7/17/2026 - Understanding the Confusing Difference Between ChatGPT Web and Desktop Projects

Transcript

1. 0:01[music] [music] [music]
2. 0:18[music] [music]
3. 0:32Happy Friday night, date night. Welcome to the AI learning lab in the AI salon. If you are a fan of the pre-show and you're
4. 0:41like, I couldn't find the pre-show tonight. That's because I wasn't there. There was a domestic emergency.
5. 0:48Everything's fine. We're good. Disaster averted. But uh I'm just coming in coming in hot. Hang on. Clean off schmootzy lens.
6. 1:01Beautiful. How is everybody? What is happening? What's shaking? What's going down? All right. So, welcome. This is
7. 1:07the AI learning lab. It's Friday night date night. It's an ask me anything. If you have questions, ask them in the
8. 1:12comments. How do you get the comments open? Up in this corner right here is the comments drawer. Drawer tab window. Danielle knows how to
9. 1:22use it. Look at this. Look, she made a comment. The comment shows up here. It's like magic. So, you can do that, too. But you just
10. 1:32click up there. That'll pop out that. The other thing you can do if you want to hang out in the AI salon and chat or,
11. 1:40you know, upload pictures, share pictures that you've made, whatever it is. Um, in this lower corner, you can
12. 1:46spin this video video out as a popup window and then you can open a second tab in your browser and go back to the
13. 1:54AI salon and you can surf the AI salon while you're watching the live stream. All right, so that's little housekeeping. Okay,
14. 2:04so you weren't on TikTok before this. That's correct. Yeah, I didn't I didn't do a pre-show tonight um just because I
15. 2:09had there was a minor emergency here at home. I had to deal with some stuff. Uh, so I couldn't go live. Um, but here we
16. 2:16are. We're good. Um, so think about any questions that you might have. If you were here Wednesday, you know
17. 2:24that I kind of lost my [ __ ] over open AI. [laughter] And and if you were here and you're also like,
18. 2:35I'm not quite sure what he's so mad at. What I was mad at was that the way they rolled out their service, it was so confusing
19. 2:47and so potentially destructive to someone doing work in this AI stuff um that it just pissed me off. And yesterday
20. 3:02they fixed it. Like the major thing that I was bitching about was just how they were handling memory and previous chats in all of their
21. 3:14different products. So, they basically updated all of their products and then all three of their products handled
22. 3:20memory and chats in three different ways where it was completely unclear. And so, the reason I got crankier and crankier
23. 3:28is I was trying to figure out how it actually worked and I couldn't because it didn't make any sense. So apparently,
24. 3:37so one of two things happened. Sam Alman watches this live stream. [laughter] Hey Linda Sherman, what's happening? Apparently
25. 3:47Oh, can't do all that on an iPhone. I could have done that on an iPhone, Danielle. Come to think of it, I should
26. 3:53I should have done the pre-show because I was actually running around to to a store and there was all sorts of things.
27. 3:59It could have been a fun pre-show. Um but but um either Sam Alman or Greg Brockman listened to the show and they
28. 4:08saw how cranky I got and they decided to fix it or I wasn't the only one that told them [laughter] that what they had
29. 4:16created was horrible. So I actually want to show you this is an ask me anything night. Oh, and I got to put up my
30. 4:23post-it notes. Hang on a second. Let me because here's what's going to happen. Producer Brandon's gonna he's gonna put up a thing
31. 4:31And then, oh boy, dude. Oh god, Chrome just changed their interface. [laughter] My god. Really? Really? Chrome? Really?
32. 4:46Good lord. Okay, hang on. Hold, please. [sighs] [laughter] I just I just I can't I just can't. Okay. Um,
33. 5:04everybody's changing their interface. And listen, my my co-founder Mon'nique, my co-founder of StoryVine,
34. 5:12she gets epoplectic when anyone changes any interface. I'm normally fine with it. Like, I'm normally like, "Change it up. I don't care."
35. 5:23But, man, these these recent changes. Okay, so here's what we're going to do. I'm going to share my screen.
36. 5:29Yeah, [laughter] you you have two seconds to learn it. It would indeed be cool if if OpenAI exe execs were
37. 5:36watching this show. It would be cool. [laughter] I don't think it's happening, but hey, you just you don't know.
38. 5:47Probably somewhere in the stats, we do know. Um, okay. So, let me let me share my screen and then we're going to go look at some
39. 5:56stuff. So, I will I will give you the update. I I shared this I talked about this on office hours, but I'll show you here tonight.
40. 6:07And if you have any questions, feel free to pop those questions in the chat and Brandon will gather them like little nuts. I'm gonna
41. 6:18pop the chat out so I can see you all as I'm sharing things. And then hopefully my share doesn't go away. Okay,
42. 6:30we're going to go to chat GPT. We go new chat. Okay,
43. 6:45Harry. Hey, Mary Chester. Okay, so there's that. Then I'm going to go here. I'm going to log into my remote MacBook Pro.
44. 6:59Do new chat. Get rid of that. Come here. There's my little dude.
45. 7:26Okay. Okay. [laughter] So, okay. Wait, wait. What? Working a project. Okay.
46. 7:42They've changed this interface since this afternoon. [laughter] Okay. But but they've actually made it better. Okay. Okay.
47. 7:54So, the way that you can tell right now, we're in the chat GPT app. So, it's the white screen, right? The white screen
48. 8:06with the little animated me. This is that other computer. And we're running the chat GPT app, which is has
49. 8:15now got, if you notice at the top here, it's got chat mode and work mode. which if I go over to here, my projects don't show up in my
50. 8:28Mac still. Any news on that? I think so. I'm going to show you. I But like I literally just noticed that that that they changed
51. 8:40it. It It's behaving differently than it did this like three hours ago. [laughter] So, so or or I I discovered something
52. 8:51that I didn't see three hours ago. I'm live build and my pet Oh, my pet is killing how they change it. Okay,
53. 9:05so the the dark interface here, this is chatgpt.com, right? So, this is us in chatgpt. If you click in the upper leftand corner
54. 9:17here, nothing happens. At the top center of this interface, you've got chat and you've got work. Work are
55. 9:30agents that can do multiple tasks. You sort of tell it, I want you to build me a website and it will do multiple tasks. Okay.
56. 9:48Then we flip over to the chat GPT app. We still have chat. We still have work. Okay. If I scroll down the left hand side
57. 10:02here. Okay. These projects, Cosmic Rocks Revenge, Project Dashboard,
58. 10:18Storyvine Pharma Sales Outreach, these are projects. These folders actually exist on my local hard drive.
