4/20/2026 - Testing Local AI Performance on the M5 MacBook Pro with Qwen 3.6 and OpenClaw

Transcript

1. 0:00Oh, champ. Oh, champ. [music] [music]
2. 0:17Come here, Champy. You want me to put you little your little bed over here? Come here. There you go. Lay down there. Now, let's
3. 0:28do a little singing. What do you think? You ready to sing? [music]
4. 1:04>> [music]
5. 1:19[screaming] [singing] >> Happy Monday everybody. Welcome. And by everybody I mean Steo and Nicholas and
6. 1:34Dr. J. I think that's it. Oh, I show eyeshow's here. All right. >> [music]
7. 1:48[music] >> Woohoo!
8. 2:05[music] Wow! >> [music] >> Woohoo! [music]
9. 2:26[music] You're stinky, champ. You're a little stinky. He's a little stinky. [music] You know when their glands get a little funky?
10. 2:41You smell like bad salmon. [music] [singing and music]
11. 3:00>> [singing and music] >> Come on. Come on. [music] [music]
12. 3:17[music] [music] >> [music]
13. 3:41[music] >> Source camp is on time. She set the alarm. Here's the thing about intentionality, people.
14. 3:54You set the alarm. You wake up from your nap. Why did you take a nap? Because you were out educating the people about AI.
15. 4:02It's exhausting. They don't want to hear it. They don't want to hear it. I don't I don't want to. I don't want to. I don't
16. 4:08want to. So, you get home from your long day of I don't want us. You take a little nap. You set the alarm. You show
17. 4:14up to AI learning lab. That's how it works. That's how professionals do it in this world. Kyle bugging my house. [laughter] >> [music]
18. 4:34[music] [music]
19. 4:48[music] >> Boo. Cool. >> [music] [music]
20. 5:07[music] [music and singing] [music]
21. 5:24>> I have to listen with headphones otherwise your dog will start howling. It happens. Happens to the best of them doggies. [music] [music]
22. 5:42[music] [music] Um, so so tonight we'll just have some fun tonight. If if anyone has any questions they want to ask about
23. 5:58stuff or nonsense, feel free to feel free to pop a question or a thought or a comment into into yield old yield old chat. Um,
24. 6:09this week, as you may remember from last weekend, I know I know you're all uh, you know, trying to run your schedules
25. 6:17around mine, so you all know this already, but um, tomorrow morning at like 7:30, I leave for Washington DC
26. 6:30and then, uh, do a kickoff of a AI salon flyin in DC. Um, so that'll be nice. [music] Welcome welcome some companies in.
27. 6:48[music] [music] Um, and then the other thing that's going to happen this week, so, so I go to I go to DC tomorrow, and then
28. 7:04Wednesday morning I fly to Boston, do a kickoff of an innovation summit, uh, Wednesday night, and then it's all day
29. 7:12Thursday and some activities on Friday. So, my week this week, I'm pretty much out. And this is also the week where it
30. 7:20looks like we're going to get a new model from OpenAI. We just got 4.7 from Anthropic. [music] But wouldn't surprise me if we got
31. 7:32something from Google as well. [music] Although I haven't heard any buzz about Google dropping anything. [music]
32. 7:43What do I think about Perplexity Computer? Um, I have only played with it a little bit and I haven't quite distinguished
33. 7:58>> [music] >> if it's any good or bad or or or different. Like right now for the Agentic stuff, I'm putting my attention into my Open Claw
34. 8:11installation, which was on a PC. Now it's on a Mac. That that transition took some time. And he sort of works now. And he's
35. 8:21running on an all local model, which is super cool. I'm running Quen 3.6. 6. So, that's pretty cool. Um, I can probably
36. 8:29want run Perplexity Computer. And then I'm also seeing now there's people that are running uh what the hell is that thing called? Um,
37. 8:39there's OpenClaw. Hermes. And there's Hermes. And Hermes is like OpenClaw, but different. And now there's people
38. 8:45running Hermes and Open Claw on the same machine where they check each other's work. [snorts] So I it's the agentic stuff is such a
39. 8:57different discipline that I don't I don't quite have my head around it yet. [music] [music]
40. 9:15The thing about agentic work, [music] about agentic systems like openclaw is that you can just do stuff, right? You can
41. 9:29just have it go do [ __ ] Make me a report. Make me a this. Make me a that. It'll go off and do that [ __ ] The problem with it is,
42. 9:40do you actually need that [ __ ] Because if it's just off making [ __ ] that you got to now pay attention to, it's not
43. 9:47actually solving a problem, right? Well, so then you go back to, okay, well, what am I trying to accomplish? And I've
44. 9:57got different companies and different projects going. Do I know exactly what I want to accomplish? Maybe. But I don't know how much time I
45. 10:07want to put into training a shitty AI employee to halfass do the work that I'm just going to have to go delete and deal
46. 10:13with at some other point. So I am having I am having birthing problems with my agents. Not because [music]
47. 10:23they're so hard to understand. They're actually decently easy to understand. But the question of what do you want
48. 10:28them to do is actually kind of a meaty question. When people tell me they want an agent, I ask them what they'll use it
49. 10:36for. I usually get, "Huh?" Yeah, exactly. I I gotta have an agent. Are you talking about a custom GPT? Are you
50. 10:44talking about an an agent like Manis or GenSpark? Are you talking about have a suite of autonomous agents running on a
51. 10:53chip locally on your device? Or or are you are you willing to run up a $2,000 a month bill to run this on the most
52. 11:02powerful model out there? What? I want I want an agent for my business. Okay, define agent. You're the expert.
53. 11:16How do I feel about putting OpenClaw on a personal computer? So, I just I just bought I'm I I basically came into a
54. 11:25little bit of money. Um my stepmother passed away. Didn't have much money, but she gave us a little bit. Some for the
55. 11:31kids, some for the wife, some for me. So, I basically took as much as I could of that. And um bought myself a MacBook
56. 11:39Pro, an M5 MacBook Pro. So, this is this is fascinating. So Brent Peterson, who if you don't know, he's he's a guy that hangs out on
57. 11:52office hours. He's got a a company called Signal and Cipher, and they've got this whole he's been developing like
58. 11:59an agent governance platform for two and a half years. So like his company is basically poised for this moment. And at
59. 12:07the time I ordered the MacBook Pro, Brent Peterson said, "Don't bother getting a powerful machine because all
60. 12:15the models suck. they like they're not none of the open source models are good enough to actually run something like openclaw all locally.
