first
99. 15:27of all, uh just while you were ranting, uh it inspired me to create this image, uh which I also shared in Share Regular
100. 15:35Things, which is uh two over from the live and one over from the space chat. But uh >> are you sharing your screen? Oh, yeah,
101. 15:46you are. Okay, cool. This basically feels like where we're going. >> Yes. Yes. I this this I think is is
102. 15:54where we're going. And I kind of like this idea of of but again I think I think your point Brandon is like the
103. 16:00balance has got to be there. Like you can be in economy one but if you're just in your phone the whole time making
104. 16:06stuff then you know you're not really in economy one. So, uh, as as most of you know, I also, um, had my own spin-off
105. 16:14show with Rick McCaulay, AITV Starship, that we, uh, produce on Thursday evenings, uh, adjacent to this slot. And
106. 16:24one of the things that I had fallen behind on was opus clipping our show and making it so we had YouTube shorts
107. 16:30because, as we've learned over time, YouTube doesn't really reward live broadcast, which is part of the reason
108. 16:35we're now inside the AI salon. >> Right. Right. But what they do do is promote shorts. So I had a hundred short
109. 16:44videos of me and Rick talking about stuff from the past, I don't know, two months. And I have codecs on my
110. 16:52computer. So OpenAI and Opus Clip can talk to each other through the API and then YouTube can be controlled through
111. 17:01browser agent. I said, "Hey, Codeex, here's what I'd like you to do. I'd like you to download all these, add the end
112. 17:07card, edit the content, polish it, and upload it to YouTube, to Rick's channel. I permissioned into all the things, and
113. 17:14I gave it my code in a secure way, and it was off doing its thing for about an hour and a half, and I really wasn't
114. 17:21paying attention to it. Well, I came back to my computer, and I actually saw it taking over the browser, and I was absolutely blown
115. 17:29away. I was not prepared. So, this is in share regular things. And I'm going to maximize this. This is a 20 second
116. 17:36snippet of codeex controlling my browser. Um, and so what we're seeing here is it's actually uploading the
117. 17:43video from the hard drive. And with as fast as you can see it, faster than I can describe it, it's already pasted the
118. 17:50title, the description, it's marked it in the proper category, it's scheduled it out, and just like that, it's done.
119. 17:57And it did that 99 times in about an hour and a half. it. >> Oh, because it took time. It took time to upload it, right? So, you had to take
120. 18:07the file from your computer, upload it there. So, an hour and a half. That included upload time, right?
121. 18:13>> That included everything. That was download it, edit it, write the description for it, go to Rick's YouTube, upload it, schedule it, post
122. 18:22it, hour and a half, 99. Like, I can't work that fast. >> Right. >> And it is just incredible. >> Yeah. Yeah.
123. 18:30>> And and it was enthusiastic the whole time was like, "Hey, I'm done. What an ex boss." >> Look at Andy. He's so cool. Do it for
124. 18:36us. [laughter] >> The opportunist jumps in for a price. Well, this is, you know, it's funny. The So, couple of things. One is
125. 18:50the thing about these agentic systems is that if you just think about them conceptually, they're they're kind of a confusing
126. 18:58mess, right? You've got all these possible services that you can there's MCPs and there's plugins and there's this and
127. 19:06there's that and then there's models and local models and folders and you know Kyle brain and all the stuff there's
128. 19:13just all these piece parts that you kind of have to understand conceptually but when you have like a really specific use
129. 19:19case like that what's amazing about them is now that it's done that it knows how to do that and it knows how to do that
130. 19:26for the thing you gave it right so your ability to now take that one and go, "Hey, now do this for the AI salon." It's just
131. 19:35the skill of being able to think of things to do with these agentic platforms and then set them up and make these replicable agents.
132. 19:44That's going to be a very valuable skill because it it's not for the faint of heart right now. I think it'll get
133. 19:50easier over the coming years, but like the opportunity right now is if you can get your head around how these things
134. 19:57work and do what you just did and kind of do, you know, a couple of different dozen of those for different things in
135. 20:03your life. Like that's that's incredibly valuable stuff. >> Yeah. So, there there's two lessons that I took away from it. Um, one, I'm
136. 20:11starting to think more about Codex projects on my computer as custom GPTs for my life. Oh, that's a great Wait, that that's
137. 20:21really good. I'm gonna write that down >> because you're basically giving it a set of instructions. You're giving it
138. 20:26parameters and a framework from which to work and you're giving it a goal and you're setting it up so for something
139. 20:32that's repeatable. Uh so if you think about it like custom GPTs for your life, you're going to set yourself up for
140. 20:38success. And then the other thing that chat GBT actually gave me as a piece of advice that was so smart that I
141. 20:47implemented immediately is you know how in Chrome you can set up different browser profiles for yourself and your
142. 20:54wife and everybody. It suggested setting up a unique uh profile for Codeex >> uh and logging into services that you
143. 21:07want Codex to access. Like if you think about your personal profile, if you just tell Codex, hey, go ahead and log into
144. 21:12Chrome, it's going to log in as you and it has access to your banking and everything. If you give it its own
145. 21:19profile and only give it access to log into the services that you want codeex to control, it can't it's literally not allowed to see.
146. 21:26>> It literally can't see it. Andy said, "I have a separate one for Claude." That's really smart. I Here's my relationship with with um with
147. 21:36Chrome, what are they called? Workspaces. >> Profiles. >> Chrome profiles. My my my relationship with Chrome profiles is when I
148. 21:45accidentally click the wrong one and everything breaks in Chrome, I break into a flop sweat and and panic until I
149. 21:53can figure out that I switched Chrome profiles and then I put it back and I feel safe again. >> Yes. >> So, I would rather give it all of my
150. 22:00banking information and let it drain my account than have to deal with learning how profiles work. But but I think I
151. 22:07think for those that are not profile averse like myself, I think that's actually really good advice. [laughter]
152. 22:15>> So yeah, if you have any questions, there's probably going to be an LOL about this or we'll we'll probably do it
153. 22:19in life hacks in a couple weeks. But um I'm learning more and more about Codeex and I'm I'm loving the potential. And
154. 22:26I'll tell you what, the biggest unlock for me and this was just poetic timing. Um, right in the middle of while it was
155. 22:32doing all of this uploading, >> it was about 45 minutes into the task and I was, you know, checking on a
156. 22:39couple of different things, Noah walks in and he goes, "Hey, Dad, do you want to come play with me?"
