8/14/2025 - Exploring Pika Social, GPT-5 Pro, and Vibe Coding

Transcript

1. 0:00Maybe sing him like slow. [Music]
2. 0:17His beard was a looking guy. She sat on a stool and he said, "What do you want?"
3. 0:33She said, "Give me a love that don't freeze up inside." You said, "I have melted some hearts in my time, dear." [Music]
4. 0:48But to sit next to you, well, I shiver and shake. If I knew love, well, I don't think I'd be here.
5. 1:04>> Asking myself if I've got what it takes. [Music] I do your blue heart.
6. 1:20has done. [Music] Turn what's been frozen for years into a river of tears.
7. 1:36[Music]
8. 1:51Thursday night. Good evening, good people. What's going down? What's shaking? What's happening? Steo, welcome
9. 1:58from Down Under. Good to see y'all. Hope you're all doing well. Oh man, I've been having fun with uh with Pa Pab Social. Pika Social.
10. 2:14That's been a that's been a fun [Music]
11. 2:36I have some thoughts on on Pika Social from a development standpoint. [Music] And from a uh how we're going to adopt
12. 2:50AI standpoint, I think there's some I think there's some really interesting things to learn from what they're doing.
13. 2:57So, we'll play with we'll talk about that tonight. Champy's in fine form. He is in fine form tonight. What a week. Looking to get my dose of
14. 3:07laughter tonight. Well, hopefully hopefully I got some comedy in me. Last night I was just a cranky little [ __ ] for no reason. [Music]
15. 3:25Oh, I forget. What does Pika cost? I think Pika is like one of the normal video things. They probably got a $10
16. 3:31tier that's useless and it's probably a $30 tier where you get some decent amount. Um, the Pika social app. So, there's Pika's
17. 3:41got two apps. They've got the Pika app where you do video generation and then they have Pika Social which is in
18. 3:48limited preview right now. You can only get in if you have an invite and if if you get invited then you get three
19. 3:54invites. It's one of those things. It's iOS only so it's the iOS app. Um and they're kind of I don't want to say they're competing
20. 4:07with Tik Tok. They have created a Tik Tok like app. Um, and right now it's free. They're I think they're just trying to understand
21. 4:16what they've built and what they've got. Um, I think it's a really interesting lesson in [Music] in um resisting the urge to
22. 4:33add more and more features. It is a It is a frustratingly limited feature set and I I actually think it's kind of
23. 4:43smart. And I'll I'll tell you what I mean in a bit. I'll show it to you. In fact, let me make sure my iPad's charged so we can play later.
24. 4:54We will play later. Yes. Yes. Yes. Yes. Yes. Yes. Danielle's been having fun with it. Yeah. It's really fun. It's
25. 5:02really they're they're really on to something here. Mr. IT, what's happening [Music]
26. 5:23through Blue Telescope. Looking at the world tonight through blue telescope. Wish I may, wish I might not see what I see.
27. 5:38Sheet metal on sheets of ice. Look through this blue telescope. [Music] Down a moon struck a road tonight.
28. 5:55[Music]
29. 6:11Yeah. Yeah.
30. 6:26[Music] Every time I see you now, get that look
31. 6:41in mine. [Music] Every time I see your mouth, I hear that smile. The early misty morning light I heard
32. 6:54the engine turning in the old f outside. [Music] You were leaving me again today.
33. 7:10You will convince me again today. You're leaving this hard time looking for someone else's cold and rain. [Music]
34. 7:26So long [Music] now. Don't you cry. [Music]
35. 7:41So long, Susanna. [Music] Don't you cry for me. [Music]
36. 7:56[Music] Do you know that Chad GPT5 Pro is now smarter than 99.9% of the humans on the planet? Did you
37. 8:12know that is true? [Laughter] Oh man. Welcome everybody. Welcome,
38. 8:28welcome, welcome. Hope you're doing well tonight. [Music]
39. 8:59Wraith desperately hating his old place. Dreaming to discover a new space. [Music] desperately hating his old place. Dream
40. 9:14to discover a new space. Buried himself alive inside his basement. Tongue on the side of his face. Working away on
41. 9:25displacement. What it would take to survive. Cuz when you're done with this world, you know the next is up to you.
42. 9:40And for once in his life, it was quiet. As he learned as he learned how to turn with the tides [Music]
43. 9:54and the sky was a flare and came up for air. [Music] All right. All right. Enough bad song playing and bad song singing. Played
44. 10:07with Pika AI today. So much fun. Yeah, Pika. Pika. I'll tell you what Pika is doing well. Um, and I think their social app really
45. 10:17exemplifies it. Um, most of the frontier model companies, most of the video companies, let me switch my videos here. So the people the
46. 10:29good people have a good view. This simal casting on Tik Tok and YouTube is special. It's something we've chosen to do.
47. 10:42Who is not smarter other than Pate? Exactly. I I would say Pate Pate's probably right up there. Um uh what was I talking about? What was I
48. 10:57talking about? Oh, Pika. What's Pika doing? Well, okay. So, most of the frontier model companies, most of the
49. 11:07creative companies, the image generator companies, the video generator companies, the song generator companies, um, are kind of assuming
50. 11:19that their audience wants to geek out. I don't know. Oh, don't worry. Okay. Um,
51. 11:40they they assume that their users are going to be people that want to geek out on prompting, that want lots of
52. 11:45controls. If you look at a tool like Leonardo.ai, it's got all sorts of [ __ ] in it, right? and all sorts of different features and
53. 11:53dials and it's like the cockpit of a 747 and you can tweak it and you can make luras and styles and whatever videos all that stuff.
54. 12:06What Peak is doing is they're kind of um they're experimenting with like the video generation equivalent of casual
55. 12:16gaming. So, so remember we used to have the these big AAA video games, right? And then like the iPhone came along and
56. 12:23the app store came along and this whole idea of casual gaming where you can just have a game that you can just play with
57. 12:29your thumb while you're standing in line, just casual [ __ ] around, right? You know, shoot basketballs at a basket
58. 12:35or ski a ball down a hill, whatever it is. Um, PA is essentially pre-prompting um, their video model and putting it
59. 12:46behind buttons, right? And so you can just go in and you type in a subject and then you can say, "Turn this into a
60. 12:52video where you peel the subject off like a sticker or you crumple it up like paper and they've added sound effects
61. 12:59in." So they're so they're basically taking a few tools, duct taping them together with a clever prompt and
62. 13:06putting a button on front in front of it. Um, what they're doing with Pika Social is you've got a thing that looks like
63. 13:21Tik Tok pika social pa. All right, I've got 14% so I should be okay.
64. 13:38Okay. So, you've got [Music] >> right, you got your share button, you got your comment, you got your heart, you got
65. 13:49your plus button, you got your profile. Everything's just like Tik Tok, right? >> All right. So,
66. 14:05what's different? Well, it's 100% AI generated and it's 100% photo to lip sync
67. 14:19or like photo to dance. So, you can't even record yourself. You can't record video of yourself and have it modify that. And I think they're
68. 14:29doing that on purpose. I find it really fascinating. So, if I go plus, it's it's saying take a picture of
69. 14:36yourself. So, I'm going to be like, you know, I take a lovely picture of myself. I can't record video.
70. 14:45There's a little animate button at the bottom. There's a little audio thing at the top. So, again, just like Tik Tok.
71. 14:51And then these things on the side. AI speech. So, you can have AI synthesize it or you can record it. change the look and add a performance.
72. 15:05Now, if I add the performance, there's lip sync, fast rap, emotional singing, TED talk, sad, playful, sassy, and then you can
73. 15:17describe your performance. So, what these do if you click on them is they well they in the background they add a
74. 15:26uh a prompt, but you can also just type in a prompt. And then if you go to looks edit look error still in still in preview.
75. 15:44Let's go. Come on. You can do it. Network connection was lost. Oh, this keeps happening on my iPad. I think it's my shitty Wi-Fi.
76. 15:54Um, let me pop out of that app. Pop back in. Go there. Let me quit the app. Come back in. Do plus. Take a picture.
77. 16:09Okay. Edit. Look. Okay. So, there we are. And then I can, you know, I can just go like hair color. Um, pink. Bang.
78. 16:20And it's going to add pink hair color. Tell the boys to get off Call of Duty. So,
79. 16:34so that's that done. And then I can go add sound. And I can do AI speech. And let's see, I can do southern dude.