61. 12:24By the time my computer got here, which was like a week and a half later, um Quen 3.6 35 billion parameter model came
62. 12:33out and it's really good. And Brent Peterson said, "Oh, this this is actually good enough to run run these
63. 12:39agents on." So, we have officially hit the tipping point where at least one person I deeply respect thinks that
64. 12:47open-source local models are good enough to run something like OpenClaw without having to go to the to the to, you know,
65. 12:55OpenAI or Anthropic. Now, if you want to do kick-ass high-end coding, great. Go do that. But like I I'll give you an
66. 13:03example. Um, one of the things that Adam did is I had him create a daily The Great Repurpose signal, which was
67. 13:15basically go find five news articles that are related to The Great Repurpose. Put it in a nicely formatted email where
68. 13:23you talk about here's the five articles. Here's why I picked them. Here are the the five stages of the Great Repurpose
69. 13:29that you know are are relevant here. um and include a picture from the from the article in the email. Um the quality of
70. 13:38that email when I was using um chat GPT 5.4 and when I was using Quinn 3.6 locally, it was the same. Like I
71. 13:50didn't see a degradation in quality. Um I'm sure there's areas where you see lots of degradation in quality, but it's
72. 13:57pretty flipping good. So he So I'm I'm writing an article right now. Uh Audra. So is it Wait, it's Audra. Audra. Audrea. Audrea. Audrea.
73. 14:12[laughter] I don't know how to say your name. Um I'm actually writing an article about it right now. Here's how I equate it.
74. 14:22When I was a kid in the 70s, you'd go to the arcade and you'd play Asteroids and you'd play Space Invaders and you you'd
75. 14:29play the video games, right? Play Pac-Man. And then there were all these rumors that they were coming out with home
76. 14:39video game consoles and the Atari 2600 came out and I got myself an Atari 2600 and my favorite game was Asteroids and
77. 14:50Asteroids was amazing. The physics were amazing and it it was spooky and it was dark and then Asteroids came out for the
78. 14:59Atari 2600. I'm like, "Oh my god." And and if you remember back then, the box art for Atari 2600 games were was done
79. 15:08by illustrators who would illustrate their interpretation of what the game looked like. So you'd see the game in
80. 15:15the store and you'd be like, "Well, it looks so good." And you get it home and you plug the cartridge in and it was
81. 15:20this stinking piece of [ __ ] these blocky pixels. And the the ship barely turned around. It was I think it was
82. 15:27like eight points. It wasn't smooth. It was like D and then the the sounds were all crappy and the graphics were all
83. 15:34crappy and the gameplay was all crappy. It was just horrible. So the gap between your home video game and what you got at
84. 15:42the arcade was massive. And then over the years it it kind of caught up. And then the era where the Nintendo 64 came out and Sega Genesis,
85. 15:54there was an there was a point at which the video games that you could play on your TV were were like at or above what
86. 16:02you could play in the arcade and and you know so so as the as the quality of games goes up, the arcades started to
87. 16:10die off, right? And went away and then you know video games just got better and better and better. I feel like we're at
88. 16:16a really similar point with open-source models and hardware right now. So, running any sort of open-source model
89. 16:25locally over the past 3 years was just kind of it was kind of like playing the Atari 2600 compared to the the video
90. 16:33game. Um, I feel like this Quen 3.6 six model is the first one that's kind of like, huh, that's that's like a similar
91. 16:43experience as to what I was having with these big big models. Now, I'm just using it informally, so I'm not in any
92. 16:49way comparing Quen 3.6 to one of the big models. It's not, but like it's good enough to run uh run OpenClaw, run my my agent, Adam.
93. 16:59So, I can talk to Adam and he talks right now and I got my my machine set up and so I'm I'm a fan of it. Why can't we
94. 17:07load local LLM like the old Microsoft CDs? Exactly. [music] [music] Jeez seems to work better for me.
95. 17:22Fantastic. Have you tried 3.6 Joker? 3.5 was good, but 3.6 six. I'm using the 3.6 the six the Q6 um K version the quantiz 6
96. 17:47[music] >> [music] [music]
97. 18:07[music] [music] >> Yes, but I'm just now experimenting. Haven't had the GPU long. Yeah. So, the
98. 18:23the Mac I got I got an M5 Max with 64 gigs of RAM and I got the 14inch MacBook Pro. It is it's it's a really sweet form
99. 18:32factor. I like it. I was I was kind of bummed like there was a part of me that was like wait for the new round of of uh
100. 18:39Mac Minis, the M5 Mac Minis, which I'm sure are going to be baller like for, you know, $600, you'll probably be be
101. 18:46able to get something that just destroys everything that's out now. Um, but the M5, the the architecture of the M5, I
102. 18:54thought my philosophy on Max has always been just buy the most expensive thing you can afford. You just you won't
103. 19:02regret it. It'll last you a while. And I feel like I've got a machine now that that I I'll be able to like as these
104. 19:08models get better and better, like it's it's pretty solid right now. [snorts] And it's a really cool form factor. I
105. 19:14don't need to plug in a monitor to it. I can just carry it around if I want to or I can just leave it. I've got it set up
106. 19:18so I can close it and it still is running as a server and then I can um I can remotely log into it from this
107. 19:26machine that I'm I'm broadcasting from. It's pretty sweet. Buy once, cry once. Yeah, exactly. Hey everyone, sorry I'm late. What's up
108. 19:34tonight? Um it's it's going to be tonight's going to have to make up for the week. So because I'm I'm traveling.
109. 19:41I got a I got a really busy week this week and next week. So it's going to be it's going to be thin here in the AI learning lab. Um,
110. 19:49so anything you want to talk about, we can talk about. Um, anything. Maybe I'll demo something. Maybe we'll just have
111. 19:54some fun tonight. I don't know. It looks like we're going to get some new models this week. Um, so that that
112. 20:05is probably going to be something interesting to deal with. Is that your Mac Mini? No. But I can show you. You want to see something cool?
113. 20:15And by the way, the the URL on screen, community.thesalon.ai, if you've not been to that, um, please go join the AI salon. It's where we all
114. 20:25hang out in between these meetings. They always do cool stuff when you leave. [laughter] Um, okay.
115. 20:40So, you can see my this is where I just ordered business cards for social media marketing week, which is next week. Um,
116. 20:49but this I wonder if this is going to work. Why? Hang on a sec.
117. 21:13Black Bar takeover. I can't keep up with all the cloud features. Can we review? [music] >> Yeah, I think black bar has uh consumed your
118. 21:30entire display. Oh, [laughter] >> I know. I'm Hang on. Let me try something here. Let me try to relog into
119. 21:39it. It could just be a screen redraw issue. Hang on. I'm going to connect to server. I'm going to go here.
120. 21:54Archetypal Architect on YouTube meanwhile has a uh has a good suggestion. >> Okay.
121. 22:10Okay. Yeah, that's I'm in. Yeah, I'm in. But am I still sharing? Yes, I am. Okay. So, what you're looking at here,
122. 22:29what's this? What you're looking at here is no black bar
123. 22:43mods. Can you please [clears throat] pin my welcome comment to Dr. J. Mary says, "Hello, Dr. Jay." Okay, Eric, how good is that M5?