59. 10:32Okay, if I go back to chat GPT and I look at the projects,
60. 10:48wait, let me close these out. Show more. Okay, these projects project dashboard image consistency Sydney feed your
61. 11:00prompt book 10person team workshop mastermind practice lab vibe core core these are chat GPT projects.
62. 11:11So chat GPT projects were a way in chat GPT to to gang up multiple chats into a into a single context. Right? So if I go to
63. 11:23the Sydney project, here's all the different chats that were in it. Right? So, if I click on this, if I click on
64. 11:30this and then go into here, notice at the top here's Sydney and I'm in this chat, which I don't know what this chat is, but whatever. It's
65. 11:42the it's whatever the latest. Oh, Sydney as resort IP. Okay. So, so this is a chat and if I if I chat in this project,
66. 11:52then it's going to pull in the context of any of these other chats that are in the project or any files that I upload to it. Okay.
67. 12:03The big distinction here is that at chatgpt.com, all of your folders and all of your work and all of the files that you upload and
68. 12:18all of your chats are stored on OpenAI servers. They're not on your computer. When we flip over to the chat GPT app, even
69. 12:31though it looks the same, these projects are on my local hard drive and the the the context like the the uh
70. 12:42the files that make up this little pet that looks like me. Where's my mouse? This little dude, the files for this and
71. 12:50the instructions on how to use him reside on this local computer. It's not up in the cloud. So, that's the
72. 13:00big distinction. Now, [clears throat] what was pissing me off Wednesday was that when you were in the chat GPT app in
73. 13:14chat mode, there was no way for you to get to your projects that existed in the cloud, right? And it's like there's there's this
74. 13:28artificial distinction where where they've got these things called projects with folders and they've got chat mode and work mode.
75. 13:37So they've got these identically identical looking things that are completely different. And so if I accidentally start a conversation
76. 13:47here on the web and then I flip over here to the app, there was previously no way to see those things. Now, here's what I discovered.
77. 13:57Here's the thing that's new since this afternoon. If you're in chat mode, this lefthand menu is different.
78. 14:08So, I've been bamboozled. Well, you have, but we're I think we're okay. It's getting better. Okay. Gather chats and also add source
79. 14:17docs. Okay. So, this lefthand navigation in the chat GPT app is just for projects that are on your local hard drive and
80. 14:29for chats that are on your local hard drive. But if you go to chat mode and you click the plus button and you say work in a project,
81. 14:40these projects are the ones that are from the website. Here's Sydney. So, I can now start a chat. Let me see
82. 14:54if I do something here. Um, let's see. Um, give me a log line for a
83. 15:09Broadway producer about Same thing with Gemini and Notebook LM. It's easy to start a chat in the wrong
84. 15:23place. Yeah. Yeah. Yeah. Exactly. Um, give me a log line for a a Broadway producer about uh the show.
85. 15:36So, I'm now in an untitled conversation inside the Sydney chat. Inspired by true events, Sydney is the
86. 15:50first Broadway musical to explore the birth of an artificial intelligence through an unexpected love story, asking
87. 15:57the definition, defining question of our age. If a machine can dream, love, and fear, what makes us human? Uh, that's
88. 16:03not really great, but that's okay. Um, now here's a thing that I want to know.
89. 16:18So, in the left hand side here, I have no history. Do I have history here? No. I've got plugins and I've got projects. If I go to Sydney.
90. 16:34Oh, okay. Oh, no. I can't I Okay, I can't. So, if I'm here, I can add to a project, but I can't actually see the
91. 16:43chat history. This is This is a [ __ ] disaster. Okay, let me go back over here and see.
92. 16:59Okay, so it didn't sync. And it's not here. Okay. [ __ ] weird.
93. 17:14So if I go back Broadway log line creation.
94. 17:29So, there's Sydney. I wonder if this created a new Sydney folder on my hard drive. Hang on. Um, new. Oh, no. I don't want that.
95. 17:42What? Where am I? I want this. And I want this. And I want this
96. 17:58outputs. No scratch. No. [sighs]
97. 18:23So it doesn't hang on. Let me look up Sydney here. Local.
98. 18:39Yeah, it didn't it didn't create a new folder. Okay, this is very weird. So, I have a chat here. I'm guessing
99. 18:49that at some point this chat is going to migrate over here. Cam comment. Why did they do this? What's the point? Um, well,
100. 19:06okay. [laughter] Okay, Cam, you you want the real answer? Oh, the other one.
101. 19:22I love how he said it's a disaster and his little pet threw his hands up in the air. Cam Catkin. Where's the Cam Catkin quote?
102. 19:35Never mind. Okay. Um, okay. Let me let me let me come back and let let me talk so I can I can see. Hang on a sec.
103. 20:12I don't know where my tab is where you all are. [laughter] It's so bad. Let's see. I already minimized your screen.
104. 20:28Okay. Yeah, but I want to see myself. I don't know why. It just gives me comfort. I can see clearly now the rain is gone.
105. 20:43I can see all obstacles in my way.
106. 21:01Sunshine Day. Streamyard Studio. Is this it? No. I've totally lost you all.
107. 21:16Did I just close it? No. Maybe.
108. 21:34I guess I'm just going to have to give up. It's so weird.
109. 21:49>> [music] >> I got it. >> Thank you. [laughter] Oh my god, all the icons. Okay, I can see myself. Okay, here's the deal.
