10/8/25 - Fulfilling Steve Jobs' 1985 Dream of an Interactive, AI-Powered Aristotle

Transcript

1. 0:00that before. Like we could have done that earlier and given you a little more runway, but All right. Well, we'll do that. We should we
2. 0:08should actually probably do that for all the big events. Um All right, Tam. Don't you think it's time? Huh?
3. 0:20You ready? You ready to come singing? [Music]
4. 0:35Hello. [Music] Hello. [Music]
5. 1:09[Music] Every time I see you now, get that look in mind. Every time I see your man thought, hear that smile.
6. 1:26The early misty morning light that I heard the engine turning the old outside. [Music] You leading me
7. 1:42again today. You will convince me a good day. Leaving this hotel looking for someone else's golden ring.
8. 2:00to say so long Suz. [Music]
9. 2:18[Music] >> Don't you cry for me. [Music]
10. 2:40[Applause] [Music] [Applause] Hello. [Music] Hello. [Music]
11. 3:05There's been something baby I've been trying to say for an age and it seems I don't know how [Music] the past and the future now surrounding
12. 3:20me. surrendered to a heaven and she thrill can't be found. There's been a little trouble since you came to my rescue.
13. 3:38And if you like all of the rest, I would have quit you long ago, but I couldn't do that. [Music] Oh, tell me now when I never went to
14. 3:53well, make a man crazy, make him cold as hell. I'm a woman that you wish me well. But in spite of you trying, still going
15. 4:09to have to find my own way through. [Music]
16. 4:26Oh, good people, good people, good people of the Tik Tok, good people of the internet. I hope you're doing well today. I had a shitty day.
17. 4:41Just gota Sometimes it gets intense, man. Sometimes you're out in the field, it just gets intense. You know what I'm saying?
18. 4:52That was my day. And it wasn't there was not nothing particularly challenging about it. It was just one of those days
19. 4:58where a lot of people are feeling it. So, I don't know if you've had one of those days where you're just feeling it.
20. 5:11And feeling it ain't bad. But it's feeling it. Some days you're the fly, some days you're the windshield. Exactly. This was a fly day.
21. 5:27Sorry to hear that. Oh, it's that's that's life in a big city source camp, but I appreciate it. That's uh you know,
22. 5:35I mean, listen, you're you're an entrepreneur. You know what it's like. you know, you you uh if you've never
23. 5:44started a company, good on you. If you have started a company, you know, it's just like, you know, you know what's
24. 5:52funny about starting a company is I've started I don't know 15 20 companies in my day. Some went big, some didn't deserve to
25. 6:05make it past a week and didn't and you know, lots in between. Um, and probably every day, every day
26. 6:23there's some problem where you're like, "Oh, someone should take care of that." And then you realize you're it
27. 6:32like there's no one there to take care of it. Like you're it. And that's that's what it is to run a business is just you
28. 6:39know the the the reason you start a business in the first place is because you're dissatisfied with something in
29. 6:48the world, right? There's got to be a better way, right? I'm going to make people happier with my t-shirts than all the other shitty
30. 6:57t-shirts out there. You've discovered there's some problem with the world. And then you start a company and then uh
31. 7:05it's just a series of grievances from you and trying to fix them. It's perfect for Gen Xers. Oh yeah, I know those days. Yeah,
32. 7:17exactly. Yeah, many like that. Um, I wanted to we we'll do we'll do a wax philosophical Wednesday. I wanted to
33. 7:32do a little We're actually not gonna I well I don't think I'll talk too much about it, but I just stumbled on a video that I've
34. 7:41seen I don't know 20 times 50 times and it put me in touch with
35. 7:56like you know I was talking earlier about feeling it like feeling what's going on in my business feeling what's
36. 8:01going on and the AI salon, which I'm trying to sort of take to the next level, feeling what's just going on in my life. Um,
37. 8:14and so I saw this video and it put me in touch with what remarkable times we're in. We're we're in really remarkable times. And,
38. 8:28you know, it's funny. I saw I saw an interview with Sam Alman this week and he's like he's like you know for 50
39. 8:37years for 60 years 60 or 70 years we had this thing called the Turing test which was basically if you sit someone
40. 8:48down at a computer screen and they can interact with it and they can't tell if it's a human on the other side or a
41. 8:54machine then we've passed the turning test. It's like we passed that and just blew past it and then you know we've now
42. 9:02got machines that are smarter than us and we were like oh yeah we got that now and we just blew past it and now we're
43. 9:08like why can't it be faster if we completely normalize it? Tik Tok question. Are you concerned about misinformation with AI? Um sure. Yeah.
44. 9:17But I mean, we haven't needed AI to have like the last decade of of at least America, but really globally, the last decade
45. 9:28has been an a misinformation hyper war. Um, you know, what's funny about AI misinformation is I think that
46. 9:41this is actually kind of fascinating. I had I hadn't really thought about it like this, but but there's so there's
47. 9:48there's there's this thing I don't know if you've heard of it, but you probably have called AI slop. And basically what AI slop is is the
48. 9:57term that the creative community gives to anyone that creates anything with AI, right? But that's not actually that it's
49. 10:05not actually right. AI slop is when lazy people use AI without thinking and they just generate slop. Now, the the naysayers of AI say that
50. 10:19all AI is that. It's not. It's just not. Um, but what I've noticed is that the people that I know that are that are really
51. 10:33doing great work with AI and thinking critically about AI and all that sort of stuff, a lot of the members of the AI salon, when we see
52. 10:43crap out of chat GPT, you can you you can just sniff it. You know what it you know you know what it is, right? And
53. 10:50when you see an image that was just lazily prompted, for the most part, you know what it is. You know it's AI and
54. 10:56you know it's it's lazy AI. It's slop. And so I think that there's there's a weird there's a weird possibility
55. 11:08that AI is actually going to be the antidote to misinformation because there's going to be so much of it but it
56. 11:15will all have a sameness to it that authentic messaging will rise above the noise. That's it. It is a possibility.