80. 16:43>> Hello, this is your clone voice. Thanks for using Pika. >> Or I can just record something. Well, you're not just special. Look at
81. 16:58me with my little pink hair. Do you like my pink hair? I like my pink hair. You know what I should do? I should put on
82. 17:04my pink bow with my pink hair. All right. So, that should be awesome. >> Well, ain't that just special? Look at
83. 17:13me with my little pink hair. Do you like my pink hair? I like my >> solid performance. That's like That's award-winning acting. I I have
84. 17:22>> And I'm just special. Sh. calm down. I have I have a I have a fine arts degree in acting. So, you can tell that
85. 17:32there, right? And so, now we we recorded that sound. We've got that picture and we'll just animate it and now that's
86. 17:39going to turn it into a thing. This is the whole app. So, the whole app is you take a picture of yourself,
87. 17:46modify it somehow, add some sound to it, and it does the lip sync. You can also take anyone else's video.
88. 17:54If you like their video, you can use their look, their sound, their motion, or all three. So, you basically just
89. 18:01say, "I want to you I want to make a version of me in their video." And then you take a picture of yourself and it
90. 18:08applies all their crap to you, right? And then off we go. Um, so here's what I think's interesting about this. They're training us.
91. 18:23They're training us to get creative with AI generated video. And at the same time, they're training us to accept that
92. 18:33an all AI video channel isn't necessarily a bad thing, right? One of the tropes in society right now is well, was it AI generated
93. 18:48with the with the pursed lips? Listen here, little mister. Did you use AI to generate this? Cuz if you did, that's
94. 18:55what you're an awful person, right? And they get that shitty look on their face and they're all judgmental
95. 19:02and you're like, "Uh, I was just [ __ ] around and having fun." Well, aren't you just the art thief of the century,
96. 19:09stealing from all those artists, world's greatest plagiarism machine? Aren't you proud of yourself? Right.
97. 19:17>> Well, and I'm just special. Look at me with my little pink hair. Do you like my pink hair? I like my pink hair. You know
98. 19:23what I should do? I should put on my pink bow with my pink hair. >> Well, and I'm just special. >> That's pretty good.
99. 19:36So, I'm going to put the the caption. Where's my pink bow? Where's Where's my pink bow? And then we're going to post that.
100. 19:52And so now, if you're on the Pika social app, you may run in. >> Ain't that just special? Look at me with
101. 19:58my little pink hair. You like my pink hair? I like my pink hair. You know what I should do? I should put on my pink bow with my fake hair.
102. 20:06>> It's >> Well, >> it's actually the the the lip sync animation is actually really good. Um, okay. So, so you've got all these people
103. 20:23who are not using AI clutching their pearls about was it AI generated or not? And listen, I understand the fear of where that comes from is
104. 20:36if someone puts out a video of me that's a deep fake that is disparaging or misleading or whatever it might be. That's dangerous,
105. 20:47right? So, I think the sentiment for where the concern comes from is fair, but it completely discounts all of the
106. 20:57other possible use cases of this kind of technology, like making a stupid silly video or making a social network where
107. 21:06people could just make silly AI videos and share them with one another and have some fun, right? Play first. There's also a lesson here
108. 21:18in we all now have access to APIs and incredibly powerful models where you could put together a social network that
109. 21:28is a combination of Instagram and Tik Tok and AI generated stuff and non-AI generated stuff and just mix it all together.
110. 21:39What they've chosen to do is just say no, we're going to lean into it. We're we're an AI video generating tool. We've
111. 21:47got this cool lip sync feature. What if we just did an app that takes that one feature and just does it really well.
112. 21:54Now, are they going to expand over time? They might, but right now that there's something about the simplicity of that they could easily add
113. 22:07text to video, video to video, image to video, but they're not. They're just saying photo to um to lip sync. In fact, you
114. 22:20can only take a selfie. You can't upload an image unless tell me if I'm wrong on that, Danielle, but I don't think you
115. 22:28can. The way I did it last night was I I made a really cool person in Midjourney and then I basically just flipped my
116. 22:36phone around to my screen and I took a picture of the MidJourney image like it was a selfie. Um, saw this online. Don't
117. 22:44be a prostitute. Yeah, exactly. All I know right now, James Dutder, all the AI customer service sucks. Oops. I mean,
118. 22:52pretty bad. Oh, you can say sucks here. Yep. Plagiarized an episode of Seinfeld, I think. Just kidding.
119. 23:01Hey, Kyle, I was watching The Irregulars on Netflix earlier. Oh, that's that's really funny. Um, man, I can't read. Oh, there we go.
120. 23:14That's better. I looked up and saw the Pika video. That Pika video, the animation in that's really good, isn't it?
121. 23:28[Laughter] Well, ain't just special. Look at me with my little pink hair. You like my pink hair? I like my pink hair. You know
122. 23:41what I should do? I should put on my pink bow with my pink hair. I'm a professional. Um,
123. 23:58oh man, couple of people still trying to find the app. Okay. So, if you go in the app store, let me go in the app store and
124. 24:06find it for you. It's iOS only. So, that means Apple Apple mobile devices only. Where's my app store? Oh, there it is.
125. 24:24P I K A search. Okay. So, if you look here, you see that one that says Pika AI video? That's not the social app.
126. 24:42These two that say Pika Art and whatever, Pika something Labs, those are knockoff apps. All right. The one you're gonna look for
127. 24:56is the one with the the Chinese I I assume Asian woman. And notice how the little pika rabbit is in 3D Chrome and it says Pika Social AI
128. 25:09video. That's the one. Now, you're going to need a code for that. If you're a member of the AI salon, which you should be,
129. 25:20and you go to the irregulars, which is down in clubs and hubs. Yeah, I changed my tab. You can share.
130. 25:35Um, Danielle, three posts down, Danielle did a post that's got a code in it. I don't know if this code is still live, but if it is,
131. 25:46Android, no, no Android. Not right now. Just iOS. So, not only do we get blue bubbles and you get green ones, which is
132. 25:56embarrassing. You don't get Pika Social. You got You got to be on the iOS train, people. It's It's a class thing.
133. 26:09It's it's digital classism. It isn't actually a rabbit. It's a pika. A small aminal that looks like a mix between a mouse
134. 26:21and a rabbit. Fantastic. I did not know that. How do they monetize the app or what do you think their game plan is?
135. 26:31That's a great question, David. I um I mean right now this this is in sort of like alpha testing or like you know
136. 26:41private beta testing or whatever they call it. Um so right now they're just testing it. Um they make money on their um video
137. 26:53generation site by charging subscriptions. Um this social app might just be an experiment. It might be a kind of thing
138. 27:01like they roll commerce into it. So imagine, so they've got all these templates. Like one of the things Peak is really good at
139. 27:11is they figure out some prepackaged things to be able to let you just push a button and it makes magical results.
140. 27:20Well, imagine them being able to do that with social shopping where you could just take a snapshot of yourself and
141. 27:27then you could put yourself in seven different locations holding up a product selling that product. So, just like people sell [ __ ]
142. 27:34on Tik Tok, there could be a whole model around social social selling. Um, could be advertising. I I doubt it's going to
143. 27:43be that, but it could be. Um, they'll likely also take what they're learning from this lip lip syncing in this little
144. 27:51social app, migrate that back to the web, so you can do more sophisticated kind of uh animations for normal video
145. 27:59projects, whatever normal is moving forward. So, I don't know. I don't I don't I don't know a ton about their
146. 28:05business model, but right now they're making money with subscriptions. All of these companies I am I I would be shocked
147. 28:17with maybe the exception of Midjourney. I would be shocked if any of these major frontier model companies or major video
148. 28:28image sound companies if they're making any money. I'm sure they're all bleeding money right now. Right. They're buying
149. 28:35as many GPUs as they can. They're getting as many users as they can. Right. right now it's the it's the
150. 28:42audience grab. They're trying to get as many audiences hooked on their platforms as possible. Right. One of the one of the smartest
151. 28:50things that OpenAI's done in the past two or three months is they turned on memory. Right? So Chat GPT now knows all
152. 28:59of your chats for the past year or so. So you can say, "Hey, based on what we've talked about in the past year,
153. 29:04what are some ideas that that I've brought up that I never pursued?" and it'll give you a list of them. Um, once
154. 29:12that's turned on, you're not going to want to stray very far from that. You want to keep adding to that memory,
155. 29:17right? So, so they're all trying to figure out what's their sticky um advantage that people won't leave. And I
156. 29:24think that social app is an interesting one where if people really get into making these crazy AI videos and it's
157. 29:31just a world of crazy AI videos, then there's none of that weirdass [ __ ] Was this AI generated? Yeah. All of it's
158. 29:39AI generated. And who gives a [ __ ] Who gives a [ __ ] right? The only reason you should give a [ __ ] if something is AI generated is if
159. 29:52someone is trying to fraudulently deceive you. And in that case, there are actual laws on the books that prevent that. It's independent of
160. 30:03technology. People can deceive you using non AI, right? I mean, we've The last 20 years has been nothing but bizarre [ __ ] strange ass
161. 30:15misinformation. Oh man. Keep them. Lick them in into their ecosystem. Exactly. It's the corner drug dealer phase. Yeah. Give out
162. 30:25some freebies. This is totally it. You want a little bump? You want a little bump? Some pretty good stuff this week. Yeah, exactly.