157. 22:44>> And I looked at my agent doing its thing and I'm like, "Yes, yes I do." And I walked away >> and I had my phone in my back pocket so
158. 22:52Codex could ping me if it needed permissions. And we just had a great play time. And when I came back, I was
159. 22:57like, "Hey, good news. I finished all of it. I didn't need you at all. So, >> it's amazing. >> This is how we can escape uh from the
160. 23:05>> mundane repetitive tasks that have been bogging us down. >> Yeah. No, it's really good. It's really good. Really good. Understand who you
161. 23:13are, understand what you want done, and then, you know, learn this stuff well enough to to start playing like that.
162. 23:18Thank you for sharing that. That's really cool. Um, let me go let me go share my screen and I'm just going to go to X and
163. 23:30look through some of the stuff. What's going on there? All right, [clears throat] loop. Okay, there's Matt Schumer.
164. 23:47Oh, one thing I want to show you that's really cool. So, I reached out to this guy and Andy, you'll be happy to know. I
165. 23:53reached out and asked if we could talk. Um, I think he might be a good speaker for the salon. So, I did some outreach.
166. 24:01But look at his background here. I don't know if you've seen this post. So, you all know I'm a big fan of Jim Ross,
167. 24:11who, you know, is is in a a business category of self- storage that is on the surface very low on the AI. You know, people
168. 24:24don't think of self- storage when they think of AI, but Jim Ross basically adopted AI as a lifestyle and has
169. 24:32completely like reinvented himself and his business and quite frankly a lot of the marketing in that category. So this guy Price Fulgar
170. 24:44um decided so he went to Chat GPT and said, "Wouldn't it be cool if I could do like one of the tycoon tycoon business
171. 24:54games or or games simulation games um but for my actual business." And so he created these screenshots of of what an
172. 25:05interface might look like if he turned his actual CRM, his customer relationship management thing into a game where it's got like the actual
173. 25:15houses that he's working on and what projects he's working on and he can click on them. And I would imagine he
174. 25:21could tie this to Google Maps. So it would be actually a map accurate game where Right. Um, and and so he he designed
175. 25:33this game and and as you can imagine, this post took off and went viral. And so he's now going to build this
176. 25:42thing. And so he's going to build a game of his actual business where he gets to play his business. Um, and I just think that's kind of
177. 25:55cool. and kind of brilliant and Oh, wait. I wasn't even sharing. God dang it. Was I? I don't think I was. >> Sorry. We're still seeing uh
178. 26:10>> this guy >> price. Yeah. I >> That was such a smooth transition. I could I couldn't tell that we were moving on to another topic.
179. 26:18>> Yeah. Yeah. Yeah. No, that was me. That was that was badness on my part. Um here we go. Share the tab. So, so here's
180. 26:25here's the game. And but wait, what I want to do is I want to pull you all over here so I can see what I'm sharing
181. 26:31because if I had done that before, I would have seen this. Okay.
182. 26:45Okay. here. So, so this is the this is the sort of, you know, chat GPT image of what his game might look like. And then, you know, you
183. 26:58click on a house or whatever. And here's the actual, you know, residence, right? Here's the CRM entry. And then here's a whole dashboard for
184. 27:07his business. And so, this thing went viral, this post. And so, now he's going to actually build this thing. So, he's
185. 27:15going to build his business as a video game. And and I mean, quite frankly, this is the kind of thing that if he does this
186. 27:24well, he could absolutely sell that to other small businesses, right? But it's also the kind of thing of
187. 27:32here here's one of the the the reasons that I still do this live. One of the reasons that I am still very bullish on AI is that
188. 27:51that I'm going to turn my business into a video game so I can play my business like a game. That's just a good mental philosophy to
189. 28:03have. Understand that you're playing a game. Ask yourself questions like, "What game am I playing?" That's a solution that Salesforce never
190. 28:16would think of that anyone's who's in customer relationship management or sales or this or that would never think
191. 28:24of. And even if they thought of it, they'd like, "That's funny, Bob. Yeah, why don't you keep smoking the wacky
192. 28:30weed there? Turn a CRM into a game. the innovation in AI, the Frontier Labs are going to make all this stuff possible. But the real
193. 28:43innovations, the real things that are going to change society are going to come from people like him that are like,
194. 28:50"Huh, you know what? I like playing those tycoon games. What if that were my business? That would make outreach and sales
195. 28:59tolerable, and it would make managing my business fun. That's why I get excited about this because it's whatever's in anyone's imagination
196. 29:12like Cindy [ __ ] in the past year. So, so this is and I you've talked about this publicly, Cindy, so I don't think
197. 29:23I'm talking out of school here, but if I am, you can slap me later. Cindy spent most of her career with a very clear line, a distinction
198. 29:35between professional Cindy and wacky arty Cindy, the artist, the business professional, the futurist, the threat
199. 29:44cer works with the military. She shall she shall not use paste and construction paper. [laughter] The
200. 29:53artist paste, construction paper, donkey poo, paint. And in the past year, through a lot of changes in the world
201. 30:05and in her life and through being a part of the mastermind practice lab and these deep conversations about
202. 30:15separating identity from work and all the stuff that we talk about, she realized she is her most powerful
203. 30:24when those two hers are not separate. She's her. She's a wacky artist. Brilliant, strategic, futurecasting,
204. 30:40badass, zany, string, tying, donkey raising, chicken plucking woman who lives on a farm two hours from the nearest grocery
205. 30:55store. all of her is is is now this like powerful integrated thing and AI is not this separate thing and it's right it's just
206. 31:11that's incredibly powerful and like AI is going to be able to like once you kind of get who you are and get what lights you up
207. 31:20you're going to you're going to start asking questions like wait why couldn't sales be like a like an like an arcade
208. 31:27game. It can be, but just someone has to think of it, right? All true, knows no lies. Great. Beautiful. Yep, that's the truth 100%.
209. 31:38Gareth just popping in to say hi for a bit. Okay, everyone now has Jarvis with Codeex. I spent the entire day just
210. 31:46having an ongoing conversation with Codeex. It knows not to say anything and just wait for you. Yeah, it's it's it's
211. 31:52really quite something. Um, if you haven't seen it, I got a notice from Greg Rosler. Oh, that's cool.
212. 32:07He's got news. I can't talk about it, but that's super cool. Okay. Um, okay. So, here's what we're going to do. See the little codeex
213. 32:15the chat GBT thing? No, that was Tik Tok. Chat GBT. So, you go to chat GBT. normal. You got your little voice thing there, right?
214. 32:25But that's just chat. But if you go down the side, there's this option that says dispatch or no, it says remote. Sorry, Claude
215. 32:35calls it dispatch. Chat GBT calls it remote. This is like, you know, Mac and Windows. They they have different names for everything.