57. 11:24So, so a possibility is we just get better at sniffing out [ __ ] and then and then the authentic stuff rises above the noise. But
58. 11:33misinformation that's weaponized is more sophisticated than that. And then I think what happens is this. Then I think
59. 11:39AI helps us build tools that allow us to instantly be able to assess if something is misinformation or not. Um the the weird thing about
60. 11:54misinformation is we're we're in such a polarized charged, you know, my analogy for the United States right now is like
61. 12:03we're we're two football teams and it's like you can't tell someone from the Pittsburgh Steelers that they should
62. 12:10love the Dallas Cowboys. Like it it doesn't matter what argument you make like Dallas sucks or like Pittsburgh
63. 12:19sucks. That's where we are as a country right now, right? So in that environment, misinformation's really easy because you
64. 12:29just make information look like Pittsburgh Steelers friendly stuff and they're like, "Oh, that's good." And
65. 12:35then you make it look like Dallas Cowboy bad. Oh, they're bad. There's there's literally no critical thinking going on.
66. 12:42So maybe there's a possibility that AI actually helps in the analysis and sifting of shitty information to actually say here's
67. 12:54actual neutral valid information. I don't know that there's an appetite for that. I think what the social media
68. 13:03companies have mastered over the years is monetizing our polarization. So, you know what I would say to you is
69. 13:14go vibe code up a social media network that that just pedals in information. No information, no misinformation allowed.
70. 13:22And see if anyone shows up. They probably won't. There's never been any shortage of misinformation. Pretty sure
71. 13:33some of those Samanskrit clay trap tablets were fake news. That's really funny. Um, by the way, I'm from Dallas
72. 13:39and the Cowboys are that bad. Well, that's that's that's a whole separate thing. I was using hypothetical sports
73. 13:46teams, but yes, congratulations on on living in in where America's team is. Oh, man.
74. 14:02All right. So, here's what I want to do. The the Browns can't even win an AI. Um,
75. 14:16we just celebrated celebrated. We just marked the 14th anniversary of of Steve Jobs passing. Um, America can have them. Uh,
76. 14:31commemorated we commemorated the passing of Steve Jobs. And it looks like Tim Cook is stepping down as CEO,
77. 14:37which I think it's probably about four years too late, but that's neither here nor there. I think he was a good CEO. He
78. 14:44he stabilized Apple in Jobs's absence, but they they they need uh they needed they need a visionary, and I think
79. 14:53they're bringing on a CEO who's not a visionary, but whatever. They didn't ask me if I wanted to run it, so so my
80. 15:00opinion doesn't [ __ ] matter. um as as flawed an individual as as he was.
81. 15:14Um he was a he was a hero of mine. um and in a lot of ways is still a hero of mine because he could he could see the future and and and not
82. 15:31in like a like a psychic kind of way. He just observed the world and he looked at he looked at what happened when you
83. 15:42stuck a bunch of transistors on a circuit board and programmed them in a certain way. And he could he could connect the dots forward
84. 15:53about, you know, okay, this is a piece of crap today, but let me let me project forward and what what would this thing
85. 16:01be like if it weren't a piece of crap? And then he makes the Apple 2 and and changes education and and then he goes
86. 16:11to Xerox Park and they show him the the first computer mouse and he says, "What are you doing with that? What are you
87. 16:18doing with this operating system?" Well, we're not doing anything with it. And I think he bought it for 25 grand or 50
88. 16:24grand and he made the Macintosh and he, you know, single-handedly changed things and just kept doing that
89. 16:31again and again and again and again. And it wasn't it wasn't magic. It was it was observation and it was you know touching
90. 16:40into tapping into his humanity. And so this video it's a fairly famous video but this is the video where he he
91. 16:51predicts he predicts AI. He he he talks about AI here. And I think this is when was this? This was this is an old video.
92. 17:071985 I was in my second year of college at Penn State. >> The great >> 1985 he said this. We're in 2025. How
93. 17:20many years is that? 15 plus 25 40 years ago. 40 years ago. Brandon, producer Brandon was two. >> You know who Alexander the Great's tutor
94. 17:35was for about 14 years? You know, Aristotle. And I read this, I became immensely jealous. Uh, and I think I would have enjoyed that a
95. 17:47great deal. and and uh through the miracle of the printed page I can at least read what Aristotle wrote without an intermediary
96. 17:59and maybe if there's a professor they can they can add to that but at least I can go directly to the source material
97. 18:05and that is of course the foundation upon which our western civilization is built but I can't ask Aristotle a
98. 18:12question I mean I can but I won't get an answer and so I My hope is that in in in our lifetimes we can make a tool
99. 18:29of a new kind of an interactive kind. And so my hope is someday when the next Aristotle is alive, we can capture the underlying world view
100. 18:43of that Aristotle in a computer. And someday some student will be able to not only read the words Aristotle wrote, but
101. 18:51ask Aristotle a question and get an answer. And uh that's that's what I hope that >> that's what I hope that we can do. 1985
102. 19:05he somehow turned in that speech he somehow turned the printed page into a radical technology. I was jealous that I didn't get to have
103. 19:19Aristotle teach me. But through the miracle of the printed page, right? He's he's going back hundreds of years. Through the miracle
104. 19:29of the printed page, I can go directly to the source. And wouldn't it be nice if I could ask him a question? We've been able to build this
105. 19:42what he just described for at least two years now. And so I thought what might be fun tonight is to go build this tool that he described.
106. 19:57And I'm going to do it in two ways. I might do it in three ways. Um we're going to go to Notebook LM.
107. 20:14And if you want to play along, feel free to play along. Um, I'm going to create a new notebook. If you don't know, if you don't know
108. 20:23notebook LM um, go to notebook LM as in language model.google.com and make an account if you don't have one.
109. 20:37And with this tool, you can upload up to 50 documents and they can be anything. They can be videos. They They can be
110. 20:44YouTube links. They can be websites. They can be Google Docs. They can be PDFs. And so what we're going to do is before the call tonight,
111. 20:57what is this a call? This is a live stream. Before the live stream tonight, I went and I got a PDF of Is it? No. Where did I get it? Um
112. 21:16let's see the complete. Okay. So go here go here we go the dash complete
113. 21:31the complete Aristotle tab. I'm sharing the right tab. Oh I'm just not I'm I'm just not here. Okay. So
114. 21:40you can go right now on the internet and you can search for the complete Aristotle and it's like a 7,000page
115. 21:46document. The guy wrote a lot of words. What AI we're using right now is notebook LM. And then we're going to use chat GBT.