163. 30:36Lock them in. That's it. Source camp. I'm loving OpenAI's memory. No more searching individual chats. I know.
164. 30:44Thank God. That was a mess. They will eventually have cloud storage memory and sell it at a premium. I agree with that.
165. 30:51Actually, I'll tell you probably a really good business is, and I'm trying to think how you would do this. I guess you'd have to
166. 31:00have all sorts of your own data storage and APIs and I I don't know, you probably do it with web scraping, but
167. 31:10imagine a service that you could subscribe to that actually provides memory across services. Like wouldn't it be amazing that I could
168. 31:21not only search through my chat GPT history but Claude and Perplexity but also things like Midjourney and you know
169. 31:29I could be like hey which video service did I did I do that video with the Civil War guys? Oh that was in Lumalabs.
170. 31:36Here's the link to it. Right. That that would be huge because one of the things we talk about here in the AI learning lab is play right and
171. 31:47just go play with as many tools as possible. play. Just figure out what the stuff does. Figure out what the capabilities are.
172. 31:57If you're anything like me, you've got crap everywhere. It's just like social media. Like, how many social media
173. 32:04accounts did you start over the past 20 years that are just orphaned out there swinging in the wind, right? Same with
174. 32:11all these AI tools. But with something like universal memory, maybe it's a blockchain enabled memory where you can
175. 32:17sub, you know, you can um opt in for your different services. That that would be pretty slick. Maybe an agent could do
176. 32:25that memory across platforms thing. Yeah, exactly. And that's why, you know, a lot of people right now are like,
177. 32:32well, you know, it's just going to take all of our jobs and what do we do? Well, you now have the equivalent of like PhD
178. 32:41level employees that work for you, and those employees are going to get smarter and smarter and more and more autonomous
179. 32:48where you can just turn them loose to go do stuff. So, as these tools get better and smarter and can do all this stuff,
180. 32:56you're going to have ideas like, "Huh, wouldn't it be cool if blank go start that business?" I think entrepreneurship
181. 33:07is is going to skyrocket. It's going to skyrocket because a lot of people are going to get laid off. They're going to be like,
182. 33:17"Holy [ __ ] what do I do? Let me go apply for jobs. Oh, there are no more of my jobs. Well, now what do I do?" Well,
183. 33:25maybe I should look at this thing that caused my my unemployment, my undermployment. and learn about it. Oh, I didn't know it
184. 33:38could do that. Oh, I didn't know it could do that. Hey, wait. If I take my skill of the thing I used to do and my
185. 33:45dream of the thing I've always wanted to do and I use these AI tools as an amplifier of my skills, holy [ __ ] I
186. 33:52could start a new business that I would love. We're going to see a ton of that. And those companies are likely going to hire
187. 34:06trusted people that can do things that the founder isn't good at. And sure, the founder can use AI to do a bunch of that
188. 34:14stuff, but that can get kind of lonely. So what we're likely going to see is two, three, four, 10, you know,
189. 34:22somewhere between two and 10 person companies that are operating like they're 150, 200, 500 person companies. It's going to be insane.
190. 34:38Now, the fact that you're here watching this means you're ahead of people. But there's a lot of people right now just
191. 34:44sitting on the sidelines like, "I don't like it. I don't like it. The robots are going to kill us. I don't like it one
192. 34:51bit." Taking the jobs is stealing from the people. The people being pissed off right now at AI. I completely understand it.
193. 35:08Um, even the gray and not so gray areas of how these models were trained that was like unethical and they shouldn't have
194. 35:21done it, but they did it. Like you can be pissed off that they did it, but they did it. And because they did it, these models now exist. And
195. 35:34because these models now exist and they're this good, they're not going away. There is there is like just
196. 35:42ridiculous amounts of money being pumped into this sector. Ridiculous amounts. 24 year old engineers getting $250 million signing bonuses.
197. 35:59Uh, I do not see a pinned comment right now. All the faces tonight are absolutely classic. I just dropped a code in irregulars for
198. 36:11the chat social app. Wait, in Ire's chat for the Pika social app. Okay, so look, for those of you who went and got the
199. 36:21Pika social app, but you don't have a code, Silverf Fox is your best friend. She put it in the chat. So go into the
200. 36:30Irregulars channel, Irregulars AI Learning Lab. And then let's see, Lori Blair, should you get it, use her TTWV.
201. 36:46All right. Use my code. I think she's got three. If she's like a normal human being like me, she got three. All right.
202. 36:57We create whatever the hell comes to mind and eventually something shakes out. Yep. It can have my job if it will also pay
203. 37:05my bills so I can do the thing that I want to do as a kid. Have fun. I think that starts to be where we go. Um the
204. 37:13transition period's going to suck. Like to be clear, people are going to lose their jobs. They're going to be more pissed off than ever.
205. 37:22Um, and over the next five years, we'll start to figure it out. People will adapt. People adapt. We adapted to the industrial revolution.
206. 37:31We adapted to the steam engine. We adapted to the the tractor on farms that displaced 80% of farm workers. Now, that took four decades.
207. 37:43And what we're going to see is probably going to be half a decade, maybe a full decade where we're going to see similar kinds
208. 37:52of displacement, but we're also being given this this knowledge tools been democratized and given given to us all. So, um
209. 38:05I'm exactly as normal as you, Kyle. Exactly. Oh man. Okay. So, let me let me share something here. Um, this image
210. 38:30this is as of today I think. Um, this is tracking aai.org. So, so basically what this organization does is
211. 38:45they take every new large language model that comes out and they have it take the the Mensah Norway quiz, the IQ test,
212. 38:55and they do it multiple times per model. And you can see most models got to the sort of center of the hump here, which
213. 39:03is basically the human average IQ of a hundred, right? And then you've got distributions that go down and you've got
214. 39:10distributions that go up. This thing all the way out to the right here is Chat GPT5 Pro and it hit an IQ of 148
215. 39:24today. 148. Smarter than 99.9% of people on the planet. So, all of the people that said large language models can't get us here,
216. 39:38um, were wrong. Uh, it got us here. Now, GPT5 Pro, it's 200 bucks a month and, you know, it uses massive computer resources for every
217. 39:49prompt. One of the things Altman said is he thinks he's going to give a handful of GPT Pro um, instances to $20 a month people. like
218. 40:01maybe we get five a month or something like that. But if you've got a really tough problem that you want to crack,
219. 40:08throw it into GPT5 Pro and it can, you know, it can likely be 148 IQ. They they said Einstein Einstein
220. 40:19was in the 160 to 180 estimated range. So, we're not quite at Einstein yet, but we're at the 99.9% of of all people's intelligence, right?
221. 40:33Um, so that's that's kind of crazy. So, that happened today. And then I want to show you another thing which is just
222. 40:40bonkers. Um, I got to find it. [Music] Here it is. Okay. You can hear that, right? Right. Brandon,
223. 40:57>> do you need to join the irregulars? No. If you if you go in and join David, if you go in and join the AI salon, if so,
224. 41:04if you go to community.thesalon.ai. So, that URL with community at the beginning of it, community salon.ai.
225. 41:17That'll take you to the community site. um just sign up for it there. If you scroll down the lefth hand side, you
226. 41:24should see the irregulars under clubs and hubs and then if you just click on that, you go in there, that's where you
227. 41:31can see those posts that have the codes. Okay, beautiful. So, what I'm what I'm showing you right here is a little minutong movie
228. 41:41about the China's in in Beijing today, they kicked off the humanoid robot games. Yes. Yes, they did.
229. 42:09[Music] soccer. Oh no.
230. 42:28[Laughter] boxing. [Laughter]
231. 42:50[Music] Now, strange as that is, >> I don't think it's going to be strange for long, um,
232. 43:08I don't think it's going to be strange for long at all. the the the rapid progression that we've seen in
233. 43:19large language models over the past two and a half and and three years. We're going to start to see in um
234. 43:28in humanoid robots and they're going to get cheaper and they're going to you know they're going to be marketed to
235. 43:32consumers. China's going all in on this. So, so part of what this event is is to get people used to seeing them,
236. 43:43used to seeing their capabilities. This to me feels like one of those videos that we will look at five years from now
237. 43:50and they'll have these robots that are running faster than any human, right? That they're this will become a
238. 43:57technology capability thing. And I could see there being different countries competing like showing off their
239. 44:03technology, right? I can absolutely see this being a thing that gets really big. Um, if if where your mind is right now is it
240. 44:16will never be normal to see a robot walking down the street. Well, if you've played with chat GPT three years ago, you would probably
241. 44:25argue it would never you would never find it normal that you can just say to a computer, "Write me you know, an epic poem in amic
242. 44:37pentameter about quantum physics and it'll just write it for you. Or it'll just you just tell it lovable to write you a program
243. 44:47of a of a video game that your kid came up with the idea on a beach and you get back home and you're like, "Hey, make
244. 44:53this game for me." And 10 minutes later your kid is playing a video game that the computer programmed. We have
245. 45:00normalized so much remarkable capability in the past two and a half years just on this channel alone. Like
246. 45:13I don't know about you, like I do this every [ __ ] night. Every night. And sometimes when I do the I I just do the the basic parlor trick.