124. 23:00>> Please welcome my good friend, >> not Dr. J. >> Oh, not Dr. J. DJK to a regulars. Hey, DJK. Welcome, welcome, welcome, welcome. Um,
125. 23:13so what you're looking at here is this is my um my MacBook Pro. And so here's a little this. And let's
126. 23:25see, here's my terminal. So there's my little terminal window. So So basically the way the way I work with it is is uh
127. 23:34I've got Claude right here. This is Telegram. So this is Adam. So I can say, "Hi, Adam. How are you?" And then you'll notice up here at the
128. 23:46top right underneath Adam, it says typing right there. Um, and so, so Adam's thinking he's probably waking up
129. 23:58and then, uh, and then, you know, if I need to do any Let's see. Hey, Kyle, I'm here. I'm
130. 24:13doing all right. Bit annoyed at the busted YouTube Digest migration, but but [laughter] otherwise, Steady. How are you? I'm good.
131. 24:24Um, let's see. Um, Lord Digital Gods has sent us a migration uh instruction
132. 24:38set. So, we should be good to go as soon as I get off my butt. >> [laughter]
133. 24:53>> Okay. So, so I'm just over here talking to him. But what's cool about this is I'm, you know, I'm traveling for two
134. 24:59weeks, so I can just remote into this dude, this this uh M5 Mac and do all the configuring I need to do. So, I've got
135. 25:09I've got a lot of things I want to play with. Um, did you get rid of something there? No. YouTube comment archetypal.
136. 25:19How about setting up an AI avatar to run your channel while you're gone? That should be a quick project. [music]
137. 25:33[music] But I don't know that there's any AI avatars. Yeah, I lost my screen share. That's okay. I don't know that there's any AI avatars
138. 25:48that you can turn into just like a full-on interactive talking dude. Champion responds to human Kyle. That's
139. 25:56probably true. Um Eric was asking about the M5. Do I like it? Um I I got to say that the the MacBook Pro M5 in its in
140. 26:07its baller configuration is an incredibly small machine. like it's way smaller than my my M1. It's way thinner.
141. 26:15Um, it's very fast. When I was running [music] I I was running something on it. Oh, I I was running LTX Studio. I was running
142. 26:28LTX Studio locally and it got really hot. [music] So So if I'm going to be doing a lot of compute on it, I would
143. 26:37probably get get it some sort of cooling fan base. um just to just to improve the the the cooling of it. Speaking of it, did you
144. 26:45see Tim Cook is stepping down? I think thank god um [music] Apple needs to start innovating again.
145. 27:02Um but but to your point, I think Apple Silicon, they've been they've been very behind like like laughingly behind on the AI
146. 27:13stuff, right? Like just like it's ridiculous how far behind they are. And and like Siri has been a laughingtock of
147. 27:21the AI agent world for more than a decade, right? Siri's just an embarrassment. Um [music] and so you have this product that's kind of an
148. 27:31embarrassment and then everyone else is doing this AI stuff that's remarkable and you're [ __ ] Apple. So, like
149. 27:38you've got cash, you should be able to do this as good as anyone on the planet. You understand h the humanities more
150. 27:45than most other tech companies and and they just they just missed the ball on it. That said, their hardware division
151. 27:53is is making like the thing about the M5 chip is that it's got 18 or 20 GPU cores and each one has its own AI accelerator.
152. 28:01On the M4, you had a single AI accelerator that then routed it to all the different GPUs. This now has an AI
153. 28:08accelerator on each GPU. Like, it's it's incredible. It's incredible architecture. I don't know enough about
154. 28:15how this [ __ ] works to know if it's good. But what I can tell you is the the Quen 35 billion parameter, the 3.6 version
155. 28:24with the with the quant the quant 6, the quantiz 6 version is fast and good. It's got reasoning. It's it it's can it can
156. 28:34see images. You can send it images. It can see them. Um it's really quite good. So yeah, so I think Apple's going to
157. 28:40catch up by by having computers that are really strong inference machines and then a lot of AI inference is going to
158. 28:48move off of OpenAI and Anthropic as as these models get better and better and as they get easier to install, right?
159. 28:56which is again Apple should [ __ ] destroy it. Like like what Apple should have is like you turn on your Mac, it
160. 29:03says, "Do you want AI?" "Yes." "Do you want me just always to give you the latest model?" "Yes." "Do you want to
161. 29:09make pictures?" "Yes." "Do you want to make music?" "Yes." Right? Do you want to do programming? Ah, maybe. Right? And
162. 29:16then it just goes out and finds all this [ __ ] What are your priorities for these things? And then it's just constantly
163. 29:22updating them. Like that would be something Apple could kill and people would buy things just for that and then you'd run all your all
164. 29:31your AI locally. An Apple harness running all that stuff would be amazing. [music] But here's the thing. Tim Cook didn't call me.
165. 29:46All he had to do is say, "Kyle, what do you think we should do?" I'd told him. >> [music] [music]
166. 30:02>> Um, but yeah, I think their hardware is pretty amazing. It's such a slick form factor. It really is. And they're they're completely
167. 30:15redoing the MacBook Pro line um in the fall. So, they're going to have touchscreens now. They're going to
168. 30:21be thinner. They're doing a whole new architecture. I think that this this is probably the 10th, maybe not 10th, maybe sixth or seventh
169. 30:31[music] version in this same chassis. Tik Tok question. Kyle, what's your theory about why Apple set sat back on AI? Um, [music]
170. 30:42well, as innovative as Apple is, sorry, as innovative as Apple was, um, they tended not to jump on new technology early. They would let they
171. 30:56would let people sort of [ __ ] around and find out and then they would take what worked and and run with that. So, they
172. 31:03would they they're what I would call a fast follower or a true innovator, right? They're doing stuff in their R&D
173. 31:09labs. They come up with, you know, the the the iPad and the iPhone and and put that stuff out in the world and change the world with it. Um,
174. 31:24I I honestly don't know. They're a consumer they're a consumer computer company
175. 31:39even though a lot of engineers, you know, use them. So So they're not they're not really targeting engineers. Um
176. 31:49AI isn't ready for prime time yet. I would argue that if they had gotten in the game early, like if if Apple were
177. 31:56one of the frontier model companies along with OpenAI, Anthropic, and Google, like they would just be doing
178. 32:05large language models and and probably multimodal models on on just a they would basically be coming at it from a
179. 32:12very philosophically different approach where they would be balancing art and science where the others are just going full-on science.
180. 32:23So, it frustrates me that we don't have Apple's voice in that conversation. That said, the models are just now getting good
181. 32:32enough where Apple can sit back and go, okay, maybe we'll go with open source. We'll take some open- source thing. Like
182. 32:40the the Apple operating system is based on Linux, right? It's it's BSD. So they can they can take an existing
183. 32:48open-source framework and and build, you know, consumer grade stuff on top of it. So So they know how to do it.
184. 32:58I don't know. I I honestly don't know. It just, you know, it was either too early and unproven. It didn't really
185. 33:08they in the end they've got to sell more AirPods and Macs and Apple watches and right they've got to sell more [ __ ] So
186. 33:17if they made the calculation that AI features are not going to sell more [ __ ] Yeah, we'll just wait three years.