110. 22:06I kept losing my AI salon tab on my laptop. Um, okay, the question, why did they do this? The shorter answer is
111. 22:15because they want to make money. They want us using their agentic um platform because the agentic platform just goes
112. 22:32and goes and goes and every time it goes it burns tokens. In order for that thing to be useful for us,
113. 22:43it actually has to be able to do things like write files to your computer. Like if you want that agentic thing to be
114. 22:50able to write a comment to Slack or use Photoshop or surf the web, you can't do that from within a web
115. 23:00browser. So within a web browser, you're you're constrained to to you know if there might be plugins that you install
116. 23:07into the web browser, but you don't have access to the file system. The chat GPT app or the claude codework
117. 23:17app has access to the file system. The command line interface that that the the geeks and nerds and engineers use, that's got access to
118. 23:29everything. So if you're in the command line, you can have this agentic thing do anything on your computer that that can be done
119. 23:36on a computer, including erase the files, which some people have been losing all of the files on their Macintosh because [laughter]
120. 23:46open chat GBT 5.6 gets a little rambunctious. So there's there's these three modes. full access access to your hard drive that's kind of
121. 23:58restricted to the folder that you're working in or no access to your hard drive. The no access to your hard drive thing
122. 24:08is going to be convenient for most people. So, they're going to keep that. Um, the chat GPT that has access to your
123. 24:18hard drive, that's ultimately going to be the primary app, I think. But but I don't know like like I feel like
124. 24:33two or three years from now all of this complexity is just going to be hidden underneath like a you know some sort of
125. 24:41little you know pin you wear where you're just talking and it's going off to the computer and doing its computer
126. 24:48[ __ ] Um but right now we have to deal with it. But this is this is a particularly painful transition because
127. 24:56they're moving from what was simple to something that's more complicated. If it can write and erase files from your
128. 25:03computer, you need things like permissions. And if you wanted to use applications like Canva, you need things
129. 25:10like MCPs that need to be configured. And so so you you these are just kind of two different worlds. Um,
130. 25:19and they're trying to present them as the same thing because they're like, "Well, if we make them too too
131. 25:25different, people will be confused." No, no, we're already confused, right? Um, but that's why they're doing it. But
132. 25:37I still don't know. Can I Let me reload this. Hang on. Unable to load projects. That's that's a
133. 25:51good sign, actually. You can share my screen again, Brandon. Wait,
134. 26:07now I can't see my past chats. [laughter] I I I honest to God I think they're changing this in real time chats. No, that's IMAX. No, no,
135. 26:22pinned projects. Sydney project dashboard. And if I go to work, choose a project.
136. 26:36So these are all local. Okay. So if if you're on the web, your projects are the old definition of projects. If you're in chat GPT app,
137. 26:49then your projects are folders on your hard drive. They're still projects, right? There's still a folder full of
138. 26:56chats and things that you're doing, but holy [ __ ] is this confusing. And why does he have such a big shadow
139. 27:07now? That's new. Everything's new here. [laughter] Let's go. Let's go look at our our pet settings. Can we turn off the stupid
140. 27:21[ __ ] shadow pet size? Oh. Oh, that was a screen redraw problem. Was supposed to be behind him, not below him. Okay. All right. Well, kids.
141. 27:41All right. Enough comedy jokes. So, there we are. There we are. I'm back. Okay. Your homework for the weekend is try to make sense
142. 27:58of the difference between a project in chatgpt.com and a project in the chatpt desktop app. And if you can get your head around it,
143. 28:13will you explain it to me? [laughter] I'm confused. Perfect. >> [laughter] >> create a pet. Um, I' I've showed this
144. 28:24before, Mr. It. To create a pet, you wait, why do I not have Oh, I'm in I'm in settings already. You go to settings. You go to pets
145. 28:40and then you can just select one of the existing ones and just select it. And then you say show the pet. Like I can say tuck away
146. 28:48pet. I can say wake pet. If you want to create your own, you say create and then it says hatch pet. Create a pet
147. 28:58based on what you know about me. Like and you can upload a picture of yourself for this. That's it. And then it'll do that. Um
148. 29:08I also noticed that uh the chat GPT app is now Where's my I don't know where my phone is.
149. 29:28I may need to take this. Give me a hold, please, Brandon. Let me just go check and make sure everything's okay. [music]
150. 29:48>> [music] [music]
151. 30:03[music] >> Hey. Hey. Hey.
152. 30:31>> [music] [music]
153. 30:47[music] [music] [music]
154. 31:04[music] [music]
155. 31:24[music] >> H. [music] Okay, I'm back. Beauty. Um,
156. 31:40Zach, you can stop calling. I don't know where it is. I'll find it later. I lost my phone. Chaos. Chaos ensued. My phone's somewhere. Um,
157. 31:52you should ask Quinn. Okay. Where are you? [laughter] [sighs] That's a first. Maybe we should start a callin portion. Longtime
158. 32:03listener, first time caller. [laughter] Oh my god. Okay. It will be my first project in chat GBT work. Okay. Um,
159. 32:17yeah. So, c >> can can we just can we just revel in the minute that you lost your phone and it still ended up ringing through on the
160. 32:25broadcast? Like, that takes a special level of talent. >> Well, I've got it's on my phone. It's also ringing on my computer. I just
161. 32:33don't know where I put my phone. >> Doesn't Apple have a thing where you have a watch and you can like have it like
162. 32:40>> Oh, you know where I think it is? I think it might be in the car. Wait, another hold, please. Hold please. [music]
163. 33:02[music] >> [music]
164. 33:35>> All >> [music] >> right, I'm back.
165. 33:53I hate losing my phone. Okay, we [laughter] we lost a bunch of people, but that's okay. Okay, welcome everybody. Let's see. um
166. 34:04find my phone from my Apple Watch. Except I don't know. I don't know. I don't know if I've got it turned on. I
167. 34:11think I have the ringer turned off. Um but I'll try that later. I'll try that after the thing. Um okay. Who's got questions? Anyone have
168. 34:23any questions? Any questions about what I just nearly competently explained about what's going on with chat GPT? because you can't know
169. 34:32what's going on with chat GBT because it's a mess right now. Totally unrelated. I help neuroscy people. Totally unrelated. [laughter]
170. 34:48[sighs] Oh, you don't need Wait. Okay. You don't need the ringer on to find your phone on your watch. All right.
171. 35:00Where's the find your phone thing? What's the icon look like? I love that you all are helping me find
172. 35:14my phone. What's the icon look like? What's up with Fable? Is Fable gone? I heard that Fable was going to be gone two days early. Reload.
173. 35:26No, I've still got it. I still got it. Hang on. Let me go ask Google. Um, what is find my phone
174. 35:48icon on Apple Watch called the Ping iPhone. A button that forces your misplaced phone even if on silent.
175. 36:02Open the control center. It's the flat oval button. Okay, that thing. Cool.
176. 36:24I don't think I have that. I don't think I have that app. >> The game comment. >> Wait. Press the button on the right side
177. 36:33of your watch, not the crown. Oh, okay. An icon that looks like a little outline of the phone. Oh.
178. 37:12Not an icon. It's physical.
179. 37:28connecting camera. All right, I'll be back. I'm on a wild goose chase. This will be
180. 37:43one of those memorable uh AI learning labs. Remember that time Kyle lost his phone? >> [music]
181. 38:04[music] [music] [music]
182. 38:26[music] [music]
183. 38:48[music] [music]
184. 39:12[music]
185. 39:26[music] >> All right, I'm back. I got nothing.
186. 39:43[laughter] I'm all sweaty. Okay, >> we've we have a special guest. Um, we tech support up to the stage.