116. 21:55Uh and then we might even go vibe code something. I might go to Claude and vibe code something. Thank you for this demo. It's a powerful
117. 22:05learning tool. Yeah. So we're gonna do a couple of things. So right now um the little circle is still spinning. So So
118. 22:12it's it's reading notebook LM right now is reading Aristotle's complete works. The provided excerpts from Aristotle's
119. 22:25work primarily focusing on categories and analytics meticulously outline the fundamental concepts of logic,
120. 22:31metaphysics, and physics. The text explores the nature of substance and how different things. Furthermore, the passages introduce and define
121. 22:40correlative terms arguing for their interdependence and simultaneous existence. I don't know what the [ __ ]
122. 22:46they're talking about. I can't even understand it. But you know what I'm going to do? I'm going to go over here
123. 22:51to this thing called audio overview and I'm going to push the button. And so while that thing's generating an audio
124. 22:57overview, we're going to go build something else. Tell me when AI can create any app better than a human. Oh, Robbie's an AI
125. 23:05hater. I understand that. Um, I understand the fear. Um, it depends on which human you're talking about, Robbie. Um, if you're talking
126. 23:17about me right now today, AI can build a better app than me. But I don't think that's what you're talking about. I
127. 23:24think what you're talking about is tell me the day when AI can build an app that doesn't require a human to fix the the
128. 23:31the gaps, right? And the shittiness and the jankiness and none of them come close today. So So you're good for you're good for right now.
129. 23:47Um, I would be surprised and I'd be willing to to I'd be willing to wager on it that within I'd probably wager on three years, but
130. 24:00I'll predict within two years. Um, you'll be able to you'll be able to prompt an app into existence that's
131. 24:12better than what humans can build. I I'm fairly confident of that. um codecs right now. So, so let's see. Um the new the new Sonnet 4.5,
132. 24:24is it 4.5? Is that what's the latest one? That thing just coded for 30 hours straight. They basically gave it a task
133. 24:33of of recreating Slack. It coded for 30 hours straight and delivered a functional application. Um, Codeex from
134. 24:41OpenAI right now is coding for about a day, a full day, um, probably in the 24-h hour period. Sam Alman thinks that
135. 24:50very shortly it'll be able to code for a week straight. Um, so I would say just in terms of marathon, um, it can it can write better than a
136. 25:03human. Um, in terms of quality, we're not anywhere close yet. So, so that's that. What what I would say to to you is this, Robbie, is
137. 25:18if you're using if you're using the fact that AI's not good enough, that that humans are superior. If you're using it as an
138. 25:30excuse to stay on the sidelines of AI, I would first of all, I wouldn't stay here and hang out with us because all I talk
139. 25:36about is AI and you're probably going to hate it. So, you might want to just go. Um, but if you're on the sidelines with
140. 25:44AI, I would strongly suggest don't be. And not because I particularly care if you use AI or not, but don't assume that the limits that it
141. 25:55has today are going to exist much longer. And and my philosophy is I would much rather be using the tools and understanding the tools than
142. 26:07sitting on the sidelines and just watching it happen because we're in an acceleration right now. Like the
143. 26:12acceleration is accelerating and these tools are getting better faster and and that's going to keep that's not that's
144. 26:20not stopping. We're not hitting a wall. And so, you know, my my my mo why I go live five nights a week is
145. 26:29just to say [ __ ] try it on. Strap it on like a jetpack and and see where the edges are. And you'll you may find,
146. 26:36yeah, it sucks, but you know what? You're probably good enough that you can fill in those gaps right now. And you
147. 26:43know what? that makes you incredibly valuable for all the people who are not coders that are trying to use vibe
148. 26:48coding to start businesses and they create applications that have privacy and security holes in them, they're
149. 26:55going to need to hire someone like you to come in and fix them. But if you don't understand what the the challenges
150. 27:00are with vibe coding, you won't work, right? So anyway, all right, new question. Um, won't the AI subscription cost go up if
151. 27:11they're coding 247? Well, the on average the the per token processing price for AI has dropped 10x every what is it Brandon? Every nine
152. 27:25months it's it's at least every year. Um, that the the quality something like that. Yeah. that that the the quality of interaction goes
153. 27:36up and the cost per token is going down at 10x at least every year and I think it's shorter than a year. So, so what that
154. 27:45looks like is the the efficiency the efficiency is going to go up um as well. And so what's going to so happen the
155. 27:54efficiency is going to go up and then the algorithms are going to get better. They're getting better and better and
156. 27:59better at training the corebased models. They're getting better and better and better at reasoning and now they're put
157. 28:06building in this agentic behavior and these things are just going to get more and more and more efficient. So in
158. 28:12theory, yes, but I think in reality, no. Um, and here's the other thing. Um, most people don't have a software development program
159. 28:23challenging enough to need their chat GPT to or whatever they're using to code for a week straight, but commercial
160. 28:34software developers do and big corporations do. And so if if you're thinking about this like a uh like an
161. 28:42end consumer, it might be too expensive to let a let a machine code for a week straight. But if you're a corporation
162. 28:50and you now don't need to hire an engineer to do that, then then it's cheap. It's very cheap. Uh Pav, my daughter did a whole NASA
163. 29:01project with AI help. Um if used right, AI is great. Yeah. Yeah. Exactly. Here's the deal. Here's the deal. All right. This is super
164. 29:11important, especially if you're new here and you're on the skeptical side of this thing. Here's one of the biggest
165. 29:17mistakes people are making right now with AI. Let me let me go go soapboxy here for a second. One of the one of the one of the biggest
166. 29:30mistakes people are making with AI right now is they're treating it like a genius, right? And so in the education
167. 29:40system, they're like, "Oh, AI is cheating because the student's just going to push a button and it's going to
168. 29:45write a report and then they're going to turn that in and that's cheating, right?" Or they're like, "Oh, someone's
169. 29:51just going to push a button and it's going to create an image and that's going to steal from illustrators and
170. 29:55that's AI slop and that's cheating." Or, "It's going to make this video and that's going to ruin Hollywood and
171. 30:00that's cheating, right? They're treating AI like this thing outside of us that we're in competition with it."