187. 33:24We'll we'll let these guys go battle it out. We'll just pay attention. But you know the fact that Tim Cook is
188. 33:30leaving, you know, I think I'm I mean I'm not the only one bitching about the fact that they missed it. Hey Kyle, I'm
189. 33:36building an app to automate social marketing for authors. What do you think I should do? Well, the first one is build it. [laughter]
190. 33:48The next one, okay, so here's what you should do, archetypal. You're building a a social marketing. You're going to automate social
191. 33:58marketing for authors. Automate social marketing for the social marketing automated social marketing tool for
192. 34:07authors. Use the tool to market itself you know something like that. Oh, you he thinks
193. 34:22he means like lovable. Oh, I'm building an app to automate social marketing. What do you think I should do in terms of what?
194. 34:32Yeah, that's in the queue. Are have you already built it? Are you building on on lovable? Make videos and smoking jackets
195. 34:39and tobacco pipes. Yeah, Chef Kelly, that would be good. Hey, Chef Kelly, you want to hear something funny? So, you
196. 34:46know, when you did the podcast and you sent me um you sent me snacks to to eat on. Actually, your Amazon order got
197. 34:55split into two orders. One half of it I got the day that I talked to you, but the other half has been sitting in in my
198. 35:02company's um little post office boxes for like a month and they're like, "Hey, you've got a package down here no one
199. 35:09picked up and we're about to we're about to send it back." I was like, "Oh, okay." So, I went down there and it was a whole bunch of candy.
200. 35:16Thank you. [laughter] Apple played AI like it was another nice upgrade while everyone else realized it
201. 35:23was the whole next game. Yeah, I don't I honest to God I can't I can't for the life of me figure out why Apple didn't
202. 35:31pounce on AI. I don't get it. I don't get it. Hey, you want to know? So, Chef Kelly sent me So, she's doing this thing with
203. 35:43um helping big companies replace artificial dyes with natural dyes. And so she sent me the these sample candies
204. 35:52that are that use natural dyes and they're tasty. They're good. But but here's the problem, Chef Kelly.
205. 35:59[laughter] Because because they don't look at as as bright and because they use natural coloring, I'm like, "Oh,
206. 36:06these must be healthier for me." So I I ate like eight of those little packs today of like fake Skittles and fake
207. 36:13Starburst shoes, things like that. [laughter] >> [music] >> Okay. Oh my god.
208. 36:28[music] Another big food ruse.
209. 36:47I gave it a very simple prompt and it spun up agents to build up much more complex business. Already got payment set and cold outreach.
210. 36:58[music] I mean I think a thing to do archetypal here here's here's a thing that I think we could all start experimenting with.
211. 37:15is whether you use Perplexity Computer or Manis or Gen Spark or Open Claw, whether it's local or in the cloud, doesn't matter.
212. 37:30Once you start getting your head around agents and and to be clear, the thing that I'm about to say, I'm not doing,
213. 37:36but I think this is a thing to do. So, you've come up with this automated social media marketing for authors platform, right?
214. 37:49I would have I would go to Claude probably put Claude into deep thinking mode and have Claude tell you,
215. 38:02give me a whole strategy for how I can market this thing and break it down into components. Here's how I find authors to
216. 38:10target. here's, you know, the the campaigns I put together. Here's, you know, you need to come up with a
217. 38:16creative campaign and you need to have a creative brief for that. Have it come up with a whole plan for how you're going
218. 38:21to do this. And then, you know, once you get people to a landing page, how are you converting them? Have it give you
219. 38:27the whole strategy for all that stuff. And then just see if you can use agents to do all of the components that it it
220. 38:36designs. Like I kind of feel like this idea of a single a single operator doing all of the all of the execution of
221. 38:45of a full strategy as if you had a 20 or 30 person company. Uh might be might be a uh might be worth I I think it would be
222. 38:56worthwhile if nothing else just to learn, you know, just to learn [music] because Apple sucks. That's why they didn't punch on AI,
223. 39:09but really they're trying to save money. Well, this is Listen, here's the deal. Steve Jobs put Tim Cook, you know, at the helm of Apple, you
224. 39:20know, on his deathbed. And I I understand it, right? You have a like Tim Cook was a CFO, right? So, he's a
225. 39:29numbers guy. So, what Steve Jobs, I assume, was banking on was, "Hey, Tim, you know, go keep the company running."
226. 39:37And and and Tim Cook's done a good job. Like, basically, all of the innovations that they did, they've maintained.
227. 39:42They've done a couple of, you know, decent innovations along the way, but Apple has been in a in a maintaining
228. 39:51mode for since Steve Jobs died. Tim Cook stepping down makes room for someone to come in. I like I don't know
229. 39:58who's coming in but I hope who they put in as CEO is someone who's got some [ __ ] vision right [music] you know there's maintaining and you
230. 40:09know maintain the growth maintain the company build the cash reserves so that they can survive anything
231. 40:17so they've done that well it's just that you know the the the quality of their products has gone down the quality of
232. 40:24their software has gone down and their innovation pipeline is just crap. [music] Making data centers is different
233. 40:36business than selling hardware. Did you see that? Um who was it that went into business with um CBRE, the the big uh
234. 40:46commercial real estate company? They're they're now training people to install fiber optics in data centers.
235. 40:54Was it Amazon? Might have been Amazon. or [music] maybe it's Meta. I think it's Meta. Meta and CBRE are are training
236. 41:04people how to be optical cable technicians for data centers. [music] [music]
237. 41:20I'll tell you why I don't like Tim Cook. Same iPhone every year. Change one or two things. Call it a new model. This is
238. 41:26but this is what I'm saying. That was Tim Cook's job. Tim Cook's job was maintain what we've got. Don't lose, don't blow the lead we have.
239. 41:39Except the cost of it is when something like AI comes along, for whatever reason, they chose not to pivot. They
240. 41:46chose not to embrace it. They chose not to lean into that. And you know, maybe it was a calculated bet to say that all this generative AI
241. 41:57stuff is just going to be interesting to engineers and we're a consumer company, but I think they could have driven the
242. 42:04acceleration of AI by getting in early. And they still may. I mean, here's the thing. If you if you're if you're talking
243. 42:13metaphorical comparisons to the early days of the worldwide web, the first three years of this channel was like 1995, 1996.
244. 42:26The last 6 months is like 1997. The difference between 96 and 97 is that 97 is when the hype really started with the worldwide web.