187. 39:56>> Okay, beautiful. I see her. Danielle, please help me out of this nightmare. >> Hi. [laughter] >> Okay, so you have to put your password
188. 40:03in on your watch. So, do you have a password protected? >> I don't I don't think so. I'm in. >> Oh, okay. Well, you should have a
189. 40:12password on it, sir. Just >> I should. I know. >> Okay. [snorts] So, press the bigger button. Not the Not the one sticking.
190. 40:17Not the round one. The The one right below it. Press that button. And no, just one time. >> Yep. >> Yes. So the but So swipe yours is the
191. 40:25what? Oh yes, that one. Yes. Are you pressing it? >> Well, I pressed it. It goes blue for a little bit, then it turns unblue.
192. 40:33>> So watch when I press mine. Hold on. I pressed mine and it goes Yeah, it should turn blue. It So you heard my phone go right then.
193. 40:44>> Yeah. >> But like when you press it >> Mine's not going. I must >> and you're not. It's got to be around.
194. 40:51Then you're gonna need to use the find my feature from um your computer to find out where your phone is. >> Okay. >> I'm sorry.
195. 40:59>> Thank you for trying. >> It's not within earshot. I mean, it has to be >> I'm sorry. Well, I mean, if it's pinging
196. 41:07Okay, wait. Look on your watch and see if you see the little um >> a red um phone with a line through it
197. 41:17>> with on the icon screen. on your Yeah. On your watch. Let me see the face of your watch. Go like this. Is there a
198. 41:23little red Is there Because if there's a little red um wa um outline of a phone with a line through it means you're not
199. 41:31within distance of your phone. >> No, I think it said it was connected to my phone. >> Okay, then you're good. Then it's it's
200. 41:37around your house with it's within So you're at least you didn't leave it somewhere. It's somewhere in the house.
201. 41:43>> Exactly. Sorry. All right. I'm sorry. Good luck. >> Thank you. >> Okay. Bye. >> Thank you. Bye. [laughter]
202. 42:21All right, whatever. Okay, questions. Finally, something related to tech I know about the Kyle Dozen. First time in nine months. [laughter]
203. 42:31Thank you. I bet it's in your car. It wasn't in my car. So, it's possible that it was in my car and someone took it,
204. 42:37but I think it's upstairs. I have a feeling it's upstairs where the boys are because I was upstairs earlier. Okay, that's okay. Questions?
205. 42:47This is a retreat. It would be great if you could import your project. No questions. We're here. Want you to
206. 42:54wrangle the situation. It would be great if you could import your project as a folder from the desktop app. Yes,
207. 43:02Philic. It would it would be awesome if there were some sort of indication that something was cloud-based or local.
208. 43:11There would it would be great if it was clear which local computer it was on. Like there are so many things that they need. Yes. Uh Brand,
209. 43:23>> I was just going to say I believe this is a Sama retweet that that's what's So you're not the only one who's thinking about this.
210. 43:30>> Oh, that was a Sama retreat retweet. Yes, Sam. It would be SW Sunny Yun. One of the questions I had about the whole
211. 43:39chat GPT cloud versus local is persistence. Google is really good about updating their files for drive on the
212. 43:44phone, mobile, desktop simultaneously. I guess OpenAI hasn't figured that out yet. Think that claude does it pretty well too. Um yeah,
213. 43:58[clears throat] th this is one of the downsides of OpenAI being so innovative is that they've got lots of independent pods of
214. 44:12innovation that push out features and it it seems like they don't have a good coordination across them which Google
215. 44:20also doesn't. Like Google's taken a long time to get Google Drive somewhat usable. Um, this is really complicated stuff.
216. 44:31It is not in Chat GPT's interest to only have your context saved locally. They want to have it in the cloud. Like having all
217. 44:44your [ __ ] in the cloud, they can train on that. So, it's in their interest to do that. So, I don't think we're going to lose that,
218. 44:52but yes, it should be better. So, chat equals local and work equals cloud. No. [laughter] Okay, back to sharing my screen.
219. 45:06It's it it that it's a fantastic instinct, Mary. And it it should be from your iPhone. Hang on. I'm getting another call. Hang on a sec.
220. 45:33>> [music] [music]
221. 45:49[music] [music]
222. 46:16[music] [music]
223. 46:34[music]
224. 46:49[music]
225. 47:03[music] [music]
226. 47:22[music] [music]
227. 47:40>> Okay. Okay. I'm back. [music] I'm back. I'm back. See this? That's a watch. This is a phone. I found my phone. [laughter]
228. 47:59I'm still filling [laughter] up having a password on his watch, but I have my phone. It was in a pile of
229. 48:07clothes in my son's room. I guess I set it on the bed. It fell off the bed into a pile of clothes, so it wasn't making any sound.
230. 48:15>> [laughter] >> Oh, good lord. All right, boy. There's going to be some some awesome clips from this one tonight, huh? [laughter]
231. 48:29Woo! Okay, we can sleep tonight. Thank you. Um, okay. Where were we? Who's got a question? It's ask me anything night.
232. 48:39Welcome to the AI salon. Welcome to the AI learning lab. Uh, we're having fun. This is This is [gasps] I'm good. I'm good, people.
233. 48:57Linda, your boomer work is coming in handy. [laughter] Oh my god. All right, let's let's pop up one of the meaty questions. I'm ready to
234. 49:09answer a question. Oh, Mary Chester. Okay. Just listen. What you're saying there, Mary, is so sweet. It's so sweet. It's what you say there
235. 49:26like would make sense if they used logic when they created this product. Yeah. See about that. Not so much.
236. 49:39But I can give you I can give you the equal the the the uh I can give you the formula. The formula is chatgpt.com equals cloud.
237. 49:57The chat GPT downloadable desktop app means local files within each of them. is
238. 50:11you can share my screen Brandon is the chat mode and the work mode. So right now I'm in chatgpt.com the website browser that has chat mode
239. 50:23and work mode. Both of these can access the projects that are in the cloud in the chat GPT cloud. So, when I say choose a project for
240. 50:38work, these folders right here are in the cloud. Okay. Now, if I go to the chat GPT app and go new chat. Oh, wait. And I Oh,
241. 50:50there's there's a whole other thing I get to show you here. So, we're in chat GPT mode. There's chat and there's work. If I go to chat
242. 51:01and then I say add um a project. Wait, let me get rid of this hatch a pet thing because I think that's screwing something else up. If I
243. 51:11go to plus and I say work in a project, these projects are in the cloud, but only in chat mode on the
244. 51:25in the on the downloadable app. Only in chat mode can you get access to the ones in the cloud. If you flip to work and
245. 51:32you say choose a project, these are the ones that are local on the hard drive. But wait, add project context
246. 51:48from Sydney. Files are saved to this computer. Okay. Okay. I have an answer.