247. 45:23Write me a poem about quantum physics and it just goes It's remarkable, right? Same thing's going to happen with robots. The first
248. 45:35one you're going to see, you're going to be like, "Oh my god." Everyone will be taking a picture of it. And then two
249. 45:40months later, there'll be 10. And then a year later, we'll be shopping with those robots. And there'll be funny videos of a robot
250. 45:51losing its mind and tripping and knocking down a bunch of cans in the grocery store. Yeah, self-driving cars in San Francisco was
251. 46:00weird until it wasn't. Now Teslas are driving around Austin. It's weird until it wasn't. So, it's just crazy. Just crazy what's coming.
252. 46:15Um, it sounds like we've got some new folks here, which I'm really excited about. If you're new here, first of all,
253. 46:20welcome. My name is Kyle Shannon. Um, let's see. Ryan kind of makes sense. Humanoid is more familiar as opposed to
254. 46:27the robot and Interstellar or HAL. I think we'll see specialty robots. The thing about humanoid robots is
255. 46:35our entire world is built for us, right? So while you could make specialized robots that are really, you know,
256. 46:44efficient at a particular thing, just like large language models being generalized intelligence, generalized
257. 46:52knowledge, um having a humanoid robot can be generalized, you know, physical physical work. So we've got knowledge
258. 47:00work with the LLMs and with with all the other models out there. And then we've got this for physical work. Um, okay. Let's see. Um,
259. 47:13Whimo Whimo driverless cars are all something. That's what God says. Yeah, I don't think this is a God thing. I I
260. 47:22really don't. The whole thing about, you know, we're acting like God and if this is sentient beings and things like that.
261. 47:29I don't think it's that. I I understand the fear, right? And I understand it's very easy to demonize the technology companies and go, "Oh,
262. 47:40they're playing God." No, what they're doing is they've discovered a technology that can provide us remarkable access to knowledge and
263. 47:55do that in this generalized way. And the way I think about it is not like us in competition with that intelligence, but more like
264. 48:07us augmenting ourselves with that intelligence. So like my experience here and and the experience of a lot of the people in the
265. 48:15AI salon is not that AI is this awesome thing that like a vending machine like you put in your prompt and out comes
266. 48:24genius results. That's not how this [ __ ] works. you put in your prompt and [ __ ] comes out of that thing most of the time.
267. 48:33The way you do it is you say, "Okay, if it's not the genius, then it's still about my intention and my meaning." Um,
268. 48:45okay. But I worry about the people that control the prompts. We control the prompts. You mean you worry about the frontier model companies
269. 48:55that are making the the uh the models the Oh, the one above it. I meant in the context that it was something to work for us.
270. 49:09Well, yeah. I mean, I think I think ultimately the humanoid robots will work for us. Um, it is it is in it is in all of the
271. 49:22robot building companies interest to sell as many as possible. And so if they if they just use the robots to repress
272. 49:32us, then they don't sell as much [ __ ] So So they want us to want robots. They want us to to use these things. Tik Tok.
273. 49:42Um, I want a robot that can do my dishes and remind me why I just walked in the other room. Exactly. Exactly. And and
274. 49:50like that's the kind of thing that I I think we'll get. Again, I think that's the kind of thing I think we'll normalize. Um
275. 49:57I don't disagree with you that you are you are very right to be skeptical that open AAI or X AI or Google or Meta
276. 50:12or any of them are going to inherently do the right thing. Whatever the right thing is, we should be skeptical that they're not
277. 50:21going to do that. But there's something happening at scale that I don't think I've ever seen before. We saw it a little bit with the
278. 50:32internet, but every time a frontier model company puts out a model, we get an open-source model either from China or from the US that is
279. 50:44nearly as good or as good and sometimes better than that model from from the Frontier Company. um fully open sourced.
280. 50:54If you've got the fully open source models with the weights, you can train them to be however you want. So if you
281. 51:03think that the the the safety measures that OpenAI is putting into their models, you don't agree with with their
282. 51:10politics or you don't agree with their definition of safety and you don't like anthropics. Well, go to one of the open
283. 51:18source models and train it yourself. do it yourself and put it in the marketplace. I mean, th this is one of the things that drives me a bit
284. 51:34crazy right now is people that are on the sidelines with AI are talking about AI like it's being done to us.
285. 51:45It's only going to be done to you if you don't use it. Right? It's not like they're making all these discoveries and and all this incredible
286. 51:59progress and they're keeping it behind the firewall and only giving it to their highest paying customers, right? That's
287. 52:07the way AI's been for decades. What they're doing is they're putting it in our hands. So, we get to play with this
288. 52:13stuff. The 98.5% of people that are not engineers have access to the best models in the world. And way more than 50% of people, at least in
289. 52:30America, are on the sidelines going, I don't like it. I don't want to deal with it. It's one of the most profoundly powerful
290. 52:40technologies in the history of humanity. It's not going away. And you have two choices. If it and listen, I could be wrong. Maybe it is
291. 52:49going away, but I I just I can't I can't see a path where it goes away. So, if it's not going away, you have two choices.
292. 52:58There's two. It's binary. You deal with it or you don't. And if you're sitting on the sidelines pissed off about it and you're just
293. 53:09watching this bullet train [ __ ] take off, AI is going to happen to you. If you're like, I don't like it. I don't
294. 53:20like it one bit, but you know what? [ __ ] it. I'll show up at the AI learning lab. I'll listen to this [ __ ] lunatic rant
295. 53:26about how it's inevitable, and maybe I'll try it. Maybe I'll make a a kids book for my kid. Oh my god,
296. 53:33that's so cool. Can it make pictures? Oh my god, it made pictures. It'd be nice if that were a song. Wait, you can make
297. 53:40that into a song, too? Holy [ __ ] Right. What what what we regularly experience in here
298. 53:56is not this adversarial relationship with AI, but this relationship where it's like I still get to have my agency
299. 54:04as a human being to say, "Here's what I want in the world. Here's the audience I want to do this for. Here's how I want
300. 54:11to impact them. Here's an idea that I have. Here's an idea that I had when I was seven. that when stepdaddy told me I wasn't
301. 54:20talented, I sort of shut that idea down. And you know what? I just remembered it. And let let me go to chat GPT. Hey, I
302. 54:28had this idea when I was seven. Oh my god, that's a great idea. Here, let me help you articulate that and expand it. Holy [ __ ]
303. 54:36And all of a sudden, people are are in here reinventing themselves. The amount of people who have come out of
304. 54:43retirement in the AI salon and in the AI learning lab is remarkable. They were like, "Yeah, I'm done." And
305. 54:52they're like, "Holy [ __ ] this is the most exciting. I feel like I'm seven again. [ __ ] amazing." All right, Tik Tok
306. 55:00pin. Oops. And black bar. I can't believe you didn't yell at me for the black bar. Will Smith is sitting over there saying,
307. 55:09"Hold my beer." Exactly. Exactly. Um, at a minimum, you owe it to yourself to just play with these things. Don't take
308. 55:24them seriously. Here's the thing. If you try AI and you hate it, cool. But, I mean, really try it. Like, really try it on for size.
309. 55:40Like go in and take something that you're really passionate about, something you really know. Like maybe
310. 55:44you're into gardening and go in and see what it knows about gardening. And then take something that you're
311. 55:51really shitty at. Like maybe you're really shitty at math. Like I'm really shitty at math. Like really shitty at math.
312. 56:01But I can go in now and it can do math. And you might have heard from people, oh AI is not good at math. No, AI wasn't
313. 56:09good at math and now it is. It wasn't good at math and then they figured out ways to make it good at math.
314. 56:19Like it can write its own Python code, which is code that you use to do math, among other things, and it can write and execute its own
315. 56:30Python code and do complicated math. Now, is it always perfect? No. But I'll tell you something that'll blow your
316. 56:37mind. If you take something where you think you you know in your heart you're horrible at that thing. I could never be
317. 56:44a coder because my mind just doesn't work like that. I don't have the attention span for it. And you go to lovable.dev.