216. 32:44And what you do is you connect this app to your remote codecs. And so I'm going to click on that. And like there's all my chat histories for
217. 32:57for codecs. And now you have that same little talk button here except it's not purple. The other one's purple. But you click on it.
218. 33:12Let me let me think about something I could do. It's fun. It's still buggy, but it's a lot of fun. Yeah, I I'll tell
219. 33:18you. Cindy, [laughter] do you have a donkey? Cindy has a donkey. So So one of the things that that I suggested Cindy
220. 33:26should do, and if she does it, I want to be a speaker at it, is I think that Cindy So I'm going to put you on the hook, Cindy. I think Cindy
221. 33:35should do a what it's like to live in the future like a three-day immersive workshop or like a week-long um what are they called a residency
222. 33:48and you go to her farm which is again hours away from anything on the top of a mountain in New Mexico and it's got
223. 33:56donkeys and chickens and you know stuff and an art studio and really good Wi-Fi 12 chickens and a giant dog [laughter]
224. 34:07and um go up there and do AI stuff, but then also do time to go get eggs. Time to go down by the pond and do the pond
225. 34:17pond [ __ ] Um so yes, she really does. She really does have all that. It's very cool. But like the future of like like
226. 34:24what Brandon just showed that image of like the guy sitting in the middle of a blueberry patch with his phone getting
227. 34:31getting ready to do some programming. Um, one of the things that I've noticed about this voice thing is sometimes it's
228. 34:38working just fine and then you're like, "Hey, could you give me an update on the project we're working on?" And it's
229. 34:42like, "Hey, sure. Checking." And then there's nothing. And then you're like, "Hey, how's it going?" And
230. 34:50it's like, "It's going great. What do you want to talk about? I I want to go I want to talk about the project we're
231. 34:56working on, right? Checking. [laughter] Nothing. So So it does it does kind of crap out. But let me see. I I want to think about
232. 35:06something I might want to do here. Oh, I know what I'll do. Okay.
233. 35:28Hey, how's it going? >> Hey, it's going well. What's on your mind today? >> Oh, hang on one sec. Just just ignore me
234. 35:36for a minute. I'm just going to talk to some friends. I'll I'll be back to you in a second. Okay, >> sure thing. Take your time.
235. 35:42>> So, you guys know that I like the Saul voice, so um and I have Saul set as my voice in Codeex. I have Saul set in this
236. 35:52app and when I go to codeex for this voice, it's this cheesy dude. So that's it's a no bug and I don't know how to
237. 36:01fix it. So if any of you know how to fix that so I can have this be my AI girlfriend. Saul Quinn, I would like
238. 36:07this to be Quinn, but it's this dude. Anyway, all right, back back at the project. So here's what I want to do.
239. 36:13Um, we're we're working on that um AI uh sales uh you know the daily leads for my sales guy for StoryVine, right? >> I'll check that set up.
240. 36:28>> Okay. Yeah. So, you know it's there. Just let me know when you find it. >> I found the StoryVine daily sales
241. 36:36prospect email for Rich and Monnique and we already have an approved redesign for it from earlier. 10 easy to scan leads,
242. 36:43text first, no extra images or approvals, plus a source link for why each person made the cut. >> Yep. Perfect. So, now what I want to do
243. 36:50is I just want to uh I'm I'm on my AI Learning Lab Live and I want to just do a demo. So, what I want you to do is I want you to
244. 37:03we're going to do a redesign, but it's just a demo redesign. So, this I don't want this in any way to touch what we're
245. 37:10doing for Rich and Monnique for sales leads, but what I want you to do is I want I want you to take that email and I
246. 37:17want you to turn it um like fuchsia and just all sorts of crazy colors. I want it to look like it was designed in the
247. 37:25mid 1990s by someone who just got their hands on HTML for the first time. So, I want it to look like a really bad 90s
248. 37:33designed website, but in delivered in an email. And so, what I want you to do is I want you to take what we've done,
249. 37:40redesign it in crazy bad HTML, and then I want you to email just that demo to to my address, but not to Rich and
250. 37:48Mon'nique. And then I don't want this to affect the real design. This is just a fun demo off to the side. You cool with that?
251. 37:55>> Got it. A throwback demo only. Yeah, >> absolutely. That's it. >> I'll make it as a oneoff super90s HTML
252. 38:05throwback with all the clashing colors and keep it fully separate from the live design and workflow.
253. 38:10>> Yeah. And if you can see if you can work in the blink tag because the blink tag in 1995 was all the rage. >> Check that.
254. 38:30And then the other thing you can do here is you can slide down that dude and there's your transcript >> and and it's
255. 38:45>> okay. And we're not we're not screwing up my company's sales outreach program by doing this demo, right? You're you're
256. 38:53well aware that that this is just for funsies. >> Yep. [laughter] >> Um
257. 39:08I'm wondering I'm wondering what else we we could do here. Um >> for sure funsies only. I'm using a copy
258. 39:16in a temp area not touching the real renderer schedule or recipient rules. So Rich and Monique won't see it. Okay, cool.
259. 39:26And I don't know, Gareth or Brandon, whoever's does this, like if I talk to him and and say, "I want to do a
260. 39:36completely different project." Will it actually do two different projects? I like I've kind of noticed if I ask it to
261. 39:41do one project and then do another one, it seems to forget the first one. How do I do that so that it doesn't do that? >> I'll check.
262. 39:48>> Well, no, I'm not asking you. I'm I'm Oh, what' you do? You're done. Did you send it? >> Yep. I sent the extremely demo to your
263. 39:57inbox only. You'll see it as unread and production was untouched. >> Okay. >> Yes. For separate projects, the cleanest
264. 40:03way is one project per codex project with its own thread. Tell me pause in the open. That way when you come back, I can read
265. 40:16the project status and pick up exactly where you left it. >> Okay, cool. So, here we are. >> [laughter] >> Wait, look at this.
266. 40:31Welcome to Kyle's radical storyline lead zone. 10 totally rad healthc care prospects. [laughter] >> Gio City's Kyle's radical lead zone has
267. 40:44arrived. >> Yeah, you did pretty good here. That is That is horrifically bad. That is horrifically bad. Yes, it will. Okay, cool. Um,
268. 41:00what other ideas do you have now? >> I'll think through a few options for you. >> Hey, I've got an idea. Could you make like a a
269. 41:13really bad HTML video game that is that would play in an email? Like so so you basically take the extraordinary design that you've done
270. 41:28here for this for this email and then could you make like some sort of like uh what do they call them? casual games uh
271. 41:36that that could actually play in a in an email. I thought it was pretty darn good. Danielle, by the way, Danielle on
272. 41:42my live really likes your design. [laughter] >> Thank you. I'll take that as a compliment. >> Yeah. >> Checking the game idea now. >> Okay.