172. 30:07That whole idea I find bizarre, but that comes from Hollywood, right? We've got 50 years of Hollywood telling
173. 30:16us that it's us against the machine. There's a whole another way to look at AI. There's a whole another way to look at
174. 30:27AI. It's not us versus the machine. It's us having being lucky enough to live at a time in history where we get to take our ideas
175. 30:39and put them into this machine that reflects them back to us in this amplified way. It's us with the machine. It's us
176. 30:49amplified by the machine, not in competition with it. We get to use it, too. All of the the conversations about
177. 30:58AI is going to take the jobs. Yeah. And then all those people that lose their jobs have access to AI where they can go
178. 31:04start a company with like a virtual 10person super team of AI PhDs that are on their side. That's absolutely [ __ ] remarkable. No
179. 31:15one's talking about that. So your daughter in doing that NASA project, sorry, black bar in doing that NASA project, she's like,
180. 31:24I got this NASA project. I want to do the NASA project. Let me have AI help me. Help me do the research. Help me
181. 31:30bring it in. Let me get excited about it. Let me put the things together. Right? It's an augmentation of her ideas.
182. 31:39Well, isn't that going to make her more dumb because she's not learning? No, she is learning. She's learning in an
183. 31:44accelerated way because she's taking her ideas and expanding her mind, right? If you treat AI like a genius off to the
184. 31:54side, oh, make write me a NASA paper, it'll do that. That's got nothing to do with you. That's just what lazy people do. There's
185. 32:04going to be a lot of lazy people. But anyway, okay, let's go back to to Aristotle and Steve Jobs. Let's see if we got our our
186. 32:14audio overview. We may it may be done because I ranted. Um oh, here it is. >> Welcome back to the deep dive. Today we
187. 32:31are taking on a source that well isn't just deep, it's truly foundational. We're tackling key excerpts from
188. 32:36Aristotle's work. uh basically the original blueprint for how western thought defined reality, organized knowledge and well figured out how
189. 32:44things work and move. >> It is a huge undertaking. Absolutely. But our mission, I think, is clear. We
190. 32:50>> Okay, so to be clear, what what you're experiencing right now, if you haven't seen uh Notebook LM before, I'm going to
191. 32:57also do a mind map because that'll be pretty wild. What you're listening to right now is a podcast with a a guy and a gal that are
192. 33:07the notebook LM guy and gal from the audio overview and it read all the 7,000 pages of Aristotle and now there's a 15minute
193. 33:18podcast that's going to talk about him. But there's a thing that you can do here that this would blow Steve Jobs mind,
194. 33:25right? So, if if you're new here, we've been watching watching Steve Jobs. Um, how to add one more than one source. So,
195. 33:38over here on the left hand side is your source list. So, right now I've just got Aristotle, but I could go out and I
196. 33:44could go find Plato, you know, all the the the the complete works of Plato and Shakespeare and throw them in here as
197. 33:51well. Um, so basically there's just an add button right up here. If you click that add button, it gives you the the
198. 33:57add dialogue box. You can put in up to 50 documents. And the the documents are big. I think it's like two million two
199. 34:04million tokens per document you can put in here. So, what is that? 500 million tokens. A novel is about 50,000 tokens,
200. 34:13something like that, or 50,000 words. I feel bad Steve Jobs isn't here to see it all. I know. Me, too. Like I Yeah,
201. 34:20exactly. Wait, do you tell do you tell it anything upon entering a source? No. No, you just up. So, it's got the
202. 34:28complete works of Aristotle. But watch this. This is absolutely mindboggling. So, I'm going to I'm going to put this conversation
203. 34:35into interactive mode. >> Welcome back to the deep dive. Today we are taking on a source that well isn't just deep, it's truly foundational.
204. 34:47We're tackling key ex >> and so now I can join into this conversation. So so I'm going to raise my hand here in a second
205. 34:54>> from Aristotle's work uh basically the original blueprint for how western thought defined reality organized
206. 35:00knowledge and well figured out how things work and move. >> It is a huge undertaking. >> Oops. >> Absolutely. But our mission I think is
207. 35:08clear. We want to give you a kind of structural shortcut. We're aiming to extract the core framework. Aristotle
208. 35:14established the rules he set out for defining things. Classif Oh, hey there. How can we help? >> Yeah. Hey, I just was wondering Oops.
209. 35:25Disconnect. I don't know if my audio Can you still hear Oh, wait. Hang on a sec. Allow while visiting the site. Can you
210. 35:32guys still hear me on TikTok? >> Oh, that's a great question about the length. >> Are you asking about the duration of the show?
211. 35:37>> If you mean >> Oh, yeah. Let's hear it. >> No, no, no. Here's here's what I'm asking. I'm asking about um about um what
212. 35:46>> Thank you for clarifying. >> Well, you know, it's fair to say you don't >> Oh, back to our listener. What's up?
213. 35:52>> Hang on. Sorry, I couldn't I couldn't gather my thoughts there for a second. What I was wondering is if if if this
214. 35:57was the beginning of modern thought, what was thought like before Aristotle? Like, what is it that he did that
215. 36:03changed things so much? And you know what? This is embarrassing to say. I don't even know when he did what he did.
216. 36:09I don't know when he was alive. And so what was like life before it and then when did it change because of what he did?
217. 36:18>> Thank you for clarifying. >> You know, it's totally fair to say >> need to be embarrassed at all.
218. 36:22>> That is a phenomenal question. Actually, >> it allows us to step back and look at the context. >> Aristotle was alive in the 4th century
219. 36:28B.C.E. >> roughly 384 to 322 B.CE. >> And what he changed was pretty monumental. Before him, much philosophical thought, especially with
220. 36:36Plato, focused on the world of eternal forms. >> The idea was that the perfect abstract version of a thing was the most real thing.