318. 56:53Let's go there right now.
319. 57:10archetypal two YouTube comments wasn't good at math until two weeks ago when it won the gold medal in the math
320. 57:16competition. Exactly. There's basically a 0% chance that everything goes fully right or or wrong direction. It boils
321. 57:24down to who is the most effective and cooperation always wins wins game theory in the long term. The here's here's a really cool
322. 57:38dynamic. Oh, do I have to change my tab? Yes. Here's a really cool dynamic. Um the frontier model companies are
323. 57:52populated with engineers and mathematicians, right? Heavy leftbrain, logical, intelligent process systems, math, coding,
324. 58:07right? Almost none of them have any creative people in them. none. They are building tools that are called large language models.
325. 58:24What I can promise you right now is that the engineers that are building these tools have no idea how the 98% of us
326. 58:33that are not engineers and heavy geeks are going to use these tools. We have a disproportionately impactful
327. 58:44um uh uh uh influence on how these tools are going to be used than we may realize. No, we didn't build them and
328. 58:53no, we're not writing the prompts that control them and make them safe and all that sort of stuff, but we're using them
329. 58:59and we're building things and we're putting things in the world. And if rather than just sitting on the
330. 59:07sidelines pissed off, we go, "Oh, you know what would make it better? If we made some things out of these AI tools
331. 59:14that really thought about humans first and technology second, because the the frontier model companies certainly
332. 59:20aren't doing that. We can do that." Well, how do you do that? I don't know. You just use things to make [ __ ] that is
333. 59:28more human centric. take whatever you're passionate about and use the tools to do that. Right? So like this is not we are
334. 59:38not passive bystanders at all. Right? At all. In fact, Sam Alman talks all the time about they put things in the world early
335. 59:54so that they can see how people use them to go, "Oh, we didn't know people would use it that way. Let's figure out how to support that.
336. 1:00:02All right. Um, relevant spotlight comment. Open AAI figured out immediately that iterating hundreds of
337. 1:00:10millions of times beats the pants off a few hundred engineers trying to guess. Yeah, exactly. The reinforcement
338. 1:00:16learning thing. I was grilling an econ professor of mine about um, black skulls theory pretty hard. I don't
339. 1:00:27know black skulls theory, but we could go there and do it. Anyway, let me show you. We We can go to to uh to chat GPT.
340. 1:00:34I just don't know what that is. I've heard of it before, but I don't know. I don't know it off the top of my head. Um
341. 1:00:40and at one point he just made an analogy. I'm just a carpenter. That's interesting. If the tool works, it works. Yeah.
342. 1:00:49And this is so um David Shapiro talks a lot about the industrial revolution and the steam engine augmented our physical
343. 1:00:59strength and generative AI is augmenting our intellect, right? Our our our uh intellectual strength. Um both, you
344. 1:01:11know, massively transformative. Um, we didn't think we would get displaced like this, you know, just like I'm sure
345. 1:01:21the farmers in in the late 1800s didn't think they were going to get displaced. Um, Dana, they they said the same thing
346. 1:01:29when computers came into homes and now everyone has one in their pocket. Yeah. Like we've got crazy ass computers in
347. 1:01:37our pocket now. Okay. If you've not seen Lovable before, and this is this is the the thing of, you know, if you've told
348. 1:01:45yourself that you can never do programming, um I'll just do my favorite one. Um, I want you to build me an exact
349. 1:02:01clone of the original Space Invaders video game with sound effects and scoring
350. 1:02:17and play ability. Um, I want you to make the color version, not just the black and white. All right.
351. 1:02:33So, I just typed in what I would consider a really crappy prompt. And now, if I click the code mode, you can actually watch this thing. So,
352. 1:02:42it's set up this whole environment for all of the different files. And for every one of these files, it's actually writing code for them.
353. 1:02:52Um, it's not copying and pasting this code. It's actually writing the code. So, let's Here's index.html that's already written.
354. 1:03:02Um, so it's spinning up the preview. So, what did it say here? It said, I'll build you a color colorful Space
355. 1:03:09Invaders clone with classic gameplay elements. the features in this version. Player ship with left right movement and
356. 1:03:16shooting grid of colored alien invaders, collision detection with explosions like like it knows what I meant when I said
357. 1:03:24make me a game. It knows what it needs to build. It knows how to structure an application. And in a few minutes here, we're going
358. 1:03:34to be able to play this game. All right, Tik Tok pin. For better or worse, I finally have a tool that can move as
359. 1:03:40fast as my brain. Yeah, that's the other thing. The other thing that is so inspiring about this AI [ __ ]
360. 1:03:46when you when you get into it, when when you when you shift out of this mode that AI is this thing off to the side that's
361. 1:03:52going to replace me, and you shift it in front of you and say, "No, no, I'm in control. I'm the conductor of this orchestra."
362. 1:04:02And the more AI tools that I can learn, the more instruments I have in my orchestra. And what I can do is I can
363. 1:04:10just sit there as the conductor and I can go, "Oh, I have an idea. Let me start out in chat GPT and organize my
364. 1:04:17thoughts and let me go over to this thing and make some pretty pictures. Let me go over to that thing and make some movies." And within
365. 1:04:2615 or 20 minutes, you can take what was just a vapor of an idea in your head and be looking at a TV commercial about it
366. 1:04:34or playing a video game or have a a 10-page investor deck, right? Maybe you were at some dinner party and you spouted off your mouth
367. 1:04:47some big idea like, "I have this idea for a thing." And there was an investor there and the investor said, "Well, do
368. 1:04:53you have an investor deck?" And you're like, "Uh, yeah, okay. Let me send it to me tomorrow." Uh, okay. In the olden timey world, you'd be
369. 1:05:07screwed. With AI, you can go, "Hey, here's this idea. I need an investor deck. Please go do some research. Find all the
370. 1:05:16competitors in this space. Find the total addressable market. do all the investor best practices and make me a 10 slide deck that that you
371. 1:05:26know follows Silicon Valley best practices and 15 minutes later you'll have something. Um, we just asked for a Space Invaders
372. 1:05:36game. Here it is. Let's see. All right. It doesn't have good logic. Uh oh.
373. 1:05:51What? We're screwed. We're screwed. Couldn't have that collision detection. But anyway, you can't code now. You can.
374. 1:06:05And you know how you fix all the crap that's broken and the fact that the shapes look like [ __ ] You tell it.
375. 1:06:12You just go, I don't like those shapes. And the bullet shouldn't go through all the things. It should do a row at a time
376. 1:06:17and they come down too fast. And the stars in the background are too fast. Go fix that. And it goes and fixes it. Wild times.
377. 1:06:29Wild times. Stacy Rodriguez, when they say you can build software, everyone is building games and calculator apps and
378. 1:06:36card games, which is cool. But can we build software that is actually useful, like a CRM or accounting app? So Stacy,
379. 1:06:42let me show you something that I made. So, if you go to project-cardboard.lovable.app. This is a
380. 1:06:57adh friendly uh project manager, like a to-do list kind of thing that that I um that I built in three hours because I was
381. 1:07:09trying to get all my projects out of my head and on a whiteboard. And I figured out a cool way to organize them on a
382. 1:07:15whiteboard. And I don't know why it's not loading. It might be It might need to spin up the server. I don't I haven't
383. 1:07:22run this in a while. Um within three hours I had a functional project management app. Yeah, you can make anything you want. Like if the
384. 1:07:35reason games are easy is you can you can know if a game works or not, right? You you know what they look like. You know what they play like.
385. 1:07:43Um, yeah, this is not working. Strange. Let me go back to lovable
386. 1:07:57and log into my dashboard. Maybe I need to up up update something. [Music]
387. 1:08:14[Music] All right. What's going on here? [Music] Oh, it did a security assessment.
388. 1:08:30[Music] But yes, you can. Where's the dog? The dog is over on the bed. Um,
389. 1:08:44yes, sir. >> There you go. Uh, first of all, uh, we're about to lose our Tik Tok bin, so hang on one second. >> All right.
390. 1:08:52>> Uh, now you got the song stuck in my head. So, thank you. Um, your Superbase uh, project probably spun down. uh it
391. 1:09:02caused it because of inactivity since it's been a while since >> inactive. Yeah. >> So that's one thing to call out. If
392. 1:09:09you're doing anything on any of these Vive coding sites that's using level that's using Superbase Superbase if you
393. 1:09:15don't actively use the project, it will shut down your server unless you're paying for it. >> Got it. >> Kelly's pins back up by the way.