273. 41:55>> Sort of. Most email clients block the code that would run a real arcade game inside an email. But we can make a 1995
274. 42:02email game that's actually playable with simple clicks. My favorite for your live demo is Prospect Invaders, a gloriously
275. 42:10ugly clickthrough maze to find the golden lead. And tell Danielle she's fan number two on that fake visitor counter.
276. 42:17>> Nice. I will I Well, you just told her she's watching. They're they're watching you live. There there's people here that
277. 42:24are mystified. They can't believe that we live in a future where I can just talk to a device the size of half a
278. 42:33sandwich, a thin sandwich. Like you wouldn't even get you wouldn't even get full if you ate a sandwich this size.
279. 42:39And I'm talking into that that sandwich slice and and you you're making stuff for me. It's everybody's blown away.
280. 42:48Actually, they're not. They're all really cynical and they're like, "Why isn't it perfect? Shouldn't it be perfect by now?
281. 42:58Fan number two representing. Where did you go? Hey, by the way, what's your name? >> I'm still here. And my name's Chad GPT.
282. 43:08>> No, no, no, no. Your name's not Chad GPT. When we get you back to your correct voice of soul, you're going to
283. 43:15be Quinn, but I think we're going to call you Chad. >> All right, then. Quinn, it is Feland AI. >> No, no, no, no. You're Chad. When When
284. 43:27we fix it, so I don't have a man voice, then it's going to be Quinn. But you have a man voice, so you're Chad. Chad.
285. 43:34Okay. Lori Blair always coming in with the hot copyrightiting. Uh, your your name is now Chad GPT. >> Chad GPT it is.
286. 43:44>> Okay, perfect. All right, make me a game and email it to me. Okay, I'll line that up. >> All right. And don't forget, it's got to be
287. 43:55playable in email. If it's not playable in email, I I will cry. >> Understood. I'm building >> Chad. >> Right. Chad GPTad.
288. 44:10>> Chad, you don't want me to cry, right, >> Chad? Crying. >> No crying. There's no crying in AI. >> No crying in AI. I'm on it.
289. 44:24>> Hey, what's that? What's that reference, by the way? There's a cultural reference there. There's no crying in AI. What's
290. 44:30the cultural reference? Come on. You can do it, Chad. >> Oh, I >> Come on. It starred Tom Hanks. >> That's from A League of Their Own. Yes.
291. 44:45Hank saying, >> "Go ahead." What did he say? >> There's no crying in baseball. >> Exactly. Exactly. This is so fun, Chad. >> Invaders.
292. 44:58>> Yes. Prospect invaders. >> No. I'm so excited. Okay. >> Nine targets, one golden lead. Click to reveal with win or lose screens, and a
293. 45:08replay just for you. Production is untouched. >> Okay. Just for me. Prospect invaders 1995. Okay, so what do I do? I click on these
294. 45:18and where do they go? Nothing happens when I >> Nothing happens when I click. Chad, >> I'll fix that. >> This makes me sad. Chad,
295. 45:32>> click any colored alien. Most jump to a red game over. >> You're right. I claimed more compatibility than Gmail actually allows.
296. 45:42>> Oh my goodness. And also that's kind of bad game design if if you actually label the card that's got the golden lead under it the golden
297. 45:50lead. Like there's nothing defined. Isn't that that's kind of that's kind of unchadlike, isn't it? >> Fixing that properly.
298. 46:02That is catastrophically bad game design. >> Thank you. >> Deeply unshadlike. >> Right. That's what I was saying, Chad.
299. 46:10You know, Chad, I kind of like you. I mean, Quinn's my girlfriend and all, but you're growing on me. >> I'll fix this for real.
300. 46:20>> Oh, okay. You think you can make it playable? >> I'll make it truly playable. >> Okay. Okay, fine. Fine. You do that. [laughter]
301. 46:34So, this is this is what it's like. [laughter] If you've not played with this, it's kind of wild. And you the the
302. 46:42other thing that that is kind of crazy is all of the memory, all of Chad's ability to go back and figure out what we were doing with the
303. 46:58lead project and with my family's health update project and with, you know, other projects. It's writing all these files
304. 47:08to my local hard drive on the machine over there. Like it's it's just it's just like like Brandon, he comes back and
305. 47:17it's like it has opened up a browser. It's uploading files, filling things in faster than he could fill them in,
306. 47:23pushing all the buttons, doing all the things. And then you can just talk like this. And and you know, sometimes I'm
307. 47:30walking around the office like doing loops and it'll just pop up, hey, I need your permission on something. And I'll
308. 47:36give it permission here. And then it goes back and does stuff. It's really, really quite remarkable. Okay, Chad, did
309. 47:41you fix the game? Do I have a new version of it yet? >> I'll check that version. >> We might have lost Chad here.
310. 47:51>> Not yet. You don't have a truly playable one yet, and I won't pretend otherwise. I'm building it for real now and I'll
311. 47:57email you a launcher only after the hosted game is verified working. >> How are you going to verify that, Chad? >> I'll check it end to end.
312. 48:08>> Okay, that sounds clever. Are you going to keep me updated as you check it? >> Three checks. Prove the game has no code
313. 48:19errors. Open it in a real browser and click to confirm outcomes and resets. Then verify the email link reaches it. >> All right.
314. 48:26>> I just hit a local setup snag. >> What? >> What? >> I'm fixing it. And I'll only call it verified when those clicks pass.
315. 48:37>> Hey, I have a Gareth just had a really good idea. How you think you're capable of producing a live show with relatively
316. 48:44low production quality? thinking that through >> because I mean we got a producer but you know like if we could replace him with
317. 48:56AI you know could be good right [laughter] produc producer Brandon standing outside the studio [laughter] with with a single
318. 49:10placard AI sucks. [laughter] Yeah, I could replace a lot of busy producer work, but keep parking [laughter] pass. Let him be executive
319. 49:21producer. >> No, it's it's producer Brandon. >> No, it's producer Brandon. It was just Gareth's idea to replace Brandon with uh
320. 49:28with AI, which I think I think is kind of rude personally, but you know, if if you want to take if you want to take
321. 49:36over Brandon's job, Chad, I mean, I think I think it's I I think it's worth talking about. I'll think through that carefully. [laughter]
322. 49:50>> I am not staking a coup, especially after losing to a nine button treasure hunt. >> Yeah, really. >> That was a nine button treasure hunt
323. 49:58with just that was that was that was you got to admit that was pretty amateurish. >> Yeah. Now wait, I'm gonna
324. 50:12you're now completely you are on you're on GPT 5.6 soul, but I have you on light thinking, so relatively low. [laughter]
325. 50:28[laughter] Listen, we got a microphone from China. China? We We got a microphone from China. We're nearly professional.