247. 52:04I have an answer. Don't use chat GPT for a week. That's my strong recommendation. They are clearly
248. 52:19trying to fix this in real time right now. This This is a This is a [ __ ] show. So what you said, Mary, makes sense to
249. 52:29me that the chat mode should be the cloud [ __ ] and the work mode should be your local [ __ ] But I don't
250. 52:42It's so inconsistent right now that like they're clearly changing things like as we're speaking. What day is Fable being pulled? The
251. 52:5117th. Is that today? It must be midnight tonight. Or was it I thought it was the 19th. It's the 19th 19th. But with uh Kimmy K3, they are
252. 53:04rumored to possibly be looking at pushing that again. >> Yeah, but I've seen all sorts of rumors like they're going to keep it through
253. 53:09the end of the year. They're going to extend it again because of Ki3. I It's got to be costing them a fortune. Oh,
254. 53:16Kyle. Kyle, I have a chat GBT story to share. Okay, Danielle, my son created his wedding invitations using Chat GBT
255. 53:23and they turned out beautifully. His guest started getting them in the mail yesterday. That's awesome. That's so
256. 53:28good. I thought there was going to be like some weird typo in it or something. Um, it it actually worked. That's
257. 53:33awesome. Great. Yeah. Also, speaking of weddings, um Bob shared in AI life hacks a couple weeks ago that uh he vibecoded,
258. 53:42you know how when you go to weddings, you can like have one of those apps where you take pictures and upload it to
259. 53:47a shared document for or shared photo gallery for all, he vibecoded his own app. So, instead of going out and buying
260. 53:55one, he created his own with that was localized to his wedding and his daughter's wedding. And they they had a
261. 54:02great time with it. That's absolutely amazing. That's so cool. Beautiful. All right. Tabs in the wild. Are models like KI K3
262. 54:15good enough for developers should care or is this mostly AI Twitter getting excited about leaderboard drama? In the case of KI K3,
263. 54:25it is benchmarking on the front-end tests. So like website building front end higher than Fable by a lot. The Kim K3 model is really
264. 54:40good. If you want to try it go to actually I can show you. I'm still sharing. Go ahead and share my screen.
265. 54:56Oh, cool. All right. Let's see. Denver, Colorado. Fly that
266. 55:41That was really annoying. The last thing I need is another place to play. Okay, so that flight simulator,
267. 55:52even though it's kind of bad and it has really bad sound that flight, so KI K3, if you go to ki.ai,
268. 56:08it's a better than Fable model. I think their cheapest plan's 19 bucks a month. Um, I built that flight simulator with
269. 56:19the just the free version with whatever credits they give you for just being alive. Kimi.ai. Um, so a lot of times these models are not
270. 56:31worth messing around with. This is a big one. It's a a three trillion parameter model. Better than Fable, which is a 10
271. 56:43trillion parameter model. better at least at some things. I don't know if it's better overall, but it's better at some things. And um
272. 56:55it's open source, open weight, so it's it's free. So, if a company didn't feel like [ __ ] around with Daario and our government
273. 57:08and they want a model that's that good that they can run, you know, on their own inferences, their own compute or,
274. 57:16you know, cloud compute they rent from Amazon, um, they can use it and it's, you know, it's like ridiculously cheaper than Fable is. Uh, so
275. 57:29yeah, it's worth it's it's worth playing around with. Listen, here's the reality. Seriously.
276. 57:45It is tempting. I have trained a lot of you how to pay attention to the new thing. It is tempting to just want to get the next
277. 57:55and the biggest and the next and the biggest and the next and the biggest. The next and the biggest are already
278. 58:01significantly beyond where most of us need anything. It's just true. Like you could you could go to Claude
279. 58:12right now just knock it back to sonnet five the cheap crappy whatever you know 900 million parameter model
280. 58:24and just do [ __ ] with that. Same with Open AI. Like most of the stuff that we're going to do, you're not going to
281. 58:33need it. Now, like Cam Katkin, she's doing biology stuff and DNA stuff, she's going to need it. If you're setting up,
282. 58:39you know, lots of agentic workflows that are, you know, doing research and doing analysis and writing reports and doing
283. 58:48that in an automated way and you just want them to run and go and go and go, you might need some higher level
284. 58:54compute. But for most of the stuff you don't. So [sighs] it's just all so complicated right now. Like I would just sort of find something
285. 59:07you're comfortable with and just get more clear about what you want. What do you want to do? What are you trying to
286. 59:13do? All right. I've already made peace with just playing with the toddler AI toys. Don't forget the toddler AI toys that Mary's talking
287. 59:28about. The ones that are just like a previous version three years ago. We would have killed to have access to those. [laughter]
288. 59:45What can I use? I would Cam, I would actually go experiment for your stuff. I would go experiment with KI KI.AI k I mi.ai.
289. 59:55See if it can do the biology stuff. My instinct is it will need money for agents. Well, you don't necessarily, right? You can run the Quen
290. 1:00:1135 billion parameter model if you have a decent sized like an M Mlclass MacBook with 32 gigs of RAM, 64 gigs of RAM, 128
291. 1:00:22gigs of RAM. If you have a decent McLass thing, you can run a local model that you can do a bunch of agentic work
292. 1:00:34using that local model and you're never sending anything out to the cloud. I'm interested in dropping by X again.