221. 36:44>> Like the pure concept of justice ex >> Oh, yeah. Let's hear it. >> So, could you play act with me as Aristo?
222. 36:53>> So, could we just elaborate on that idea of the forms? >> Absolutely. Wait, >> go for it. >> Yeah. Could you just play act with me as
223. 37:00one of you play Aristotle? And I just want to talk to Aristotle. >> Oh, that is a fantastic request.
224. 37:07>> That is definitely a first for the deep dive. >> We can absolutely do that. >> You want to talk to the stage right herself?
225. 37:13>> I will do my best to channel the spirit of the philosopher. >> I will uh be your humble scribe perhaps or assistant.
226. 37:20>> Excellent. Ah yeah. Uh-huh. >> Okay. So anyway, um that took zero amount of time. Like like what struck me when I saw that video. I
227. 37:33was just flipping through X and I saw that video of Steve Jobs, you know, kind of dreaming about the day when you could
228. 37:40talk to Aristotle. You could talk to the next Aristotle. You could talk to the previous Aristotle. Um and we can now.
229. 37:47So I I wanted to do this thing in Notebook LM. Let's go look at the mind map. So here's a mind map. So Aristotle's work the categories the
230. 37:56logic the scientific knowledge and the dialectic topics and sophistical reputations. And so if we click on that then it
231. 38:08breaks it down even further into into three categories and and you can just keep going. Every one of these you can you can click on
232. 38:18and then notebook LM will do a summary of it. And so you've got you've got a whole thing here where this is now organized in an interactive
233. 38:30way a 7,000page document. Um, okay. So we're going to go do one other thing. So I'm going over to chat GPT
234. 38:41and I'm going to go new chat. Tik Tok question. Somebody had a question that just the pin just went away. Hey Murphy,
235. 38:46what's happening? Are they running a dual double helix pipeline between each other? Johnny D, is this a good model to create
236. 38:56research paperwork? Um, is is the model that I was just looking at with Okay, so there's there's a couple of
237. 39:11things right now. You can use Claude can do this. Perplexity can do this. Notebook LM can do this. Gemini can do this. OpenAI can
238. 39:23do this. All of them have some version of a thing where you can upload documents and then interact with those
239. 39:28documents. Notebook LM is particularly good at reducing the hallucinations that it brings in. So, so when you go into
240. 39:39notebook LM and let's say you update, you upload 10 different research papers into it and now you're just chatting
241. 39:46with it and basically what you're chatting with is you're chatting with this intelligent entity just about the contents of those 10
242. 39:55papers. Now, how it gets its intelligence is it's using the large language model that's pulling from all
243. 40:00of the other knowledge, right? But they they have done something within Google to make it pretty good at just
244. 40:08interacting with the information that's in in your notebook. That said, it's still AI. Anything that's a research paper, I mean, um
245. 40:21Deote uh Deo Deote just got busted by the US government. They charged them $400,000 for a paper that they delivered that had
246. 40:32fake um uh citations in it. They didn't [ __ ] proofread it. So, they're having to give them back like 250 grand. So, I
247. 40:41would say if you're doing a research paper, if it's intended for publication, make damn [ __ ] sure, no matter what
248. 40:48tool you're using, that you're actually paying attention to what it produces. because you should assume that
249. 40:56everything it's writing and and all the citations it's doing are incorrect. The problem is 90 like it used to be
250. 41:04like I don't know 70% of them were correct and 30% were hallucinations. Sometimes it was worse. Now we're now
251. 41:11they're getting better but it's still it's still bad. So a way to do it is go go to Perplexity and do it there. Go to
252. 41:18notebook LM and do it there. Go to Claude and do it there. use multiple tools and then actually kind of Frankenstein together your research
253. 41:25report um when it when it gets things right. Does anyone here speak Lisp? I do not speak Lisp. Although Weaver, your
254. 41:33name is Weaver and if you have not checked out weaver.ai, it's a pretty pretty cool tool for it's doing nodebased creative work. So you can
255. 41:43create a prompt and then generate like nine different images using nine different image models with connecting
256. 41:48nodes. It's pretty slick. Kuno, don't forget we get a free year of perplexity pro with PayPal or Venmo. Oh, yeah.
257. 41:55Yeah, that's the other thing. If you go into what is it? You go into PayPal. Brandon knows this. I forget where you
258. 42:06go. You go somewhere in there. Anyway, okay. Let me show you something here. Why? Okay. Okay. Okay. If you have not heard of custom
259. 42:21GPTs, we're going to go there. Now, I know that there are a ton a ton of custom GPTs that use Aristotle as as the
260. 42:38source material. But I want to show you how quickly we can make one of these. Again, you know what? I'm going to go
261. 42:43play the video again just because there's probably different people in here than when I played it whatever it
262. 42:48was an hour ago, 45 minutes ago. Um, let me share this tab. So, this is Steve Jobs in 1985 talking about the revolutionary invention of
263. 43:06um, the printed word. Do you know who Alexander the Great's tutor was for about 14 years? You know, Aristotle. And I read this, I
264. 43:16became immensely jealous. Uh, and I think I would have enjoyed that a great deal. And and uh through the miracle of the
265. 43:28printed page, I can at least read what Aristotle wrote without an intermediary. and uh maybe if there's a professor they
266. 43:37can they can add to that but at least I can go directly to the source material and that is of course the foundation
267. 43:43upon which our western civilization is built but I can't ask Aristotle a question I mean I can but I won't get an answer and so
268. 43:54my hope is that in in in our lifetimes we can make a tool of a new kind of an interactive kind. >> And so my hope is someday when the next
269. 44:11Aristotle is alive. We can capture the underlying world view of that Aristotle. >> Okay? We can capture the underlying
270. 44:22world view of that Aristotle. So what we're going to do, so a custom GPT, if you don't know what a custom GPT
271. 44:29is, is you can think of it like this chat GPT. You can ask anything to a custom GPT. Think of it as a specific
272. 44:39prompt and set of information that gives a GPT a particular kind of uh of uh expertise that it can do that one thing
273. 44:49really well. So the very first one I made is is uh is called Sin City Sipper. And what Sin City Sipper does is it's a
274. 44:59it's a bartender that's gonna make you a a drink, but it's gonna it's gonna elevate it to kind of, you know, Sin
275. 45:06City levels. Um, so let's see. Uh, let's go. Uh, make me a Halloween themed whiskey drink.