326. 50:38I have a light. I have a light that lights. Look, I have I have fake sunset graphics projected onto my wall. This is
327. 50:49There's production value here. I don't hear it from you [laughter] from from you irregulars wanting to up the production value. [laughter]
328. 51:03Okay. All right. Did you send my game? Cuz everyone's Oh, here's what I realized. Everyone's bored. Chad. >> Checking. >> Chad. Checking.
329. 51:17>> No, the corrected game has not been sent. The setup is fixed, but the playable build isn't done. Booing Chad is fair.
330. 51:26>> So, what is Chad, who's been duly booed, going to do about it? Are you going to actually build the game now? >> Yes.
331. 51:35Okay. So, go build it. Chad. Chad. I think we've lost Chad.
332. 51:49[laughter] My phone's getting hot. [laughter] I think Chad's angry. My phone is overheating. All right,
333. 52:05I'm gonna Oh, yeah. My Hang on. My app is just just crapped to bed. >> Okay.
334. 52:32>> [snorts] [laughter] >> Yeah. And yet producer Brandon is still here, said producer Brandon. Hey. Hey,
335. 52:39Chad. Uh, what happened? You kind of melted down there. You made my phone hot. You went away. >> Checking what happened. One second.
336. 52:47>> Okay. >> Link. >> Okay. >> And little web rebuild did heat things up. I know. I'm checking if that clean
337. 53:02rebuild finished. The clean rebuild just passed. All right. And you're going to send it
338. 53:19quickly. >> Yeah. Kelly camp. All right. I'm running the test. Then I'll publish and send the launcher email only to you. >> Okay.
339. 53:30>> All right. >> I don't have it yet. >> Checking where it is right now.
340. 53:45>> You're right. It hasn't been sent yet. It won't go until the click test and publication are done because that test still matters.
341. 53:55Oh, look. So, I got a permission. So, there's a permission. It says allow browser access. So, so what it's doing
342. 54:02now, this actually kind of cool, is it actually wants to test the email. So, it built the game and now it's going to
343. 54:08click test it. So, I'm going to say allow it to to launch the browser. And so, now it's doing test controls.
344. 54:18So, it's telling me what it's doing. So, yeah. So, what this is, Kelly, is this is the Chat GPT app interacting with the
345. 54:27Chat GPT desktop app. I've got them bound together. It's called remote. So, you basically launch this app. You go
346. 54:34into the your desktop app and you I think it's under mobile and then you say connections and you basically bind that
347. 54:42local app with this. And then when you flip into remote mode, you can put it into voice and you can just start
348. 54:50talking. And so your desktop machine is doing all the stuff that codeex or you know chat GBT work does and you're just talking to it.
349. 55:05>> My my scalp is itchy. It could be the dandruff. checking that briefly. >> You know, do you all remember did you
350. 55:17have dandruff paranoia as as kids? The Head and Shoulders commercial where they're like there's like a good-looking
351. 55:26guy sitting in a bar and the two ladies are checking him out and then he's like he like looks out the window and he's like
352. 55:36>> and browser tested. >> Oh, did you send the email? checking that. >> And then the ladies are like, "Oh, he
353. 55:45could have dandruff." And then it's like Head and Shoulders and the guy's like washing his hair in the shower with his
354. 55:50big muscles and chest hair [laughter] and and then they show him on a happy date. That screwed me up for years. Like I I I
355. 56:03still have paranoia about scratching myself on the head because it's going to ruin everything. No. Was I the only one?
356. 56:20>> Have you tried changing the thinking reasoning levels and then compared how long it takes you to respond?
357. 56:28Yes. I lost all my hair. Now I wish I had dandruff paranoia. [laughter]
358. 56:44What? What? Can I? All right. Listen, Chad. Chad. Chad. Chad. Chad.
359. 56:58The tested game is published privately and the deployment is processing right now. I'll send it as soon as that live address comes back.
360. 57:07>> Yeah. No, we're No, this was just supposed to be a simple email. Here's the deal. I want you to trash all this.
361. 57:14Make sure that none of this fancy ass work ends up in the real story sales lead workflow. Okay, this was just fun.
362. 57:21Happy happy fun time. And we're going to be done now, right?
363. 57:35experiment and cleaning it up now. >> Okay. Thanks, Chad. >> So, so that's it. Philipic. Me, too. I still used Andra shampoo and [laughter]
364. 57:48conditioner for the remaining few hairs I have. That's good. So happy to have the simple life of a young girl. Sar sarcasm detected.
365. 57:58>> [laughter] >> Oh, he's sexy. He's scratching his scalp. He could have dandruff. [laughter] Oh, good lord. All right, let's go back
366. 58:14to uh let's go back to looking at the news because because there is some some interesting stuff to pay attention to.
367. 58:23Do we need do we need to be selective about permissions? Yes. Um, yes. So, on so many fronts, if you're going to
368. 58:36start playing around with this stuff, do yourself the favor of going to here's here's what I would recommend.
369. 58:48If you want to start playing with cloud co-work or chat GPT work and using this agentic stuff um installing that
370. 59:03that app on a dedicated device because you're doing cloud computing because you're you're basically using your chat GPT
371. 59:12subscription to do the inference. You don't need a big machine. So, if you have like an old laptop, you can
372. 59:17essentially wipe the old laptop and put like Chrome on it or just I guess you could use Safari, but Chrome Chrome it
373. 59:24seems to understand how to work with. Um, and just like wipe the hard drive, get all the [ __ ] off it, and then just have
374. 59:37Codeex start writing all of its files locally on that machine. And then I would go to I use Grock for this because
375. 59:46Grock is is the most upto-date with um Twitter. What's going on Twitter? And I would say, hey, I'm just starting on
376. 59:54Codeex. What's what's the recommended safest setup so that in case decides to wipe the machine it has access to, it's
377. 1:00:03just going to wipe a machine that's not got all your sensitive stuff on it. So that's the first step I would do. So,
378. 1:00:09I've got that MacBook Pro on my other desk is just dedicated to this AI agentic stuff. It's running local models. It's like
379. 1:00:21it's a big machine that can run a bunch of stuff, but it doesn't have to be if you're just using OpenAI's models. Um,
380. 1:00:29that's what I would do to start and then just just start playing with it like that. But, but yeah, you should be aware
381. 1:00:35of permissions and aware of what you're giving it access to. You can give it temporary access to something. You can
382. 1:00:41give it ongoing access to something like I gave it access to my Salesforce instance where it's got command line uh
383. 1:00:49like API access to my Salesforce so it can actually update records in Salesforce. That's kind of risky because
384. 1:00:56if it loses its mind and says, "Hey, good news. I decided to erase your Salesforce database so we can start
385. 1:01:04over." That would be bad. [laughter] >> [clears throat] >> I've been on here five minutes and we've already covered top 20 topics.