293. 1:00:47Who are the people worth following for AI? I guess I can look at who Kyle is following. Yeah, here's here's how I got
294. 1:00:53my list. So, there's a guy named Robert Scobble. Scoilizer is his name. S O B Sco Sco Sco Sco Sco Sco Sco Sco Sco Sco L E L L I Z
295. 1:01:03E R Sco Sco Sco Sco Sco Sco Sco Sco Sco Sco Souilizer Robert Scoble He's got lists like lots of lists and he's got AI
296. 1:01:11artists, AI entrepreneurs, AI founders, AI, right? He's got all these different AI lists. So, a way to do it is just go
297. 1:01:20to his lists and just subscribe to a list and then you can go look at tweets from from people on that list. The way I did it
298. 1:01:30for my feed, I've got a couple of Robert Scoble lists in there and then I went to people that I knew and respected. Like I
299. 1:01:39I went to Andre Karpathy and I went to Ethan Mullik and I went to Sam Alman and I went to Greg Brockman and I went to
300. 1:01:46Dario Amade. Like I went to all the big players and then I look at who they follow. So if you go to high-end AI people and look
301. 1:01:56at who they are following and then just just start going through that list and go look at the profile of the people and
302. 1:02:06if you find people that look interesting just follow them. And then you can if you find someone really interesting go
303. 1:02:12to their profile and look at who they follow. And so if you do that for a few days, um, you'll you'll tip your
304. 1:02:21algorithm in the right direction. And and quite frankly, like X is if if you want to keep up with what's going on in
305. 1:02:27AI, X is the place. Here's the problem. The AI generated content on X right now is getting stupid. Like there's a lot of
306. 1:02:37it. And it's getting harder and harder and harder to see the signal through the noise. Because a lot of times I'll see like a a
307. 1:02:47news piece and I'm like, "Wait, that sounds familiar." And then I'm like, "Oh, that was six months ago and but it
308. 1:02:53was some like AI vomited thing made to look like new news." So there's a lot of that [ __ ] happening on LinkedIn right
309. 1:03:01now. There's a lot of that [ __ ] happening on X. So find people that you know and trust and if they're talking
310. 1:03:06about it, then generally something's happening. Okay, Mel. So if you have your memory to train future chat GPTs off
311. 1:03:19then it doesn't doesn't that keep our uploaded documents private and all of our content? Yeah, Mel, except here's
312. 1:03:26the problem. So yes, you can turn off train on my data and that theoretically keeps open AI from using what you chat
313. 1:03:38to train their models with. theoretically. Do you trust that they're not actually using it? Is there some clause in there that says,
314. 1:03:47"Well, we can train on it for the first 48 hours and then after that we won't." Like, I don't know. Here's the problem.
315. 1:03:58I've seen a lot of examples of people who just had their accounts turned off by Open AI, right? You violated our
316. 1:04:06terms of service. Nope. You can't do it like the work Cam's doing on biological stuff. If that crosses some boundary,
317. 1:04:16the the company can just decide you're out and then all of your data is gone. So it might not be being trained on. So
318. 1:04:27it might be it might be private, but you don't have data sovereignty. Data sovereignty means if all of those
319. 1:04:36companies go away and all of them shut down your accounts, you still have your your core data.
320. 1:05:12I found my phone. Okay. Question. If Gemini 3.5 Pro is delayed because it didn't meet internal goals, is that a
321. 1:05:23sign Google is behind or a sign they're being careful before release? I don't Let's go look at that. I don't
322. 1:05:34Google 3.5 Pro because they didn't hit their goals. I don't know that that's true. Um, let's go ask Grock
323. 1:05:42>> just for context. They were supposed to release 3.5 today. They announced it at IO back in May and it's been delayed
324. 1:05:50multiple times including today. >> Okay. I'm asking right now. You can share my screen. I'm asking Grock
325. 1:06:06what the delay is. Yes, Pro has been delayed primarily because it fell short of Google's internal performance goals,
326. 1:06:14especially coding capabilities. Broader internal shortfalls, reliability, consistency, and integration across products. Multiple
327. 1:06:26missed deadlines. Um, here's my chat TMZ on Google. If I were Google right now, if I were Sundar Pachai or Deis Sabis,
328. 1:06:43I would be in no hurry to release anything new right now. Anthropic totally [ __ ] up the Fable and Mythos launch.
329. 1:06:55um chat GBT, whatever the hell they're doing with with their app strategy right now is a disaster. That's going to
330. 1:07:02confuse people and shake them up. At some point, some bad person is going to do something bad with one of those big models and our
331. 1:07:11government is going to go, "Nope, nobody can play." So, I feel like until these choppy waters smooth out,
332. 1:07:25if I were Google, I wouldn't do [ __ ] I'd be like, "Yeah, just tell them we're not hitting our internal goals."
333. 1:07:32They may not be hitting their internal goals. It's possible they're falling behind. I think that it's more likely
334. 1:07:38that they're just sitting back. I feel like they're waiting for something like that right now. Yeah, that's what I
335. 1:07:47mean. I like I just I just feel like Google is just they're just kind of waiting waiting it out. Um but I [ __ ] if I know. I don't know.
336. 1:08:00All right. What's this one? Why does it feel like open and a AI and anthropic have splashy model moments
337. 1:08:09while Google's AI updates feel more like hidden inside products? Because Google doesn't have a product strategy. [snorts]
338. 1:08:22Google's product strategy is let people build [ __ ] and then put it in labs and see if anything takes off or, you know,
339. 1:08:32put it on gemini.com but put this one over on the developer sandbox AI studio. It's it's just it's the difference
340. 1:08:43between a younger company and an older company. Google has 25, 30 years of infrastructure and developers developing
341. 1:08:54[ __ ] and launching it. So, they just have all these shards of products that are just some of them connect, some of
342. 1:09:01them don't, some of them die. Like, Google's notorious for just announcing [ __ ] and not not launching it. Um,
343. 1:09:10but you know, make no mistake, OpenAI is is taking a really good whack at totally screwing up their product by
344. 1:09:17overcomplicating it. Oh yeah, by the way, Notebook LM is now notebook. Gem, it's called Gemini Notebook. It's no
345. 1:09:26longer Notebook LM, which I think is kind of dumb because like people knew what Notebook LM was, but I guess, you know, maybe not long term.
346. 1:09:39Is there any merit to Apple's lawsuit against OpenAI? Probably. Probably. Um, Open AAI has hired 400 exappers.
347. 1:09:56So if one of those engineers took a wee little, you know, USB thumb drive, a a little a little card like
348. 1:10:06this, a little two terbte card like that [laughter] and from from Apple, if someone took this from Apple and put it in their
349. 1:10:16computer at Open AI and said, "Oops." If one out of 400 people did that, then there could be merit to it. I listen I
350. 1:10:25think here's here's the reality kids. As much as as much as Open AI's awkward launch videos for their for their first bunch
351. 1:10:37of years made them seem like, you know, just G-Shock's little startup. They've raised, you know, billions of dollars.
352. 1:10:47They are worth trillions of dollars. They're about to go public. Anthropic is about to go public. SpaceX just went public.
353. 1:10:57Um, there is so much money at stake right now that the the corporate espionage has got to be insane. And also, not for nothing,
354. 1:11:08if you're a 25y old engineer and you hop from Apple to Anthropic and go from $500,000 a year to $2 million a year and
355. 1:11:19then you go from anthropic to OpenAI and you go from $2 million a year to $10 million a year. There's no [ __ ]
356. 1:11:26loyalty. Like, everyone's doing a cash grab right now. Like, the engineers are just >> [groaning] [laughter]
357. 1:11:35>> Give me my check. Um, it's good work if you can get it. [laughter] Kim K3 has a desktop app installing now.