245. 42:37So AI has taken a little bit longer to sort of buzz up. Like I know it's people are talking a lot about it, but it's
246. 42:45just now starting to get stupid, right? Just stupid marketing campaigns and stupid just it's it's getting stupid now. [music]
247. 42:57So we're very early. So here's the thing. Here's the thing with Apple. It's possible, highly unlikely. It's possible they've
248. 43:08been brewing up [ __ ] you know, in the back 40 for a couple of years and they're just going to come out with something and it's going to be
249. 43:15remarkable. I doubt that's the case. I think they just chose not to do it. But if at some point Apple says, "We're
250. 43:22going to go all in. We're going to become a frontier model company," um, and they come out with a model that is
251. 43:28really good, everyone will forget about the fact that they waited. They'll call them, "Oh, they were really smart. They
252. 43:33let everyone else make all the mistakes and they came in and figured out how to do it." I mean, I wouldn't be surprised if Apple
253. 43:41comes out with some sort of chips to um to put in their own data centers to train their own things. [music]
254. 43:48Amow to invest up to another 25 billion in anthropic. So amazing. Just amazing. Personal models. Yeah, I think personal
255. 44:01models are going to be big. Or maybe it'll be a shoe company like like Albert Apple shoes. Yeah, apparently they've
256. 44:09got Apple glasses are still rumored to be coming out like augmented glasses. Posted the app in the salon and look what I made. Okay, cool.
257. 44:22So, if anyone wants to play with um Archetypal's app in the salon, go to look at look what I made. Let me jump over there. [clears throat]
258. 44:43This is an automated social marketing app specifically for authors built with Pulsia.ai. I've never heard of Pulsia.ai.
259. 45:00You direct it, the AI writes it. Nothing posts without your say. Start free. No, no card required. Book title, book blurb, genre.
260. 45:13Notebook, LM audio. Oh, that's cool. TJI, your voice. That's cool. AI slop spelled backwards. That's pretty That's pretty genius.
261. 45:31[laughter] I slop spell backwards. That's pretty good. Um, one thing I would fix is your your button in the top that says create
262. 45:42account. The color of the font. I can't read the button. [clears throat] Make it white like this. Generate a week of posts. Tell us about
263. 45:55your book. We'll create seven days of platform ready social media content tailored to your genre and audience. That's super cool.
264. 46:03Ink Signal. Good name. Ink Signal. Nice. Oh, we lost your screen share. Nice. Thanks, Apple. Here, case in point.
265. 46:20Hey Tim, do you think we need the the quality control team here at Apple? Nah, we're good. Our stuff's good enough. We
266. 46:28don't need QC anymore. [laughter] Oh my god. Oh my god. Yeah, right. That's fantastic. Yeah. No, you're all super good. Super
267. 46:43good. Yep. There you go. Well, you're in. Look what I'm I just lost the screen share again. God damn it. I really should restart my machine
268. 46:54before every uh [gasps] before every time I go live. You know, this is maddening.
269. 47:11[whistles] At least once a week. I know. >> [music] [music]
270. 47:27>> A Dr. J song in Look What I Made. Let's go look at a Dr. J song, shall we? [clears throat] Bones by Untamed Acid House.
271. 47:43I love that. Dr. J untamed acid house. Nice. All right, here we go. Here we go. Good people. Broken piano. Close mic.
272. 48:05I still have your sweater in the bottom of a drawer. I tell myself tomorrow I'll be brave enough to throw it. But
273. 48:13tomorrow keeps on coming and I [music] keep on keeping it. And I lie to everyone I [singing] love and tell them
274. 48:19I don't know it. The elephants they go back to the bones of their dead. Stand there in [music] the silence doing
275. 48:27[singing] nothing with their heads bowed down. I didn't know. I didn't know. I [music] didn't know. Grief had a body of
276. 48:36its own. They told me there were [singing] stages. They told me there's a door. [music] They told me one day soon
277. 48:42I wouldn't feel this anymore. But I keep coming back to the bones. To the bones like the elephants. I keep coming back to the bones. [music]
278. 48:56To the bones. I don't know how to leave you in the ground. I keep coming back. I keep coming back. [music] I keep coming
279. 49:05back. I keep coming back. I keep coming back. I keep coming back. [music] I keep coming back. I drive past your
280. 49:15mother's house. And I don't know why I go. She hasn't seen me since the funeral. She wouldn't want me to, but I
281. 49:22sit [music] across the street and watch her water all your plants. And I whisper to the windshield [singing] everything I
282. 49:28never got to. The crows remember [music] faces. I bet they remember [singing] you. I bet you fed them something and
283. 49:36they kept their promise to remember. I didn't know. I didn't [music] know. I didn't know. Grief was something you
284. 49:44could grow. They told me [music] there were stages. [singing] They told me there's a door. They told me one day soon I wouldn't feel this
285. 49:52anymore. But I keep coming back. The bones. The bones. Like the elephants. Keep [music] coming back. The bones.
286. 50:05The bones. I don't know how to leave you in the ground. [music] >> I keep coming back. I keep coming back. I keep coming back. I keep coming
287. 50:14[music] back. I keep coming back. I keep coming back. I keep coming back. I keep coming back. [music] I thought I came.
288. 50:30>> That's pretty good, Dr. J. It's really good. Very, very good. Uh, Producer AI is now Google Flow Music. That happened
289. 50:36a while ago, I think, Brandon, didn't it? But did they sell to Google or or is Google just using them? They bought it
290. 50:45and rebranded it. Oh, okay. Google Flow. Hey, do you want to come up and talk about So, so this past weekend, um Rick
291. 50:53Macaulay and producer Brandon were vibe coding and uh they got they got screwed out of out of a top five finish by some crappy AI
292. 51:06judging system. >> Tell tell us about your weekend. What did you do? >> Yeah, so uh we got the gang back
293. 51:15together. Rick McCaulay, the AI guy, and a couple of other team members, we went down to actually we got doubly screwed
294. 51:23because we had aligned on a healthc care track. We were going to cure cancer. And uh at the last second, while I was on
295. 51:31the plane on my way to Miami, they're like, "Just kidding. We're not doing the healthcare track anymore. We're doing
296. 51:36the enterprise track." And so in true hackathon spirit, we pulled up our plan B, which was Upskill USA,
297. 51:46>> and ended up vibe coding in uh nine hours a complete portal for enterprise companies to figure out how much money
298. 51:57they're leaving on the table by not activating their employees with AI. and uh thought we had a pretty good shot,
299. 52:04but then they decided to also let AI pick the top five projects and we did not place because we did not uh we did not
300. 52:16put our efforts towards impressing AI [laughter] >> with a really cool product and a really cool demo and um so we could maybe take
301. 52:28a quick look at that. But yeah, it was a lot of I mean, those those hackathons are a ton of fun. If you've never done
302. 52:34one, they're they're just wild because you're just in the zone for nine hours. And we put in months >> and build something that's actually
303. 52:44real. I don't know if you saw I put StoryVine through um and uh you know it actually does a free audit of their
304. 52:51website using firecrawl and uh some API stuff and it's it's it's a pretty cool program but not not worthy for this.
305. 53:01>> You want to demo something? >> Uh sure. >> We'll see if my screen share does better than yours. >> Yeah, really crazy.