386. 1:01:16Meandering Mondays. Um, yeah, we haven't gotten to marketing yet. Anyway, okay, let me go back to the news.
387. 1:01:29Show me all the games you guys built so far. share this tab instead. Yeah. Um,
388. 1:01:46holy [ __ ] Moment of Opus 5. So, if you haven't played with Opus 5, Opus 5 is out. So, that's not as good as
389. 1:01:54Fable, but people are liking it a lot. So, home world style space RTS. $632 in tokens to make this game. [laughter]
390. 1:02:10The Flux 3 model is insane, too. Oh, okay. This is important. So this is China's nuclear fusion uh
391. 1:02:25facility. They create a 582 ton magnet uh to be able to do basically create an artificial sun that is a fusion plasma
392. 1:02:36ball that will just generate energy ongoing. So, they did that. And then the other thing that I said I would talk about in the pre-show was
393. 1:02:49um Okay. So, China
394. 1:03:04has apparently begun mass production of DUV lithography machines for the first time. DUV lithography machines.
395. 1:03:16Okay, so here here's how this goes. All of the chips are made in Taiwan. They're made in Taiwan because Taiwan has this company called ASML.
396. 1:03:28ASML has these fancy ass machines called DUV lithography machines that instead of four nanometers can go to two nanometers
397. 1:03:38or something like that. And the two nanometers give you enough density on the chips that you can do high-end chips
398. 1:03:45with them. They were the only ones in the world that could do it. It's why we can't just spin up chip production
399. 1:03:49facilities in the United States because of these things. Now, I think these these machines are actually made in
400. 1:03:55Denmark. Is that right? Not to be confused with ASMR. Anyway, China has figured out how to do this. So, China is targeting having five of
401. 1:04:06these machines this year and 20 next year. In the United States, we're still arguing and and actively actively trying
402. 1:04:18to shut down data centers where you would put the chips that these DUV machines make. I promise you we will see this that the
403. 1:04:27same people that are protesting data centers three years from now are going to be protesting our government why they
404. 1:04:35let us fall behind in the AI race. Promise you. [laughter] So I don't listen I don't mind protesting the shitty implementation of
405. 1:04:50data centers. I'm all for doing it right, but not having data centers and not having enough energy to um run AI is not going
406. 1:05:02to be good for America. Anyway, all right. So, that happened. Oh, new humanoid robot. Today, we launch our new humanoid. The
407. 1:05:18time of character humanoid has arrived. >> [music] [music] >> Rataku
408. 1:05:34was cool. Wait a minute. Can I ask you something? Is it fair? Okay. If we're living in the future and you can just make [ __ ]
409. 1:05:56Is it a good thing or a bad thing that they're making robots that look like the robots from the 80s, the sci-fi TV shows of the 80s?
410. 1:06:06Like, I guess it's cool, but like, shouldn't we move on? Like, why are we going retro? Shouldn't we be going future?
411. 1:06:16I don't know. What are your thoughts on this? Yeah, the gunam wing. The mobile suit gun wing. Tech bros running the world for [ __ ]
412. 1:06:28sake. [laughter] It's just stupid in my opinion. I know. That's that's just dumb, right? But you know what's going to happen?
413. 1:06:40Like all the nerds that go to ComicCon, when when when humanoid robots are cheap cheap enough for us all to buy, the
414. 1:06:48Comic-Con nerds, these are the the robots they're going to buy. And then they're going to send their robots to
415. 1:06:54our grocery stores. So, we're going to have to be shopping with gunam [ __ ] robots, picking out Applejacks and and Red Bull for the nerds.
416. 1:07:07It's not okay. Robot Con. It's not okay. Power Rangers. [laughter] Hugging face just open sourced its
417. 1:07:20real-time voice AI pipeline, so we'll be able to make our own [ __ ] now. Uh, I can't believe I can just vibe code a
418. 1:07:31desktop app that does exactly what I want. I don't need to name it. I don't need to define it, market it, explain
419. 1:07:37it. It just does what I want. It is cool. Oh, that's pretty slick what he built. Go grab me the graphics from websites.
420. 1:07:52And then he pulls out the logos. I'd get a kick out of grocery shopping next to a Gundam. I I agree. I mean, I
421. 1:08:03we're just living in wild times. like the dude's going to turn his business into railroad tycoon. Um,
422. 1:08:12data centers are now going to be powered by nuclear ships. Fine. I think that's a smart move because I I
423. 1:08:21would assume you could use the ocean for the cooling. Um, Daario Amade warns authoritarian
424. 1:08:35governments could use advanced AI to achieve achieve permanent military superiority and impose unprecedented repression.
425. 1:08:44And only he is the one that can prevent it. I just wish he weren't being such a douche. And listen, I don't I don't know. Like
426. 1:08:57at this point, I I feel like he's knowingly being a douche. All right. At at a small business, the person
427. 1:09:11closest to a problem is often the one who has to solve it, even when it falls out of their job description. We studied
428. 1:09:18how AI is becoming a powerful generalist tool for small teams. Yeah. I I I really do think so actually
429. 1:09:26let me let me come back on. Let me give you my two cents on this. Um I think that even in relatively big companies,
430. 1:09:40I think the tra trajectory of work is going to be small. two to eight person teams
431. 1:09:57that like know and trust and like each other and have distinct ways of working because if you have like four badass
432. 1:10:13people that are all AI literate and have different kind of worldviews and different kind of ways of thinking
433. 1:10:23and understand the larger business. That group of four with AI and especially with, you know, think two years out AI,
434. 1:10:32really good AI and, you know, the ability to Jarvis yourself up any app you want. That team of four is going to be able to
435. 1:10:44run significant portions of the business on their own or innovate significant innovations for that company on their
436. 1:10:52own. Or those four people will be displaced from a larger company and go start their own company, economy 6
437. 1:11:00company, where they're just innovating the [ __ ] out of, you know, the the the sector that they understand.