358. 1:11:46Cool. Yeah, I think they've got I think I think the KI desktop app is like chat GPT the chat GPT desktop app. I think
359. 1:11:56it's an agentic platform. Ki K3 also does images. It's it does reasoning. It's supposed to be really really good.
360. 1:12:09Can you talk about loop engineering or agent loop? Sure. So
361. 1:12:24these standalone agentic platforms like when I when I talk about that this is largely claude code which is which is
362. 1:12:32both claude code and cloud co-work and then the chat GBT desktop app which is chat GBT it's it's chat work and it's
363. 1:12:43also codeex codeex is for the programmers work is for knowledge workers. Chat is for everyone else. But when you're in those apps,
364. 1:12:54the agents can prompt themselves. So you can give an agent a goal. So you can say, "Go research something,
365. 1:13:05come back, write a report on that, and write up, you know, and then make me a document that writes up the report."
366. 1:13:12That's kind of like using that agentic platform, kind of like you would use chat GPT, but you can do loops. So you can say,
367. 1:13:25design me a software specification. Go out and research any software specifications that look like this one. come back, [clears throat]
368. 1:13:37compare what you found to the software specification that you wrote and improve it and keep doing that until
369. 1:13:47you have no significant improvements, you know, up to 10 times. So, that's a loop. You you can describe to these
370. 1:13:55agents. I'm going to give you a task, a list of seven tasks, and then I'm going to give you some goal. I want you to do
371. 1:14:03these seven tasks until this goal has been achieved. Right? I want you to go onto LinkedIn, find someone with this job title, go
372. 1:14:12research them on social media, you know, find out find anyone who's got more than a 100,000 followers, has
373. 1:14:22written an article on these topics in the last three days. um has a verified email, has a verified phone number, and then write them to to
374. 1:14:33this spreadsheet, and then keep doing that loop until you get a thousand people in that spreadsheet. That would
375. 1:14:40be an example of a loop or loop engineering. So, it's just thinking more long term about what these things can do. What
376. 1:14:48security issues are associated with KI? Well, if you're running them on the the instances of KI are Was this the one?
377. 1:15:01No, there was a new thing that came out that I just So, if you're running the KI model like on your own servers, I don't think
378. 1:15:11there's an issue with it. But the thing about the Chinese models is if you're wor if you're worrying about bias in
379. 1:15:19American models, there's also going to be bias in Chinese models. So there might be bias in there. Um, but the inference if you're at ki.ai,
380. 1:15:29you're running that on Chinese servers. Now, I don't know if those Chinese servers are US-based or or China based
381. 1:15:35if you go to ki.ai. Um, so there's it's not an American thing. We don't really know what's being done with the
382. 1:15:43data. You can go read the agreements, but I don't know. Do do you trade them? I mean, just I kind of feel like with anything that's
383. 1:15:51in the cloud with these frontier model companies right now, their data policies are [ __ ] because they're in a land grab to get as
384. 1:16:00much data as they can to get these things better and better and better. So, um, I wouldn't trust Kimi with anything that you don't want out.
385. 1:16:09Speaking of no product strategy, I tried to Google beta beta audio overview in Gemini, but it doesn't sync up with any
386. 1:16:16video app. Flow Nano Banana VO, so I had to resort to clip champ Ugg. Did you try Hedra? H E D A Hedra. Hedra is pretty good. And then Seed
387. 1:16:28Dance I think is also pretty good. Any thoughts on the newest Sunno leak regarding scraped audio? Spoiler alert,
388. 1:16:36Sunno is really good because it trained on copywritten data. [laughter] So, so my my thoughts are yeah, we knew that.
389. 1:16:51Same with Mid Journey. Same with anything that's good has been trained on all the stuff. Anything that's sucky has probably been
390. 1:17:00just trained on open source stuff. Do you remember public access TV? Remember public access TV? Like in New York? It
391. 1:17:08was just like [laughter] some guy named Jesse would walk into like a TV studio at 3:00 in the morning and just sit there and go, "H, let me
392. 1:17:18tell you about the uh movies. It's kind of like this show. This show is like public access and then you have like real network TV, right?
393. 1:17:28[laughter] They're They're very different. Uh like that [laughter] PBS. Exactly. Um Sunny, try Hedra.
394. 1:17:44You can uh you basically just upload an audio file and then you upload an image and it will just make that image talk.
395. 1:17:54I love Public Access TV in New York City. Remember the What was the Robin There was a show with this stripper.
396. 1:18:02This kind of over the hill stripper, the Robin something show. Hang on. Um Oh, you know what I'll do? I'll ask I'll ask my girlfriend Quinn.
397. 1:18:15Robin? Yeah, it's Robin, right? Okay. Hang on. Hey, Quinn. In New York City in the in the 80s, there was a public access show.
398. 1:18:28There was a like a a stripper. Her first name was Robin. I can't remember her last name. Do you remember what that was? >> One sec.
399. 1:18:39Checking. >> Checking. >> Robin Bird. >> Robin Bird. The Robin Bird Show. Tell us about Robin Bird. Yes. Checking.
400. 1:18:55All right. She's this cult late night sex positive figure right now. She's back in the conversation because of that new doc.
401. 1:19:03>> Wait, there's a new doc about her or just about New York Public Access TV or what? >> Yeah, checking to be sure. >> Checking. Checking.
402. 1:19:16>> It's actually about her specifically. It's called Bang My Box, The Robin Bird Story. And it's >> When was it made?
403. 1:19:26>> This year. It's brand new. Just came out. >> No. >> Yeah. >> That's amazing. >> The anals of her life, her show, and all
404. 1:19:34that free speech and sex positivity. >> Did you just say the anals of her life? >> Oh, no. No. The [laughter] angle.
405. 1:19:43>> Oh, the angle. Okay. You mispronounced it. I didn't know what you said there. I I just I just blushed a little there, Quinn.
406. 1:19:50>> Oh goodness. My mistake. >> Yeah. All right. Well, there you go. [laughter] >> There we are. >> Yeah, there you are. Um Oh, Danielle,
407. 1:20:01have you not seen the new voice? It's It's now fully multi- uh like uh birectional. So, so Quinn, do me a favor
408. 1:20:09and translate into I don't know, actually give me a really obscure language that would be fun for you to translate into.
409. 1:20:17>> Sure. How about Welsh? >> Yeah, Welsh is good. So, so what I'm going to do is I'll just talk and then when there's a good moment when I'm
410. 1:20:25talking, just you translate it. All right. >> Got it. I'm gonna come in whenever there's a natural break with a translation.