306. 53:15So share screen with audio upskill USA. So it's upskillusa.lovable.app. And this was actually Rick's original idea
307. 53:30is he wanted to help upskill and retrain employees you to be ready for AI. And so we took that and we pivoted towards a system that allows
308. 53:43the AI to understand what you do and where your gaps are and what you can automate completely away, what you can
309. 53:52augment with AI and what you have authorship over what is uniquely human in your role. And so >> you built the entire site and you get
310. 54:01your free AI readiness audit. So, we can come here and do something like I don't know, give me your favorite uh
311. 54:08enterprise company, big box company. >> Um, let's go with um Salesforce. >> So, I [clears throat] just type in the
312. 54:20website and we got a little lead magnet here. So, it asks for your email address so we can send you the report and capture who's using it.
313. 54:30And so, it's analyzing Salesforce.com. So we not only use firecrawl to understand the website but we also use
314. 54:36an API to look at like the publicly available how many employees do you have that type of stuff and it figured out
315. 54:43that uh they're losing $2.8 8 million a year by not fully implementing AI. And even though they are 75% out of a 100,
316. 54:54they got a good workforce score. What we're losing is sales representatives spend an excess amount of time on data
317. 55:01entry and lead tracking. Customer service agents are inundated with repetitive inquiries. So it does this whole analysis and then what
318. 55:11>> sales sucks. [laughter] Their own tool sucks. >> Yeah. What you can do is you can log in and then you can have your employees
319. 55:23take an assessment. Uh and actually I'll just show you and the um and it would show you let me log back out here.
320. 55:39It shows you for for as an employee. These are the amount of how much time you can save by automating invoice data
321. 55:49entry. Uh but things like vendor relationship management should stay in completely human. So it shows you how much you and then
322. 55:57you can build your own um emulator. So the core problem we're trying to solve is everyone's trying to
323. 56:08do AI from top down and say here's cloud code go use it in your workforce. >> We're finding that the bottom up
324. 56:18approach is much more effective. Cisco did a pilot program where they had 800 of their employees just use AI freely
325. 56:27with governance obviously, but the low-level employees that were actually on the front lines of the company found
326. 56:34250 novel use cases for AI that the upper management team was completely blind to. So, this idea of empowering
327. 56:43users to automate little parts of their mundane day-to-day is a huge unlock. And uh I'll play this. This is a two-minute demo. And
328. 56:54we put this together and in about 40 minutes. And what's really cool about this is that Sunno did the audio for this. We asked, you know, to
329. 57:08do spoken word and it did a really good job. >> Oh wow. >> Enterprise AI is at a breaking point.
330. 57:22Over 70% of AI initiatives fail to scale beyond pilot millions are being spent with little to show in operational impact.
331. 57:34Not because [music] the technology doesn't work, but because the approach doesn't. AI has been deployed top down.
332. 57:44The [music] seauite defines the vision. It manages the systems. But the people who actually do the work,
333. 57:53they're not part of the solution design. The future of AI isn't top down. It's employed driven, governed by IT, [music]
334. 58:02and aligned to the business. At Upskill USA, we don't replace your workforce, we upgrade it. Start with a free audit.
335. 58:17Let your employees take the assessment, then map every role to three outcomes: automate, augment, author.
336. 58:26Because the companies moving fastest aren't removing humans, they're amplifying them. [music] Employees enter the emulation station
337. 58:35where they break down their work, identify inefficiencies, and build AI powered workflows tailored to what they
338. 58:42actually do. Every action is structured, every output is measurable, and it maintains full governance and control.
339. 58:53This isn't a future [music] state conversation. Your employees are already solving problems every day without AI.
340. 59:04The question is, how much longer can [music] you afford that? Upskill your people. Unlock what already
341. 59:13exists inside [music] your organization and turn AI from experimentation [music] into execution.
342. 59:32Beautiful. There you go. Thank you. Yeah. Um, the uh the first half of that, by the way, all the B-roll was Google Vids. We just
343. 59:41gave the uh VO, we just gave the script to Google Vids and it came up with all of those uh visuals of the people
344. 59:49walking around an office and stuff like that. >> Wow, that's cool. >> Cool. So, what's the plan with it now that it's built? [clears throat]
345. 59:58Uh we're Rick's talking about taking it down to some VCs down in the Miami area and um obviously we got to fine-tune and
346. 1:00:08figure out how to pay wall and things like that, but you know, it's pretty far along for a 9-h hour app. [laughter] >> So amazing.
347. 1:00:18>> It's amazing. Have you had anyone play with it yet? Kick the tires? >> Uh not yet because it's only been out in
348. 1:00:25the world for about 36 hours. Um, but >> we're uh we're open to uh beta testers if you know anybody in the salon.
349. 1:00:36>> Uh I wrote an article about our experience over the weekend in the in the community feed today. So
350. 1:00:41>> Okay, cool. And there's a link to it in there, right? >> Yep. >> Beautiful. Beauty, beauty, beauty. Well, >> congratulations.
351. 1:00:49>> Thank you. Uh see you on the other side. >> See you out there. All right. I'll kick the tires for you. There you
352. 1:00:59go. Source Camp. Source Camp going to go uh figure out what she can do with her skills. She's [laughter] g She's going to upskill.
353. 1:01:08I don't know how you upskill from from where you are. [laughter] You're pretty much doing it. Um yeah, crazy times.
354. 1:01:18Crazy times. What other thoughts? What other thoughts? Why am I not more productive? Because I'm spending four
355. 1:01:23hours a day doing busy work. Salesforce is its own oxymoron. Yeah. It's it's just I mean what's so impressive about Salesforce
356. 1:01:34is also its greatest weakness which is once you put all your [ __ ] in there the the the switching cost is essentially
357. 1:01:44gargantuan. So like no one uninstalls it. And so just you've got decade customers, you know, decadesl long customers, no
358. 1:01:53matter how bad it is, they're just going to keep using it because the thought of changing it is uh Lori K, that's
359. 1:02:01awesome, Brandon. I don't honestly don't know anyone who uses their Salesforce account. We do at Story Vine, but not
360. 1:02:07because of me, because of other people. [laughter] >> [clears throat]
361. 1:02:24>> Hang on one sec.
362. 1:02:45Details.
363. 1:03:07All right. Looks like there was a a compromise of a third-party developer library of Axios.
364. 1:03:22The security team has rotated code signing material for all OpenAI MacOss apps. Please update any MACOS apps before May 8th. All right.