394. 1:09:24>> Okay. I came up with a new biz idea and one hour later I had a full power powerpoint presentation to I say show
395. 1:09:33for it. Oh to investors. Yeah, there you go. Book them, Kyle. Yeah. So, so what what happened there? Oh, wait. It's
396. 1:09:41back. It looks like it's back. Maybe it spun it back up. Um, so I'll do kylestorymind.com. Um,
397. 1:09:58it failed to fetch. Yeah, it's it's not working. So, uh, my, uh, thank you for that, Brandon. Yeah, my my, uh, the
398. 1:10:04database here because I haven't used it in a while. It spun that thing down. But, yeah, you can you can do any app
399. 1:10:10that you want. You want to do a CRM? Uh, someone someone did a CRM the other day. Actually, let's let's go let's go do
400. 1:10:17one. We'll do one. I I thought this was a really clever thing. Um, I want to make a CRM where um the interface
401. 1:10:33looks like Netflix where each project is a movie poster and um the people are cast members
402. 1:10:49something like that make it um easy to use, easy to manage and look absolutely
403. 1:11:07fabulous. And and if you wonder, can you say things like that? Yes, fabulous. Yes, you can. You can just
404. 1:11:24you can talk like an idiot to these things. You can have them talk like an idiot to you. You can talk very sternly and very direct to it.
405. 1:11:33It understands language. It understands intent. It understands nuance. It understands jokes. If you're joking,
406. 1:11:43ah, that's funny that you called me a dum dum. It's it's remarkable. >> Does it understand how to read sticky notes?
407. 1:11:52>> No. Tik Tok question. Robbie, when will AI be able to create completely new types of AI for you? Um, when will it
408. 1:12:01engineer? I think we're we're we're at the beginning of that right now, Robbie. So, okay. So, here are some here
409. 1:12:10are some hints that that's happening. For the first two years post chat GPT, all of the frontier model companies were
410. 1:12:19talking about wanting to achieve AGI. And AGI is artificial general intelligence. And that's the point at
411. 1:12:25which, you know, an AI system can do most of the tasks of most economically viable jobs. Um about four months ago, five months
412. 1:12:38ago, they all kind of like they coordinated, but it it it was within like a three-week period, they all
413. 1:12:46stopped talking about AGI and started talking about ASI, artificial super intelligence. That's where the the these
414. 1:12:54things are basically smarter than all humans at all things. Chat GPT5 Pro, the the high-end model today just hit 148 on the Mensa IQ. It
415. 1:13:07took, you know, they did multiple passes at it. It hit a 148. Um, part of the acceleration that's happening is they're using reinforcement
416. 1:13:18learning to to um have these things recursively self-improve. They're starting to do that process. So right
417. 1:13:26now in the labs, these things are starting to teach themselves how to get better and better and better. I built a layoffs.fyi
418. 1:13:34type tracker for higher ed actions in USA with some knowledge of that's pretty cool. with some knowledge of coding. Yeah, it
419. 1:13:45like having just even a little bit of knowledge of coding just just understanding how code works and how it's structured and it requires
420. 1:13:52different files and that there's a server and a database and a front end and things like that and you know
421. 1:13:57security and authentication. Um if you know the basic components of software you can vibe code. Now, I think there's going to be a whole
422. 1:14:07industry. If if you're in here and you're actually a coder, if you're really good at coding, I'll tell you an
423. 1:14:13I'll tell you an industry that I think is going to be huge. And it's going to be um vibe coding cleaning crews.
424. 1:14:25vibe coding cleaning crews who come in and some, you know, showoff CEO decided to vibe code a CRM just like I did and
425. 1:14:36they're going to launch it and it's going to they're going to lose all their data and they're going to want someone
426. 1:14:41who knows coding to come in and fix it. I think that's going to be huge, huge, huge, huge. Mobile first app an AI
427. 1:14:49analytics suite. So, if we click on one of these projects, people Alex Rodriguez, that's pretty funny. But
428. 1:14:58anyway, here's here's a a a Netflix-like CRM thing. Four or four projects. I don't know how to make a new one.
429. 1:15:07All right. So, so right now this doesn't work. So, I'm going to say um make this functional um without uh
430. 1:15:21without a database. Um, I'm fine with it losing losing info, but I want to be able
431. 1:15:38to create new accounts and new people. All right. And now it'll go off and do that.
432. 1:15:54And I I you can there's just all sorts of stuff you can do. You can change the look and feel. You can, you know, one of
433. 1:16:00the things about vibe coding is a lot of the apps all look identical. Well, that's just because the people making
434. 1:16:06them were too lazy to say, "Hey, I want it not to look like everyone else's app. Make it look more like this or that."
435. 1:16:12Like this already looks different than most vibe coding apps. They all have that sort of white interface with the
436. 1:16:18purple rounded corner blocks, things like that. Um, all right. Echarium. Let's see. The real. AOL. Great name. Oh, AOI. When it comes to
437. 1:16:34issues of security, how can non-coders implement this? I'm sure you hear. Yeah. Um, okay. So, just like
438. 1:16:47there was a famous um about two years ago, there was a famous instant where a lawyer Tik Tok question. Yeah, I got it.
439. 1:16:55That's what I'm on. Uh a lawyer wrote a legal brief and handed it to a judge and the judge looked at the legal brief and
440. 1:17:04he looked at the citations and he said, "None of these cases exist." and the lawyer used chat GPT to write
441. 1:17:13the legal brief and didn't check the citations, right? The the prior law. They were fake cases. That's not on Chat
442. 1:17:23GPT. That's on the lawyer, right? Same thing with vibe coding. If you're if you're going to vibe code a CRM just for
443. 1:17:32a little fun experiment, you don't need to worry about security. Like even put in authentication, things like that. But
444. 1:17:39if you're going to deploy this app into your company or into the world, um you damn well better understand the the
445. 1:17:48implications, right? So there are absolutely going to be scores and scores and scores of people who put apps into
446. 1:17:57the world. There there's I heard two two or three stories today. They put apps into the world. They don't secure the
447. 1:18:04data. they don't, you know, whatever you, you know, put it on on uh specialized servers that, you know,
448. 1:18:12aren't easily accessible or or they do backups of raw data, what, you know, what whatever the the security obvious mistakes are. Um, and
449. 1:18:22sophisticated hackers now have AI tools. They're going to be able to just take a Manis agent and say, "Go find me anything at lovable.app.
450. 1:18:31Go find me any app at lovable.app app that has the following 10 security vulnerabilities and go exploit them for
451. 1:18:38me and it'll just go do that. So if you don't know what you're doing, then get someone who does, right? Or
452. 1:18:49have Chat GPT teach you what you need to know. The other thing that's going to happen is the the um the application
453. 1:18:58developers like Lovable, like they just raised $200 million on a what was it a $4 billion valuation. This is an
454. 1:19:07eight-month-old company. They just raised $200 million. Um these things are going to get better and better and better at plugging those
455. 1:19:17security holes for you. But that's not the stage of the industry we're at right now. The reason there's so much
456. 1:19:22opportunity right now is all of the AI tools suck. They all suck. They're all janky. They're all broken. They they
457. 1:19:30hallucinate. They do crappy work. It's on you. Now, if you want to wait until these tools are bulletproof, that'll be five years from now.
458. 1:19:41and all of the opportunity, all of your friends that decided to get into it now and figure it out and duct
459. 1:19:48tape stuff together, they they will have successful companies and you'll be like, "Damn, I should have done that." The
460. 1:19:55opportunities right now, but also the risk is right now. So, you gota you gota just hop on your horse and get rolling.
461. 1:20:03Get rolling. Yeah. Get disbarred or Oh, got disbarred or reprimanded. He certainly got reprimanded. I don't think he got disbarred.
462. 1:20:11Um, Rick Olsson, when it comes to issues of security, how can coders implement this? That's what I just talked about.
463. 1:20:22It's already happening. I just had to clean up a beast of a vibe code of that. There you go. Yeah, I was talking. I I I
464. 1:20:28I'm serious. Wait, Archinum. Archenum. Cool name. Um, if you've got the skills to be able to go in and clean
465. 1:20:39up a vibecoded app, [ __ ] hang a shingle right now. And like I like you could go to Lovable and Vibe Code your
466. 1:20:46website for I'll fix your vibecoded app. I'm telling you, man. All right, these things still don't work. Completed.
467. 1:20:58How do I make a new one? Yeah, dumb dumb. All right, it didn't do it. Cinema RM here's stats. Um, let's let's do this. I'm gonna go to
468. 1:21:13chat GPT. Watch this. We'll do something fun for you our keenum. Um Um, here we go. So, I'm going to go um I'll I'll do
469. 1:21:22it with with talking. I think I can dictate to it. All right. Can you hear me? Yes, you can hear me. Okay. So, I want to um create a website
470. 1:21:34uh that is marketed toward vibe coders who have built an application using vibe coding that they're super proud of, but
471. 1:21:44all of a sudden realize that it's got big security holes in it or big parts of it don't work right and they don't know
472. 1:21:51what to do. Period. I'm an expert coder and I can go fix vibe coding apps and I want a website that's super compelling
473. 1:22:01um super clear and awesome. All right, let's see if it heard what I Yes. All right, can you hear me? Yes. Okay, here we go.