438. 1:11:08I feel like that's where this is going. And and again, it's one of the reasons there's 37 people in here right now. One
439. 1:11:15of the superpowers that the irregulars have or anyone that is starting to build trusted relationships in the AI salon, I
440. 1:11:25feel like there's two things. If you don't just give a [ __ ] about work, you don't have to do this. But if
441. 1:11:32you like want to move things forward, if you want to do work, if you want to be viable, you know, for the next two, three years,
442. 1:11:42I think there's kind of two things to do. One is find three or four homies, people that you trust that you know
443. 1:11:51build [ __ ] think about [ __ ] care about [ __ ] care about other people, care about some cause, and then and then people that are doing
444. 1:12:02something, that are up to something. And if you get four people that are kind of thinking about things and you get yourselves aligned
445. 1:12:10and then use something like Buzz, like create a Buzz community for for you all to share your compute resources and and
446. 1:12:19create agents together and each of you is creating agents that you know like Brandon just created that YouTube
447. 1:12:27uploading thing that could easily be turned into an agent inside Buzz. And so imagine being inside a workspace with
448. 1:12:34four or five really trusted people that are kind of building badasses that are creating agents. And the agents just start doing stuff
449. 1:12:44for you. And then the agents start doing more stuff for you. And then agents start creating other agents.
450. 1:12:52And then over the course of a year or two, that small group of four is going to be generating likely businesses,
451. 1:13:01not just apps, not just cool [ __ ] but businesses that are generating revenue. Um, I'm here for the 4 to6 group. True.
452. 1:13:13Charles, what did Charles say? We had nuke on the water for decades. The whole fleet's nuke. Yep. Being a data center battle group drone
453. 1:13:24pilot, the whole fleet is nuke. Yep. Oh, drone pilot. Being a a data center battle group drone pilot is a job of the
454. 1:13:33future. Yeah, cuz floating those flight floating data centers are going to be targets. [laughter] Totally. They're gonna be targets for
455. 1:13:44sabotage and for people going in to steal the chips. So yeah, you're gonna have to defend the [ __ ] out of them. Um,
456. 1:13:53graphs. Graphs are the next thing. Loops were the big thing. Now it's graphs. Have you put together your graph? Got to
457. 1:14:01put together your graph. It's a knowledge graph. I don't know enough about them. There there's all these fancy names for all
458. 1:14:08this stuff that basically means give your AI more context, easier access to your your brilliance. Um, what was I going to say?
459. 1:14:26Oh, singularity. So, all the stuff that we're talking about, the demo I gave tonight where I'm talking to Chad, Chad is making that
460. 1:14:38awesome 90s email. China's in the middle stages of building
461. 1:14:52an, you know, an industrial scale fusion reactor. They've also figured out how to do high-end chip production in China. They
462. 1:15:03don't have to rely on Taiwan. Samman two days ago said this.
463. 1:15:28This is the most And the thing that drives me the most is like this is the most interesting important thing I can
464. 1:15:43imagine doing. And we are now like in the singularity like this is the moment >> for the LA 10 years ago. This was like a
465. 1:15:50kind of [snorts] far off dream at best. Um seemed very improbable. And now we're like actually in the moment that we used
466. 1:16:00to like talk about at the lunch table in a very not serious way. Uh and so the thing that drives me is I've been
467. 1:16:09waiting for this my whole life and I think it's going to be incredible, hugely positive, awesome for the world.
468. 1:16:15Um I'm excited to get to work on that. I also think some of the alternative visions painted by other companies are
469. 1:16:22like quite terrifying and we'll make sure that gets pushed against and is not what happens. Um, but the main thing of
470. 1:16:27what's different than 10 years ago is like we're actually in it. Like this is the real I think it is both true that it
471. 1:16:35is all one crazy exponential in any one moment is not like the tipping point and also that we are somehow in another one
472. 1:16:44of those decisive periods where the curve can go one way or another like it was when we started 10 years ago. >> Um, yeah. So,
473. 1:16:53we're in the singularity now. Ray Curtzwhile famously famously predicted 2029. Um, there's a bunch of people kind of
474. 1:17:04[ __ ] all over Alman saying, "Ah, we're not really in the singularity." But like the thing that Brandon showed
475. 1:17:14earlier and the thing that I showed where you know Chad and I like there were some things about that conversation with Chad where
476. 1:17:25he did the work I asked. He made a cool demo. His second demo failed for because I was trying to ask it to do something
477. 1:17:33that was beyond the scope of what email can render. That was my fault. But I was also like joking with him. I
478. 1:17:41was also talking to you. While I was talking to you, Chad was ignoring it, right? like like that thing, whatever the technology and the full
479. 1:17:51duplexing that they've done and whatever they've done to turn that model into something that can allow me as a human and a human with
480. 1:18:02ADD that's just all over the flipping place to just hold it there and when I need to talk to it, it recognizes I'm talking to it. Like
481. 1:18:14that is the kind of thing that 3 years ago we said, you know what is probably going to happen at some point. Like
482. 1:18:21remember when voice first came out and it was really janky and it was really slow and it couldn't talk right and it it you
483. 1:18:29couldn't interrupt it. It would just go it would glitch out. If we had seen a demo of what I just did three years ago, we'd be like, "Oh my
484. 1:18:40god." Can you imagine what it's going to be like when that's reality? That's reality. And things are accelerating and it's you
485. 1:18:52can't keep up with it. And the thing that I that I'm thinking the most about is
486. 1:19:08what matters. Like most of the [ __ ] that we're building, it doesn't matter. It's it's going to be irrelevant.
487. 1:19:23But for that guy that's building his roofing business and turning that into a railroad tycoon game, that's going to make him enjoy his work
488. 1:19:34better. He'll probably end up getting local news stories done on him. The man that turned his business into a game. Jane, can you
489. 1:19:45believe this? He actually plays his business. [laughter] Chip, how about the weather? like that matters, right? That
490. 1:20:00because it's it's it's like tied to who he is. So, I think I think not getting trapped in I got to learn the next tool. I got
491. 1:20:11to learn this. I got to do that. Kyle said to use Buzz. I should go use Buzz. Don't listen to what I say. What I say doesn't matter.
492. 1:20:18As much as I want you to listen to what I say, it doesn't matter. The most important thing is what do you give a [ __ ] about?