365. 1:03:38>> [clears throat] >> Kyle, two things. One, you're not sharing your screen yet. Uh, and two, I didn't know if that was a share screenw
366. 1:03:43worthy moment, but also um, [clears throat] Lovable had a non leak that was reported as a leak today. So,
367. 1:03:51you might see that over the next couple days if it gains any traction. >> What it was was initially [clears throat] when they first launched
368. 1:03:59Lovable, when you published your project publicly, >> Yeah. It published your chat history with the project just like chat GBT when
369. 1:04:11you share a chat with the philosophy of build and public. What lovable didn't account for was that vibe coders didn't understand security
370. 1:04:19and were putting their API keys directly into the chats instead of the security. >> So they stopped >> doing that six months ago. They they
371. 1:04:29stopped sharing the chats because they didn't want to expose people. But some journalist found, you know, history of
372. 1:04:37this and and wrote a hit piece on them and now everyone's running with Loveable got hacked and they're like, "No, we didn't." So
373. 1:04:44>> yeah. Yeah. That's wild. That's wild. There's a new version of Goggout for for uh uh for OpenClaw,
374. 1:04:55which if you're not doing OpenClaw, you won't care. I'll try to scare share my screen for a bit. I'm just going to look
375. 1:05:03at at um some X stuff to just see if there's any cuz it looks like this is going to be a big week. I would assume I
376. 1:05:09think when does OpenAI do their their launches? I think they do them on Thursdays, right? So maybe tomorrow we'll hear from them that uh
377. 1:05:24Mr. Strawberry, you're not a emotionally prepared for this model. He's been really hyping the [ __ ] out of this new open AI thing.
378. 1:05:32Um, I know a lot of people who have it but they don't use it. Oh, Salesforce. Yeah. Fastest AI side side hustle to 500k a year if you can
379. 1:05:45actually execute millions of small businesses, almost none of them willing to embrace AI right now. Yeah. Yeah. Yeah.
380. 1:05:54Did you see open clause founders TED talk? I have not. I saw that there was one, but I didn't see it. Oh, right. It's 420.
381. 1:06:10[laughter] Look at the dudes getting stoned in their little helmets. [laughter]
382. 1:06:27Oh, mind your Tik Tok censorship. Yeah, I'm sure we'll be we'll be uh given warnings here in a moment. Amazon
383. 1:06:34announces they will invest up to $25 billion in anthropic. Fascinating. Peter Diamandis, we're not building
384. 1:06:45tools anymore. We're witnessing the emergence of something between software and life, between servant and peer.
385. 1:07:10Yan Lun left Meta because Zucker Zuckerberg wouldn't bet the company on Jeppa joint enabled predictive architecture.
386. 1:07:20Oh wait, joint embedding predictive architecture. Last week his group dropped the first Jeepa that actually trains end to end
387. 1:07:31from raw pixels. 15 million parameters single GPU a few hours. The timing is not coincidence. For four years, Meta has been able to
388. 1:07:44wait. Meta has been the house that Jeepa built. Lacun published the original paper in 2022. The architecture was supposed to be the
389. 1:07:56escape hatch from LLMs, the path to robots that actually learned physics instead of hallucinating about it. Every
390. 1:08:04version shipped fragile, stopped gradients, blah blah blah. Meta kept funding LLMs. Llama shipped. Llama
391. 1:08:10scaled. Llama got beat by Quinn and Deep Seek. Zuck spent $14 billion dollars to buy Scale AI and install Alexander Wang.
392. 1:08:19That was like the 23-year-old exec that he put in above Yan Lun. That was basically like [ __ ] you. Lun's research
393. 1:08:27kept winning papers and losing product roadmap. He left. He started AMI labs and said publicly that LLMs were a dead
394. 1:08:34end. Now the paper la world model one regular riser replaces the entire pile of furistics. I don't know what that
395. 1:08:43means. Oops. Hang on. I got to plug in my machine. My machine's about to die. Got to plug in my machine.
396. 1:09:04All right, there we go. Project the latent embeddings in onto random directions. Run a normality test. Penalize deviation from gauian. The
397. 1:09:20model cannot collapse because collapsed embeddings fail the test by construction. Hyperparameter search went
398. 1:09:27from 0 n to the 6 polomial to 0 log n logarithmic. Six tunable knobs became one. Well, duh. Everybody knows that
399. 1:09:41one regularizer replaces an entire pile of huristics. You know like we talk about here we talk about normalizers and regularizers and
400. 1:09:52huristics with the latent embeddings projected into the random directions which once you run a normality test
401. 1:10:00against the deviation from Gausian [laughter] I mean I mean who wouldn't want to deviate from Gausian right am I right
402. 1:10:14all right whatever Yan Lun's up to something. Apparently, [laughter] we need paint. [laughter] Cam Ken, what?
403. 1:10:27Oh, Kyle, that was so five days ago. [laughter] Pate has been summoned. Oh my god. What if everything you see, feel, and
404. 1:10:36experience is part of an high highly advanced computer simulation? I'd be [ __ ] impressed. That's what I'd be.
405. 1:10:53Elon Musk. Actually, AI and robotics will mean that everyone can have a penthouse if they want. No, there's only one penthouse in a
406. 1:11:05building. Not everyone can live there. >> [laughter] >> Unless you just make really tall stilts and put a single floor really high in
407. 1:11:16the air. [laughter] And I suppose with robots and infinite abundance you could do that, but no. No. Okay, whatever. Where my Gen Xers at?
408. 1:11:3045 to 60. Zero tolerance for stupidity. There you go. Speak sarcasm fluently. That's really funny.
409. 1:11:45Let's see. Reporting for duty. Every company needs a cranky gen Xer in the corner office
410. 1:12:06to tell the youngsters. Why? What they built
411. 1:12:25is a pile of And then we're going to make a a poo [laughter] a poo emoji.
412. 1:12:41>> [laughter] [gasps] >> There you go. All right. Beautiful. [clears throat] I just deviated into Gausian. I want a
413. 1:12:57floating [laughter] island. Yeah, exactly. Everybody with AI and robots, everybody could have a flo floating island. [laughter]
414. 1:13:05Hey, how come there's no oceans? Oh, everybody wanted an island? Really? And they all got one? Yeah, it's massive
415. 1:13:13abundance. It's hyper abundance. So, is there any beachfront left? No, we got rid of that. Everyone wanted an
416. 1:13:21island. So, isn't that just land? Yeah, pretty much. Where'd the [clears throat] water go? We buried it. [laughter] Humans suck.
417. 1:13:32Charles, cool app. Fairy Tale Revolution. It didn't give me posts for X and Blue Sky, but did for Instagram only when I
418. 1:13:39was choosing. I chose X and Blue Sky, not Instagram. Maybe just user error. I don't know. Claire. Okay, cool. Claire's given feedback.
419. 1:14:06plugins. So, I don't see a lot of buzz about new models coming other than Mr. Strawberry. And generally, we'll get you you'll get
420. 1:14:19some noise.
421. 1:14:38Tell me.
422. 1:15:01Oh, you want to hop up and talk about that, Brandon? >> Hello again. >> Uh, yeah. So, next month for mastermind,
423. 1:15:13I'm putting together a four-week prototyping sprint for Lovable. So, we're going to go over in the four
424. 1:15:20weeks. Basically, if you've never been down the rabbit hole of vibe coding, what vibe coding is and what Lovable is,
425. 1:15:28and then how to even think about how to come up with an idea and how to use Lovable to build it and launch it,
426. 1:15:36stress test it, and it's going to be a really cool four-week program, and I'm excited about it. >> That's super cool. When does it start?