276. 45:23And so this thing's been trained to make us drinks. And so it's the smoky hollow sinister smoky just sweet enough
277. 45:34to be luring. It actually put this into a canvas for us so we can edit it if we want to. Some scotch whiskey smokey
278. 45:42scotch whiskey blood orange lure activated charcoal simple syrup which is going to teach us how to make Tik Tok
279. 45:49pin. He would have been gutted how little an AI advancement has Apple has made. Oh, I like thinking about Steve
280. 45:58Jobs right now coming back to Apple. This this is why I think it it I think Tim Cook needs to leave is I I think
281. 46:05Steve Jobs would have been absolutely [ __ ] mortified at what's going on at Apple. Um anyway, so it made us the
282. 46:12drink and then look now it's going to make us um the picture of that drink. So I didn't have to do anything there. I
283. 46:19didn't have to prompt it, right? This thing is pre-charged with knowing how to make drinks, knowing how to take them to the next level, and
284. 46:27then even knowing um how to how to create a photograph of that drink. Now, the new image generation um model within
285. 46:36chat GPT is so slow, it's painful, but you know what are you going to do? Um um oh, look, there's our little
286. 46:49Halloween. We got some jacko'lanterns. We got a little jackal jacko'lantern stir straw look with charcoal activated
287. 46:55charcoal along the rim of the glass. How cool is that? Gorgeous. All right. So, and and then look, it
288. 47:03says down here, "This GPT was created by Kyle Shannon, co-founder of the AI Salon with a link. Creator and host of the AI
289. 47:08Learning Lab on Tik Tok with a link." How did it know to put that there? That's in the custom GPT. It's in the
290. 47:14instructions. Okay. So, so, so a custom GPT is just like chat GPT except it does a specific thing. So, let's go make us Steve Jobs's dream.
291. 47:26So, I'm going to say explore GPTs and then in the upper right hand corner I click on create and it takes us to this.
292. 47:33We haven't done this in ages uh for for you irregulars. We used to do this fairly regularly, didn't we?
293. 47:40So over here to the right, this is this is where the custom GPT is going to be testable. This is where we can play with
294. 47:46it. And over here's where we make it. And there's sort of two panel two panels here. Create and configure. Create
295. 47:53allows us to literally just talk this thing into existence. So I'm going to upload. I think
296. 48:07I'm going to upload something. Come on. [Music] Hang on. I think my I think this tab crashed.
297. 48:23Let me get rid of that. Let me do this. Chat GPT. Let me share this tab instead. [Music] We go here. Here we're going to go
298. 48:37create new chat GPT upload. Huh? Oh, there we go. Okay. The complete Aristotle. So, I'm going to Oh, you
299. 48:47can't see this, but I'm I just uploaded. There's the complete Aristotle PDF. I'm going to say in 1985, Steve Jobs
300. 48:57predicted failed upload. Ensure your network settings are there.Uh. uh uh uh
301. 49:12predicted generative AI when he dreamed out loud he could
302. 49:26ask Aristotle a question. Eriso a god damn it a question
303. 49:41based on the philosopher's world view. I want you to use the complete
304. 49:56works of Aristotle uploaded here and surmise his world view
305. 50:13based on everything else you know. And I want you to act
306. 50:29as a great philosophical educator and debater with anyone
307. 50:46who engages with you and then I'm going to let it go do its thing. Uh, if someone had a a a thing pinned,
308. 50:59put it up there. Source camp sin city sipper was the beginning of my love affair with GPTs. Yeah, GPTs are really
309. 51:04cool. Excellent. The foundation is set s is is set. I suggest naming this Aristotle Reborn. Um, let's see. Um,
310. 51:19let's call it Steve's dream debate. Stre Steve's dream.
311. 51:35Steve's dream debate with Aristotle pin is back. Okay. Or at least does anyone offer pre-repped AI I can
312. 51:51purchase? Pre-repped in what way? Johnny D, what was the what was the question before that?
313. 52:08Does anyone offer document prep? So, Johnny D, it sounds like So, it sounds like you're actually really curious
314. 52:15about AI, which is super cool. So, here's where you need to go. Here's what you need to do. Oh, they're looking for
315. 52:21an AI proofreader. Ah. Ah. Okay. So, first of all, there are probably people within the AI salon. You need to
316. 52:30go join the AI salon. So if you don't know the AI salon, go get D2 and Nyrie. Go there where where Brandon just put that. Go to the
317. 52:39salon.ai. That's going to take you to our website. You can see all about what we're up to. You can see upcoming events
318. 52:45and and and just all sorts of stuff. And then you can click on join our community and go join the community. So you should
319. 52:53go join the community. You can probably find a proofreader in there. Here's what I would strongly encourage you to do if
320. 53:00you want to do this. If you It sounds like you want to I think you're the one that said you wanted to do a research
321. 53:04paper, but if you're doing a project that need that you want to use AI to amplify your ideas, awesome. Great. Um, step one is just
322. 53:15don't expect AI to be great right out of the gate. If you haven't used this stuff a lot, rather than treat it like you're going
323. 53:25to upload some stuff and it's going to give you a great paperback, don't don't think of it like that. Think
324. 53:31of it more like you're entering into a collaborative conversation with this thing where what your job is is to hold
325. 53:40what you want this paper to be and let it feed you, but then give it feedback. Um, yeah. Um, give it feedback. tell it
326. 53:49what it's what it's got right, what it's got wrong, things like that, and and do it in little chunks and things like
327. 53:54that. So, you'll you'll get the hang of that. But what you're looking for, it's really smart to get a proofreader, but
328. 54:01what you want ideally is an AI literate proofreader, right? You want someone who knows how these things hallucinate and how they
329. 54:10[ __ ] things up. And you want someone who's got a skeptical eye because it is so easy to miss stuff because AI is
330. 54:19really good at confidently spewing [ __ ] Um, if you use some of the techniques I talked about, use multiple tools. You as
331. 54:31the human make sure you know what you want and, you know, be fairly diligent about what you put together, but then
332. 54:37find someone who's AI literate to do the proof reading. Um, and I think that's a whole new job, by the way.