474. 1:22:16Think longer. Okay, so so we're we're in auto mode of chat GPT5. And I'm trying to think where I want to go do this. Do I want to do this at
475. 1:22:28lovable or do I want to do this somewhere else? I could do this right within chat GPT. Why don't we do it within chat GPT?
476. 1:22:37Oh, it's actually just making me the website. Oh, that's interesting. All right. Well, [ __ ] it. Vibe Vibe App Rescue onepage website.
477. 1:22:48[Laughter] All right. It's It's writing us our website now. And we'll be able to look at this here in a second. Um,
478. 1:22:59do you use Claude code at all? I haven't I haven't used claude code. Um, any of the things I'm not a real coder, so any
479. 1:23:07of the things like cursor or or wind surf um I've used replet agent a little bit. Um, but but not a ton. Um, I'm more
480. 1:23:16of a front-end guy and just an idea guy. So once I get into needing to track all of the different components myself, I'm
481. 1:23:27like, "Fuck it. I'm tapping out." Um, but um, was it Manis? Yeah, Manis today. Today added Claude code into Manis and they're
482. 1:23:41doing all of that stuff for you. So, you can go into Manis right now into their developer mode and just tell it you want
483. 1:23:47it to write an app and it will do all of the database setup and all all of the stuff that you normally have to
484. 1:23:54configure within um within cloud code and within cla uh within cursor. Um it just it's doing that for you. So, I
485. 1:24:03haven't played with that yet, but again, these these things are going to get, you know, better and better and better. All
486. 1:24:08right, let's let's look at our website. Here we go. Your vibecoded. Your vibecoded app ships fast. I make it
487. 1:24:19safe, solid, and fast. Well, that's bad copy right there. There's case studies. Sound familiar? So, that looks like
488. 1:24:28[ __ ] So, here's what we're going to do. We're going to close this. We're going to say, "Chat GPT, that's [ __ ] boring
489. 1:24:33and awful." I'm going to say, "Why don't you write me a creative marketing brief
490. 1:24:48I can give to an app developer?" Okay. YouTube question. Out of curiosity, how does it know when to cut you off
491. 1:25:02if it's based on compute power it takes to answer the question? I don't I don't understand what you mean when it knows
492. 1:25:09how to cut you off. Um, what's happening in chat GPT5 right now is it's got these three different modes. Well, four if you
493. 1:25:19if you're a pro member. It's got fast like thinking light. So reasoning light, normal reasoning and then pro-reasoning.
494. 1:25:29All of these reasoning models are um you give it a prompt and it's got sort of an internal dialogue that it does. So
495. 1:25:39all of these are basically just the amount of cycles that it allows itself to do before it gives you an answer. So
496. 1:25:46pro it it just goes for a really long time before it gives you an answer. Mini is actually a different model. Um but
497. 1:25:54it's it's just a a smaller lighter model that does the thinking. Fast is just like chat GPT40 where you give it a a
498. 1:26:01prompt and it answers you. The auto mode is basically dynamically switching between these. And I I don't like I
499. 1:26:09don't know exactly what they're doing to analyze that, but that's what auto mode does. It's it's a routing model that
500. 1:26:16basically says based on what the person's asking for, I'm going to choose one of these different models to to to
501. 1:26:21use. Okay. So, here's our creative brief for the Vibe app rescue. So, we're just going to copy this and we're going to go back to Lovable
502. 1:26:36and we're going to do a new a new app. And I'm just going to paste that uh brief into Lovable and say go. And so now this is we'll see
503. 1:26:51if this is any better. This should be better. Lovable is pretty good at this stuff. Do you do the same prompt in Grock to
504. 1:26:59see the difference in the output? Carl, that's a great great yes. So, um, one of the things I love to do is take
505. 1:27:11this same prompt and and play in a bunch of different models. So, go into Gemini, go into Grock, go into um, Anthropic,
506. 1:27:23Claude, go into Chat GPT. Um, or if you're doing if you want to play around with these agentic tools, do a prompt in Manis, do a prompt in
507. 1:27:33GenSpark, and do a prompt in chat GPT agent mode and just see what they all do. See how long they take. See how
508. 1:27:40sophisticated the answer is. Like we just did here. We were just over in chat GPT. It made us kind of a crappy website.
509. 1:27:48Then I just said, "Hey, Chat GPT, instead of writing me code, why don't you write me a creative brief?" And so I
510. 1:27:54just pasted that into Lovable. We'll see how it does. And yeah, you could paste this into other vibe coding apps for
511. 1:28:01sure. That's a great way that that's a great one of the things we talk about in the AI salon is this idea of play first.
512. 1:28:09Play first, mindfully create, generously lead is the what we call the cycle of AI readiness. And so that first one, that
513. 1:28:16play thing is rather than feeling like you've got to master a tool and you've got to make your work more efficient.
514. 1:28:23No, no, no. What you have to actually understand is what these tools make possible. Because it's possible, it's likely that
515. 1:28:32these things are way, way more capable than you think is even possible. And the only way you can find that is to
516. 1:28:43play outside of the thing that you know that's this small. Just push the edges. Do creative things if you're not creative. do coding things
517. 1:28:54if you are creative. Do writing things if you're visual. Do visual things if you're a word person. Um because in doing that, what you'll
518. 1:29:05realize is, oh my god, I didn't know I could do that. And then all of a sudden that's a new possibility in your brain.
519. 1:29:13So when you get into the next phase to mindfully create, hey, I've got a problem I want to solve. Now you
520. 1:29:19understand the capabilities here and you understand what you're asking is like don't these things aren't some of these
521. 1:29:25tools better at some things than the other. Yes. How do you know that? I just you got to play with them enough to understand. Oh,
522. 1:29:33okay. Claude's better at this thing. I like the way Claude does this. I like the way this one writes. I like the way
523. 1:29:38this one does research. When it comes to images, okay, I've got one that I use for pretty art. I've got another one that I use for text
524. 1:29:47and logos, right? The only way you're going to get there is to play. And seriously, think about yourself as as a
525. 1:30:00an orchestrator, a conductor. And the more tools that you can learn the capabilities of, it's like you've
526. 1:30:08got virtuoso m musicians all around you. Right? You need to know where to go and which musician to point to. But you can
527. 1:30:18just sit there with your idea and go, "Oh, I'll go try this tool for that. I'll try this tool for that. Oh, those
528. 1:30:23two work really well together." Right? All right. What's this thing doing here? She's making us a stupid website. Come
529. 1:30:32on, man. Come on, man. Where's our website, man? It's got a bunch of JSON. It's got some tailwind action in it. Oh, here we're done.
530. 1:30:50Working on Galaga. Yeah, exactly. The reason we do video games is video games are fun. There we go. This looks like a website,
531. 1:31:00right? Your vibecoded app ships fast. I make it safe, solid, and fast. Fast is wrong here. So, we can actually go
532. 1:31:07change this. Let me go edit this. Go edit. Then I can go I make it make it sa safe solid. Wait, I make it your vibe coded.
533. 1:31:21I make it safe, solid, and secure. Right. Boom. 30 minute triage call. Problems I fix fast. Security holes.
534. 1:31:35Performance bottlenecks. Flaky features. scaling nightmares. Here, let me get out of edit. Um, user trust issues, technical debt,
535. 1:31:46right? From crisis to confidence, free 30 minute call to triage, how bad you [ __ ] it up. One to two days, I'll
536. 1:31:56do a health check. One to two weeks, I'll secure it and stabilize it. Two to four weeks, I'll actually make it not suck.
537. 1:32:04This is a [ __ ] business. Boom. done. I made a a Galaga a Galaga game in Perplexity Labs. Yeah, that's pretty
538. 1:32:16cool. Perplexity is pretty good. Um, but anyway, Arum and anyone else here who's a coder, I'm telling you, the orphan the
539. 1:32:25orphan rescue industry in vibe coding is going to be big. It's going to be big. Now, you should also learn to vibe code,
540. 1:32:42right? You can go in and you can fix [ __ ] up vibe coding things, but you should actually understand which of
541. 1:32:49these things are the good things. I would imagine that cursor makes similar kinds of mistakes over and over again. I
542. 1:32:55would assume that lovable makes different kinds of mistakes over and over again. How do you know that? By
543. 1:33:00playing with them, by doing it yourself. At some point, one of these things is going to fix their vibe coding platform
544. 1:33:08so that they may not need you anymore. That would be a good thing to know if you're basing a whole business on fixing stuff, right?