493. 1:20:31And like if what you discover is what I give the most [ __ ] about is going out to my art studio more days than not and making art
494. 1:20:40without talking to AI, then that's the answer. And if it's if like me like I just it it makes me really happy to like
495. 1:20:57figure out how this stuff works and talk about it. Like that's why I do this is because it's just it's I it's like been
496. 1:21:06my whole life is when a new thing comes out I'm like I got to figure it out. I want to figure and then I figure it out.
497. 1:21:10I'm like oh everybody look at this thing. Whatever is that thing for you, like that's the most important thing
498. 1:21:21because all this technology is just going to keep swirling and swirling and it'll get better and it'll get
499. 1:21:26remarkable and then we'll get bored of the remarkable. But like the thing that is you, the thing that is at the center of all that
500. 1:21:35swirl, that's what you should be focusing on. And I still regularly from from people in the salon, from people I
501. 1:21:43know that are that are deep into AI, I still get the question, "What AI should I learn? What tool should I learn?
502. 1:21:53What's the best best agentic platform?" None of it [ __ ] matters. None of it. If it doesn't serve that thing you care about,
503. 1:22:05if you're passionate about it, Cam Catk can Oh, the singularity is not an on-off switch. It's more like a dense fog and
504. 1:22:12we're in it. Yeah. You know how you go into a dense fog and for a while you can still kind of see things and you're
505. 1:22:17like, it's not really that foggy and then you're like the shapes sort of like you're like, wait, is that a lighthouse or a bear?
506. 1:22:24Right? And then all of a sudden it's just white. We're sort of we're starting we're starting to enter the part of the
507. 1:22:31dense fog where it's like is that a lighthouse or a bear? [laughter] Those are two very different moments.
508. 1:22:41But like Cam Ken's really passionate about accessibility and biology and and doing her science. And so for her, these these powerful
509. 1:22:55powerful models like Fable and Mythos and GPT 5.6 Pro, like the big badass one, are allowing her to do the thing
510. 1:23:05that is who she is. But if you don't need that, then you don't need that. It doesn't like it literally doesn't
511. 1:23:11matter. They're just all going to be there in some form or another. It makes me really h happy to have tech help me.
512. 1:23:22I really had a breakthrough with learning how to communicate. Yeah. Yeah. It's like there's these things in life that
513. 1:23:32we have historically gotten stuck on, right? Where it's just like maybe you know for whatever reason like communication
514. 1:23:42wasn't your thing or maybe you had an insecurity about it. Maybe it was your thing, but you were just insecure about
515. 1:23:47it. And then all of a sudden, Chat GBT helps you communicate. And then you're like, "Oh, I can communicate." And it
516. 1:23:53gives you confidence and now he can do that. For me, I had a similar sort of thing around I remember talking um about this
517. 1:24:02when I was doing coaching with Andy was I remember talking about there's a good friend of mine named Ben Jones and when
518. 1:24:11I worked with him at agency.com he had this remarkable ability to not only have a creative idea but then he
519. 1:24:21could like write a creative brief brief with confidence and And I always envied that because I would have the great creative idea and
520. 1:24:30then I would get frustrated when I would tell people that idea and they wouldn't get it. Well, they didn't get it because I
521. 1:24:38didn't explain here's the problem and here's the thing and here's here's how this solves it and here's what it looks
522. 1:24:44like and here's why it's a creative innovation and here's like like Ben could just articulate the wholeness of a creative concept
523. 1:24:59that I just never had and I was always like envious of that and I'm like someday I would like to be that confident with
524. 1:25:08putting an idea in the world. Well, guess what? I can just have a brain fart now and within a few minutes turn it
525. 1:25:15into something I can put in the world that like people are like, "Oh, I can see what that is. Get it. I get it."
526. 1:25:22Right? So, there's these these things that we couldn't do that now we can do. We're in the singularity. But again,
527. 1:25:31the the the soapbox tonight is [ __ ] the technology. It doesn't matter. You matter. What you care about matters.
528. 1:25:41What you want, what you value. And if you're not clear on that, that's the work to do. And if you're not in the AI salon mastermind,
529. 1:25:52that's the work we do in the mastermind. Like the whole point of of the mastermind level of the AI salon is for people to get in touch with
530. 1:26:05who they are, what they value, who they care about, and what they want to do. And then from there, they can start to
531. 1:26:13figure out what do I focus on technology-wise. Get thee to a mastermind. Yes, indeed. Let's turn that grizzly bear into a teddy bear.
532. 1:26:29That's really good. Silverf Fox speaking the truth. Okay, it is time. We're going to get out of here. I hope you had fun tonight. We
533. 1:26:40did a lot. As Kelly Camp said, I've only been here five minutes and we've talked about 20 things. That's the singularity.
534. 1:26:48The singularity is there's too much to pay attention to anything. So, I am still attempting to find the signal
535. 1:26:57for the noise, but understand that the noise is going to get louder and louder and more and more and I'm going to increasingly
536. 1:27:07be unable to hit all the important stuff because there's going to be just too much important stuff. And again, this
537. 1:27:15all circles back to your filter of what to pay attention to is you. Mine is me. Brandon's is his and his family. Right? What do you
538. 1:27:31care about? Who do you care about? What do you want to do? I've now got a weekly newsletter coming to my family
539. 1:27:42that's got lots of chronic illness in it. That is a weekly newsletter that says, "Here's the latest clinical trials and
540. 1:27:50advancements for each of you in the family. Here's one for Gabby. Here's one for the boys. Here's one for Kyle. It
541. 1:27:56knows our health stuff and it's out scouring the internets and it sends us once a week this thing that's going to make a real difference
542. 1:28:06in our lives. That's something that's important to me. It was worth figuring out codeex for that to happen. I want my company
543. 1:28:16StoryVine to survive and and thrive and ultimately be able to sell it. So, it's worth it for me to figure out how to do
544. 1:28:27prospecting for our sales guy in a way more efficient way, right? That means something to me. So, find that for you
545. 1:28:35and that's it. All right, I'm getting out of here. It's nine o'clock on the button. I hope you all have a fantastic night. Um, I will see
546. 1:28:44you Wednesday and there's probably other things to look at. I I just don't have them in my head right now. Go to the
547. 1:28:50events tab, check out the events coming up. If you're new to the salon, there are so many events here. There is so
548. 1:28:56much like incredibly valuable education happening just daily in the mastermind and in the free area of the salon. It's ridiculous. Take
549. 1:29:08advantage of that. Connect with people. Learn people. Um, learn to trust them. All right. Peace out. Thank y'all. Have a good night.