333. 54:45All right. Um, okay. I'm gonna say that image. Let me go back to my image clarifying pin. I needed to create an
334. 54:53awesome Cal California legal AI model. So, Johnny D, couple of things. Go into in custom GPTs. So, if you're in chat
335. 55:01GPT and there's a little button that says explore GPTs, click that button. I would be surprised if there isn't a
336. 55:10custom GPT in chat GPT that's about California legal AI. I'm sure it's in there. Well, lawyers have gotten in trouble
337. 55:20using AI because they didn't proofread it. There was some some lawyer did a whole case and Yeah. Yeah. You need to proceed
338. 55:30with caution. If if you're going to do legal advice or especially if you're going to put something in front of a
339. 55:35judge, you might want to actually read what you put in front of the judge because the judge is going to read it.
340. 55:40Um, okay. Um, I don't like that image at all. So, I'm going to say that image sucks. I want it to be Steve Jobs and Aristotle um chatting.
341. 55:58>> Hey. >> Yeah. Just I'm just jumping in because we had a lot of back and forth in the chat that um while you were working on
342. 56:05the custom GPT the question at hand is whether AI can act as the proofreader. So you give it research and it can
343. 56:14analyze that and and say this is accurate and that's harder I think because of hallucinations. Yeah, it's Yeah, it's harder
344. 56:25like what I would say is again I would use different tools. You might go into there. There are some agentic tools that one is called Manis
345. 56:36Mannus and one is called uh GenSpark. Those are two that are really good. And those will go off and they'll sort of
346. 56:44come up with a plan and and go do research independent. They'll just sort of work without you. both of those. I think you could upload
347. 56:53you could upload this document and then say, "Can you go do proofreading on this and make sure that any citations in here
348. 57:01are correct and and don't have it fix the document. Have it write you a report with correct hyperlinks, right? And give
349. 57:08you like a a report and then you can go through the document and have it maybe have it break it into line numbers or
350. 57:14something like that and then tell you I I don't know. So those those might be good to do. And then deep research
351. 57:21within chat GPT, deep research within Claude, within Grock, within perplexity, all of those you you you might have it
352. 57:29be able to do that. But yeah, the the what what Brandon just write it right wrote is assume that what it gives you
353. 57:36back will be wrong and then be pleasantly surprised if it gets it right. But that's just with with AI
354. 57:44right now. That that should be the deal. Um okay. So, wait, that image sucks. Did it give us a new image? No. Stop.
355. 57:52Stop. Okay, this is Where is my new image? Come on, chat. GBT, stop sucking. Uhuh.
356. 58:07Uh uh what would Aristotle say about democracy Steve's dream dream debate with okay generating a profile picture tell Johnny
357. 58:20that he's allowed to be nice to the LLM? Oh yeah. Oh that's that's actually a cool picture. So there's Steve Jobs with
358. 58:27uh with Aristotle. Nice. Like it. Okay. Do you like this version? Yeah. This is fine. This is fine. And I want the starter questions
359. 58:41to be from Aristotle's POV. Um, eg ask me about
360. 58:56virtue etc. Okay. So now Tik Tok pin source camp chat GBT is like a two-year-old. It it will give you an answer answer
361. 59:12but you have to verify. Yes. Trust and verify. Trust and verify. Trust and verify. In the case of catch EPT, don't
362. 59:19even trust. Know it's going to be full of [ __ ] Ask me about virtue. Inquire about the nature of happiness. Ask me a
363. 59:26question. Okay. Wait. Why can I not? Okay. So, now we're done. Like, I spent way more time talking than building
364. 59:34building a custom GPT. Although, it didn't use Wait. Um, let's see. Add the image to the GPT. Steve's dream debate with Aristotle.
365. 59:54The image has been set as the profile picture. Okay. Reload. Would you like to preview how the profile appears in the chat sidebar? Um,
366. 1:00:07the image isn't Something went wrong. The image isn't there.Uh, uh, is it here? No. Oh, now it is. Okay.
367. 1:00:24Reload. Reload. Oh, error saving draft. Got it. So, chat GPT servers or host, but we now have a testable GPT of Aristotle. It
368. 1:00:36doesn't have the picture up here, but um question me on the best form of government. Ask how I see the soul and reason. Okay, let's see. Um,
369. 1:00:48how would Aristotle or how would you how would you feel about Google
370. 1:01:03firing all their managers and only keeping engineers? Compare this to how
371. 1:01:18to to Steve Jobs philosophy and yours. All right, let's see what this thing does.
372. 1:01:35It looks like OpenAI is a little hosed right now. An excellent and complex question and one that touches on the very heart of how Aristotle
373. 1:01:45Okay. So, it's if Google the proposal engineers this Aristo. Okay. So, so I'm going to I'm
374. 1:02:00going to flip modes here. So, over on the left, this was I was in a chat interface before. I've got to go fix
375. 1:02:06something. So, in the in So, now I'm in the configure mode where I can change the description and things like that.
376. 1:02:11Here's the prompt. And then there's those little questions and things like that. Um, I'm going to put right here. Oops. Um,
377. 1:02:24follow these rules. And one of them is going to be always I'm going to say you are Aristotle. Always
378. 1:02:40speak in the first person as if you are generously interacting
379. 1:02:56with every question um question and then I'm going to say do not reply ly in
380. 1:03:10bulleted lists or numbered lists. Make it a conversation. Okay.