545. 1:33:17Lovable.dev/partners. There you go. Um, I assume what that is is a whole referral system that they've got
546. 1:33:28for people to come in and unfuck up Hire a lovable partner. There you go. Go sign up for this. Perfect.
547. 1:33:43Yeah. Yeah. Here's companies. How many are in here? Six. Eight. There's eight companies in here that will go fix lovable things. They just
548. 1:33:55raised $200 million. They've got a hundred million. They got to Didn't they get to 100 million users, Brandon? I think they did.
549. 1:34:08No, 200 million. 200 million in investment, I thought. And I thought it was 100 million users. Anyway, whatever.
550. 1:34:15But this is one company. There's only eight eight people that can fix stuff. Go sign up. Starting at $100 an hour, 75
551. 1:34:25bucks an hour, starting from 1,800 bucks, starting from three grand, starting from a grand. Starting from $4,500 a month,
552. 1:34:35starting at $10,000 a month, 200 $200 an hour, right? Everyone's going to get laid off. Well, yeah. a bunch of people that are going
553. 1:34:46to get laid off are going to start these companies and other companies and other companies and other companies, right? Because
554. 1:34:55every one of these new things is going to create new problems that we'll need to go in and fix. We'll adapt. Adaptability. Adaptability.
555. 1:35:06It is the the skill of the next five years. Don't get too attached to what you built that's genius because someone's going to steal it
556. 1:35:20or or just AI will just do it, right? The thing it can't do right now, it's just going to do. And then every three weeks there's new
557. 1:35:32technology. The technology that you thought you knew two weeks ago is now different. It's now different again. Just adaptability,
558. 1:35:41being nimble, being zen about, oh, [ __ ] changed. That thing that was broken, not broken anymore. Next. Move it on. Move
559. 1:35:52it on. Move it on. Okay. [Music] Who stole your song? I wonder why someone smarter than me isn't thinking about different
560. 1:36:06refrigerant than water. Oh, nothing creates opportunity quite like companies that underestimated how much they needed the people that they
561. 1:36:16just laid off. That's going to be another one. Oh, we [ __ ] up. We shouldn't have fired all those people, right? I'll come
562. 1:36:24in with my SWAT team of 10 people and we'll, you know, we'll write your company. That That's a whole another one. Okay. So,
563. 1:36:35um I gotta go. My voice is going. I haven't drunk enough water and I'm I'm uh I'm crispy. But um I'm really glad
564. 1:36:42there were some new people in here tonight. Um go check out the AI salon if you haven't. Go to community.thesalon.ai
565. 1:36:54and that's going to take you into our community. Sign up for it. Here's why. Nobody right now, nobody knows what the future's going to look
566. 1:37:06like, knows how these tools are going to evolve, knows how the future of work is going to evolve, knows what are going to be the new jobs.
567. 1:37:17The only way that I know that you can insulate yourself a little bit from what's coming is to be in a community of
568. 1:37:24people that are trying to figure this AI stuff out as well. You don't want to be hanging out with people that are on the
569. 1:37:31sidelines just pissed off about AI because AI is going to hit them upside the head and it's going to be ugly. You
570. 1:37:37want to be in in a community with people that are trying to figure this out so that as new things come along they're
571. 1:37:43like, "Oh, I know what to do with that." And that'll give you an idea. And then you'll give someone else an idea.
572. 1:37:50I talked before, play first, mindfully create, generously lead. The AI salon is about people sharing what they learn and
573. 1:37:59and being generous in in how they support other people. I don't know if you've noticed, if you're new here
574. 1:38:04tonight, have you noticed that there's a lot of people in here that are like answering questions and doing things if
575. 1:38:10I miss them? Those people are called the irregulars. And they're irregular. We call them that because they're weird.
576. 1:38:19They show up here night after night after night after night. A bunch of stuff that I've talked about tonight,
577. 1:38:26they've heard me talk about before. You know what they do when they hear me talking about stuff they've heard
578. 1:38:33before? They're off in AI playing with it or they're in here answering questions. They're making relationships.
579. 1:38:39They're building trust with people. That's what the AI salon's about. Okay. Next Tuesday, we've got an AI salon meet
580. 1:38:48and greet. So, go join the online community. And then next Tuesday we have uh what we call a meet and greet and
581. 1:38:55it's a very special one because we're doing the official launch of the AI readiness training program which if you
582. 1:39:02want to check it out the site is now live. It's are you ready for a.com and I think we have a banner in there don't we?
583. 1:39:16Yes. Is this it? Yeah. If you go to are you readyforai.com right there, that's the AI readiness training
584. 1:39:28program. This is a very comprehensive program that was inspired by a bunch of um we we did a 24-hour thing called AI Festivus last January
585. 1:39:40and the the talks were really remarkable. And so this is all of those talks, but we we extracted out of it
586. 1:39:47what were the universal truths, what were the universal lessons we learned over that 24-hour period that can get
587. 1:39:54you into the mindset to be successful about AI. That's what that that's what that program is. And then the last thing
588. 1:40:00I'll I'll talk about tonight, last little promo is tomorrow night's what we call Friday night date night. So, so we'll do one of
589. 1:40:09these again tomorrow night and and it'll be normal time, 8:00. Um, it's just a Friday night. Bring nachos or or hot
590. 1:40:16pockets, whatever, whatever you enjoy. Um, but tomorrow at 11 a.m. Mountain time on LinkedIn is what I call AI
591. 1:40:26office hours. And it's basically just a one-hour hangout with people who are part of the AI salon part people that
592. 1:40:35are part of this some people that just come you know from LinkedIn and just hang out there. It's really a remarkable
593. 1:40:40group of people. Um so if you want to meet some people um come tomorrow for AI office hours. So how you find it is you
594. 1:40:48go to my uh LinkedIn. So it's Kyle Shannon on LinkedIn. Uh I'm the CEO of Storyvine and the AI salon. So you'll be
595. 1:40:56able to find me. I look like this. My picture's there. Um, so yeah. So, so come to AI office hours tomorrow and then come tomorrow
596. 1:41:07night for this uh AI learning lab. All right. That was a rant tonight.
597. 1:41:23But thank you all for hanging out. I really appreciate it. And uh All right, looks like let me just look at some comments here.
598. 1:41:38[Music] Thank you, Ryan. Thanks for the nice words. Thanks for the good questions. By the way, I love questions in here. I
599. 1:41:46even love questions. You know, sometimes people come in and like, "Isn't AI just evil?" Like, you know, they'll come in
600. 1:41:51and be a little bit trolly. I don't mind trolly kind of answers or questions. Um because one of the things I recognize is
601. 1:42:00that this shit's scary. And especially if you're not in it, it's scarier. But even if you do get in it
602. 1:42:11there, you might have ethical concerns, you might have safety concerns, you might have, you know, security concerns,
603. 1:42:17you might have future of work concerns. Those are all really valid. Um, my antidote to that is to be in the
604. 1:42:25conversation. And so I absolutely love it when people come in here with those kind of questions so long as they're
605. 1:42:32willing to actually like give it a shot. Like try it on. Like come in, have those questions, have those doubts, and play
606. 1:42:41with AI. Play with it long enough until you have your first what I call Kevin Mallister moment, right? Remember Home Alone? that moment.
607. 1:42:52Once you have that moment with AI, you kind of can't unhave it. Once you actually get the generative part of AI, the G in GPT
608. 1:43:06stands for generative. Once you get what that actually means and what it can do for your ideas and and your projects and your work,
609. 1:43:17it's hard to go back from it. Um, and that's all I ask is like try it enough that you have one of those moments. And
610. 1:43:25then if at that point you're like, you know what, [ __ ] it. I'm tapping out. I'm going to the woods. I'm getting I'm
611. 1:43:31going to dust off my prepper kit and go move to the woods. More power to you. But I think what you'll find is it can
612. 1:43:39be very very inspiring if you get into it. All right. The real just joined the salon. Awesome. That's great. Great. Great. Great.
613. 1:43:51Oh, I was I wasn't saying that you meant you meant it in any way negative, Ryan. I I was just saying sometimes people
614. 1:43:57come in here and they do, and I'm I'm fine with that, too. I didn't I didn't I thought your questions were were great
615. 1:44:03and very productive. Um, good job teacher. Thank you, Peter. I appreciate that. My grandson thinks AI is creepy and he
616. 1:44:13works in pest control. There you go. All right, everyone. Have a good night. I'll see you at office hours tomorrow at
617. 1:44:2111 and then I'll see you back here tomorrow night for Friday night date night. All right. Peace out, everyone. Have a good night.