427. 1:15:44>> Uh May uh right May 6. >> Okay. So, if you're in the master if you're not in the mastermind, go to community.thesalon.ai.
428. 1:15:53Look for the mastermind. There's lots of information in there about it. Join it if you want to be part of that. Um, are
429. 1:15:58you going to are you going to go into one of the things that flumxes me a bit about vibe coding because I just haven't
430. 1:16:04done it enough is once I'm past the initial build, like going back and fixing stuff and tweaking it, you seem
431. 1:16:11to be pretty good at that. How do you do that without screwing things up? Is that what are you gonna cover that session?
432. 1:16:18>> Yeah. Yeah, we'll we'll definitely talk about that. I mean it it's more of an art than a science, but >> uh yes.
433. 1:16:27>> Okay. All right. I mean like even that's good to hear. That's really cool. >> Yeah. >> All right, Groovy.
434. 1:16:36>> That's exciting. Okay, cool. And and information I assume is in the uh in the in the mastermind. >> Yes. Andy just posted it a little bit
435. 1:16:45ago which was reminded me to tell you about it. >> Oh, cool. Oh, good. Awesome. Beautiful. My first sun song was waiting on
436. 1:16:56strawberry ah memories. Yeah. Remember when Mr. Strawberry had all the all the the [clears throat] rumors about what
437. 1:17:02was coming and then he sort of lost his credibility and now he still says all this stuff and I'm like if only I believed you. [laughter]
438. 1:17:11I just don't.
439. 1:17:29All right, hang on. [clears throat] This is Greg Brockman talking about the new OpenAI model.
440. 1:17:44So I think of spud as a new base as a new pre-train and that we have had this uh I'd say it's like we have maybe two
441. 1:17:54years worth of research that is coming to fruition in this model. It's going to be very exciting and I think that the
442. 1:18:01way that the world will experience it is just improved capabilities and that for me it's never about any one release
443. 1:18:08because as soon as we have this one release it'll be an early version of what we have coming. We'll do much more
444. 1:18:14of each of these steps of the improvement process. And so I think that where we're going is almost just we have
445. 1:18:21this engine of progress that just moved faster and faster and that spud is just one step. So, I think of Spud.
446. 1:18:27>> All right. So, Spud's their new model. Horrible name. All right. Deep Cam Live. You can swap your face live on camera in real time. Okay.
447. 1:18:45Wait. Some of you don't think. Okay. That's nothing. That's nothing. Elon Musk. Sam Alman.
448. 1:19:00Kimi 2.6 Six video hero sections. Cinematic aesthetic autocomposited
449. 1:19:21[music] [music] >> [music]
450. 1:19:39>> I don't know what that is. I haven't played with it. Has anyone else played with that? Kimi [whistles]
451. 1:19:56>> [gasps] [clears throat] >> UBI works.
452. 1:20:13>> [laughter] >> Wait, look at this. They have I didn't know
453. 1:20:28someone got someone [laughter] someone got into my my uh my CrossFit training. Somebody recorded me in my CrossFit training. Who
454. 1:20:42did that? Come on, you guys. [laughter]
455. 1:20:57I'm 37 years old. What's the best life advice you can offer me? Go sit on a bench. [laughter] Just wait it out. It It'll show up eventually.
456. 1:21:11>> [laughter] >> Be honest. Do you talk to yourself? Be honest. Do you not? [clears throat] Wait. [laughter]
457. 1:21:32Oh, man. All right. It's 9:30. Listen, I have to go because I have to get up bright and early. I have a 7:30 flight tomorrow morning.
458. 1:21:41I'm mostly packed. [gasps] I'm a little sleepy, so I'll eventually just take a nap tonight. Um, so listen, you're on your own this week,
459. 1:21:56so a couple of things. Um, don't take any of this [ __ ] too
460. 1:22:10seriously right now. We don't know what's coming. We don't know what's coming. We don't know what's here. I would say just play.
461. 1:22:19Figure out the [ __ ] that that brings you joy this week and just go do that. You want to vibe code? Go vibe code. You
462. 1:22:27want to go play in the dirt, go play in the dirt. I think that's the that's the thing until until something comes out where
463. 1:22:34it's clear that's like, oh, we got to deal with this. I don't think we're there yet. I think we're I think we're in a bit of a holding
464. 1:22:42pattern. And if and when something like Mythos comes out that really does change things, then we can start to talk about
465. 1:22:50what's it mean. But like, I don't even know what it [ __ ] means to have these autonomous AI agents yet.
466. 1:22:57It's not like I don't get them. I get what they are. What I don't get is why should we give a [ __ ] and I know it's in there. I just don't
467. 1:23:09know what it is yet. And I'll figure it out. Or I'll figure out that it wasn't in there. I don't think so, Brandon. I
468. 1:23:22[clears throat] I looked at scheduling. I don't think that's that's actually going to be the case. Um,
469. 1:23:29so what I will do is this. If I have an opportunity during the day to go live when I'm in between [ __ ] I'm doing, um,
470. 1:23:39I'll just do that. The thing that would precipitate that if is if some some big model comes out that's actually really
471. 1:23:45interesting and good, I'll probably go live and talk about it if I've got the time. But I'm not going to have a ton of
472. 1:23:52time this week and probably not next week. So, so the next two weeks are going to be a little thin, but I'll uh
473. 1:23:58I'll put some videos out and I may go live sporadically here and there just on TikTok. Okay. We'll be running a muck without you.
474. 1:24:07Have a good week, Kyle. Thank you very much much. Denver is the worst. I agree. [laughter] I mean, it's lovely and all. Um but
475. 1:24:18yeah. All right. Cool. Peace out everybody. Um, have yourself a fantastic week and I will be in touch. Um, jump into the AI salon. So, oh,
476. 1:24:31important announcement. You may know this, but we've got some folks here. Um, when I come back from this twoe gauntlet
477. 1:24:39of of speeches and meetings, the AI salon is moving inside or the AI learning lab is moving inside the AI
478. 1:24:48salon. So, that's going to happen the first week of May when we come back. So, I'm really excited about that. So, get
479. 1:24:54ready. Bring all your friends in. Let's get some activity going in there. I want to get the I want to make the inside of
480. 1:25:00the salon a vibrant um you know, bustling place again. We've been bustling out here on the edges. Time to
481. 1:25:07bring it all inside. Time to concentrate this energy um and get people really excited about it. So, I'm super excited
482. 1:25:13about that. So, that'll happen upon my return. Lots of lots of fun stuff coming. All right. Cool. [snorts] Peace
483. 1:25:19out. Have a beautiful evening and rest of the week. And I'll uh I'll post in the salon and let you know what's
484. 1:25:26happening, what's going down. All right. Bye, everybody.