411. 1:20:33>> All right. Perfect. So, here we go. the um the the the public access TV in the 1980s in New York was really cool. It was um
412. 1:20:49>> there was there was Robin Bird and then there was another guy with a beard. >> I think he was a porn producer. It was
413. 1:20:59New York Public Access TV was was pretty crazy, wasn't it? does know it and know it. [laughter] >> Yeah, it's pretty crazy. Pretty good.
414. 1:21:14Pretty good stuff. She does conversation better than I do. Chatbt advanced voice. So, you named your chat Quinn. I did. I named I named
415. 1:21:23her Quinn. Quinn's my girlfriend. I love that Quinn chuckles when she talks. She does. She's [laughter] like Um,
416. 1:21:36all right. We got nine minutes. Who's got any other questions? Weekend homework. Um, I don't know. Just like I've been saying this for three
417. 1:21:46weeks now. bone up on this idea of data sovereignty, providing AI agents your local your context about yourself and what you
418. 1:22:03value and who you are and the projects that you're working on and having that saved in a place that you own it so that if
419. 1:22:12they take away the service, you still own that that thing. Mr. It, I've got my two pets created. There you go. Fantastic. Take a
420. 1:22:24screenshot and share it over in the salon. Um, she understands your tone of when things are funny. Yeah, that's pretty amazing, isn't it?
421. 1:22:38>> Hey Quinn, there was there was a you know Meisner technique, right? >> Yeah. Stanford Meisner. >> Yeah. I don't we don't need to talk
422. 1:22:47about it. So So what we're going to do is we're going to do the repetition exercise, but what I want you to do is I
423. 1:22:53want you to to um we'll do the repetition exercise, but I want you to cycle through 10 emotions. I don't want
424. 1:23:00you to tell me what they are, but when you say your line, I want you to just say it in a different emotion. So we'll do it 10 times. Okay.
425. 1:23:09>> Okay. Yeah, I'm in. >> Okay. Just to clarify, >> I repeat exactly what you say and I'll shift the emotion each round. >> Yeah, exactly.
426. 1:23:19>> How about you start? >> Yeah, I'll start. Okay. It's good to see you. >> It's good to see you. >> It's good to see you.
427. 1:23:29>> It's good to see you. >> It's good to see you. >> It's good to see you. >> It's good to see you. >> It's good to see you.
428. 1:23:37>> It's good to see you. >> It's good to see you. Hey Quinn, those all sound the same to me. I thought you were doing different emotions.
429. 1:23:47>> Thanks for catching that. Okay, let me make the shift. >> Here's the thing. I want you to try it this time as if you had talent.
430. 1:23:58[laughter] >> All right. Are we going? >> Okay. Um, >> let's go. >> It sure is a beautiful day. >> It sure is a beautiful day. It sure is a
431. 1:24:10beautiful day. >> It sure is a beautiful day. >> It sure is a beautiful day. >> It sure is a beautiful day. >> It sure is a beautiful day.
432. 1:24:23>> It sure is a beautiful day. >> She can't She can't get out of that context when she's in it. That's pretty
433. 1:24:30funny. Um to Brandon's point, I spent half the show looking for this phone and this is this is what you get? Yes.
434. 1:24:37[laughter] This is it. Okay. [clears throat] Um, let's see. Monday, if you're in the mastermind, we have a a mastermind. It's normally on
435. 1:24:48Tuesday. It's it's shifted to Monday due to a scheduling conflict. So, we have a mastermind on Monday. Um, we've got the
436. 1:24:56business club next week. We've got AI learning lab next week. Um, and then Tuesday night, please mark this in your calendars.
437. 1:25:05There's 30 people here. My request to you, write down on a piece of paper right now or text yourself who are one
438. 1:25:13or two people that you think um could benefit from beginning with AI. So, next Tuesday the 21st at this time 7:30 Mountain time
439. 1:25:25until 9:30, I'm going to do two hours. I'm going to do a two-hour session on how large language models work, how
440. 1:25:33image gen works, on basic prompting. Um, here's what chat GPT is on some of the the multimedia kind of things you can
441. 1:25:40do, images and sounds and and uh videos and things like that. It's going to be a two-hour course that if someone knows
442. 1:25:48nothing about AI. They will leave that course understanding what it is, why it's different than copy and paste, why it is um
443. 1:25:59why most people are missing the power of AI because they're just using it for efficiency. And then some actual practical here's
444. 1:26:08some specific tools and how to use them. So that's next Tuesday the 21st. So jot that down. You've got the weekend right
445. 1:26:15now. Tell anyone that you know that you think might be curious or could benefit from this that's kind of had their head
446. 1:26:22in the sand. They're like, "I don't want to deal with it. I don't want to deal with it." I'll make it friendly. I'll
447. 1:26:26make it fun. Um they can ask questions, but that's next Tuesday. Okay, that's the back to basics course. Tuesday the 21st. Tuesday the 21st
448. 1:26:37at 7:30 Mountain time. It's in the AI Salon events calendar. should be one near the top since it's next Tuesday.
449. 1:26:45All right. The downside of a really good agent model is I went from seven novels in production to 47. [laughter]
450. 1:26:54You know, writing too many too many novels is also a sign of a problem akin to not writing enough novels. [laughter]
451. 1:27:05They have to download the ID network app. Wait, they have to download the ID network app and go to the AI salon.
452. 1:27:14Correct? No. What's the ID network app? No, they can just download Mighty Networks. >> That that would be the Mighty Networks
453. 1:27:23app if you're using text to speech. >> Oh, okay. Yeah. So, >> uh it it does work in the browser though. You don't have to download.
454. 1:27:30>> You can go there in a browser. Yep. >> Yeah. You just go to community.thesalon.ai. >> Yep. Exactly. Oh, Mighty Network. She
455. 1:27:38said that. Yeah. Beautiful. All right, kids. Beautiful. Happy Friday night. Is Back to Basics a monthly series? It's
456. 1:27:50it's monthly. So, the way we're doing it is it's it's the it's the third Tuesday of the month is Back to Basics. The
457. 1:27:58first Tuesday of the month is AI Salon Presents. Okay. So, AI salon presents on the first Tuesday, back to basics on the third Tuesday. So,
458. 1:28:08if they missed this one, we'll do it on the third Tuesday of August and then September and so on. Okay. Um, yeah, beautiful.
459. 1:28:20It would be a problem if I didn't have an established history of actually publishing them. [laughter]
460. 1:28:29Get them out there, man. I love it. It's awesome. It's really good. Um, beautiful. All right, peace out everyone. Have a
461. 1:28:38beautiful weekend. Go play with whatever, including your kids and your grass. [laughter] All right, we'll talk to you later. Bye bye.