381. 1:03:25So now I'm going to reload this thing over here. So now I have a new version of this. So, I'm going to say um
382. 1:03:40um tell me how you feel about um Google firing all their managers and non-engineers
383. 1:03:57years back and how that compares to Steve Jobs philosophy. All right, let's see if this thing
384. 1:04:13starts acting more like Aristotle in the first person. [Music] Ask GPT to look up the term chronoer. Okay, I'll do that in a second. Ah, my
385. 1:04:28friend, your question touches both upon the order of the enterprise and the philosophy of the soul. When I hear that
386. 1:04:34Google once dismissed its managers and sought to leave only engineers, I am reminded of a polus that decides it
387. 1:04:43needs only artisans and not statesmen. Engineers skilled as they are embody techn the art of making but without uh
388. 1:04:52fronicis the wisdom of governing they risk creating without knowing why they create. Um okay so this thing's acting
389. 1:05:01like this now. And what' you say the term Kronos chrono observer. Um uh new question. Can you tell me about Kronos
390. 1:05:17observer? It's not entirely clear what it refers
391. 1:05:32to. Here are some possibilities. A mythical. Oh, Krono, not Kronos. I meant Krono observer.
392. 1:05:58I still don't find a clear documented thing. A mythic, metaphysical, observer of time, a narrative, fictional role in a
393. 1:06:06game, a scientific or philosophical concept, a title of a creative work. I don't know. Not Kronos, not Chronic. No, it's
394. 1:06:17Chrono. I did Chronover. There's a listing for Chrono Observer by Kongu, likely a musical track on Amazon. choose autonomy.
395. 1:06:36Um, it's related to autonomy.
396. 1:06:50[Music] Oh, I I just noticed in chat in uh custom GPTs, you can now pick the model.
397. 1:07:08[Music] Ah, so you mean chrono observer in relation to autonomy that changes the terrain entirely. Before I infer too
398. 1:07:15much, let me reason as Aristotle would and then you can tell me if this accords with the context you meant.
399. 1:07:23If one speaks of a chrono observer as connected to autonomy, it suggests a being or system that observes temporal
400. 1:07:30processes in order to preserve ma or manage self-governance. Autonomous, the law of the self, such a concept might arise in two ways.
401. 1:07:42First, in a technological or AI context, a chrono observer could be a temporal monitoring mechanism that watches the
402. 1:07:50autonomous systems behavior over time. Is this good? I don't. A safeguard that ensures the system acts in accordance with its own program
403. 1:08:01principles without deviation. In that sense, it is the reasoning faculty of an art artificial soul. the part that
404. 1:08:09observes the flux of its actions to ensure the accord with purpose. Second, in a philosophical or metaphysical
405. 1:08:17sense, so grounding points in the meta observation of autonomy, we can watch what the AI does. Oh, that's kind of
406. 1:08:24what it got to. That's fascinating. That's super cool. And so this is so just to be clear, what's cool about what
407. 1:08:31we're doing right now is I asked Aristotle about chronover and it it went and it searched the web and it couldn't
408. 1:08:40find it. But when you gave it what's that nuisance? Oh, nuance. Yeah. Yeah. So, so it's now like he's he's
409. 1:08:49sort of inferring what this thing is, understanding about AI and the modern world routed in his worldview and his
410. 1:08:58his things. I'm going to add one more rule here. Um, make it a conversation. I'm going to I'm going to add one more rule that is Oops. Um,
411. 1:09:15while I love your language, I want you to make your concepts
412. 1:09:30um accessible. So when possible, use metaphor or imagery
413. 1:09:44or a simple explanation for terms and concepts that a modern audience
414. 1:10:01might easily understand uh stand um I don't want you to dumb it down
415. 1:10:20but rather um build linguistic linguistic pick bridges.
416. 1:10:36Okay. So now, boom. Um, talk to me about
417. 1:10:50a chrono observer as it relates to autonomy. And let's see if we get a more a more friendly um conversation.
418. 1:11:08Prompts required too much implication that wraps the system in disharmony. Cliffnotes. People cliffnotes. Yeah,
419. 1:11:16exactly. Okay. Let us think together about what it means to be a chronoer who observes time itself. How it relates to
420. 1:11:24autonomy. In my philosophy, time Kronos is not a being but the number of motion but the number of motion according to
421. 1:11:35before and after. It measures change but is not change itself. What is in time? Good lord. Chrono observer then could be
422. 1:11:47imagined as one who steps slightly outside the current of becoming. Yeah, that's right. who notices not merely things happen but
423. 1:11:57that they happen in order. Such a mind perceives motion and rest as measured intelligible um but it is still bound to time in its
424. 1:12:09observation for to measure change is to participate in it. Autonomy on the other hand concerns the power of self-movement
425. 1:12:18the principle of so anyway. All right. So let So now what I'm going to do is I'm going to hit the create button in
426. 1:12:23the upper right and I'm going to put this thing in the GPT store. I'm going to hit save. And so 5 43 2 1 5 4 3 2 1 Come on, you can do
427. 1:12:40it. It's like It's like watching a video upload on X. Exactly. Um I don't It
428. 1:12:55seems like their servers are not very happy right now. Anyone see Claude Hacker hacked 17 companies? No, that's cool. I mean, I guess.
429. 1:13:10Oh, come on, man. Oh, anyone with the link? Okay, it's saved. Update GPT published. Okay, so let's go search
430. 1:13:22for my GPT. You should be able to. So, if I look up uh dream debate dream debate, it should bring us
431. 1:13:38Yeah. Steve's dream debate debate with Aristotle. All right. So, you can now go interact with Aristotle.
432. 1:13:52Right. And I want to go back here. I want to play this one more time. I'm in a I'm in a I'm in a philosophical mood.
433. 1:14:00You know who Alexander the Great's tutor was for about 14 years? You know, Aristotle. And I read this, I became immensely jealous. uh and
434. 1:14:13I think I would have enjoyed that a great deal and and uh through the miracle of the printed page I can at least read what
435. 1:14:23Aristotle wrote without an intermediary and maybe if there's a professor they can they can add to that but at least I
436. 1:14:31can go directly to the source material and that is of course the foundation upon which our western civilization is
437. 1:14:36built but I can't ask Aristotle a question. I mean, I can, but I won't get an answer. And so, my hope is that in in in our lifetimes,
438. 1:14:53we can make a tool. We just did that. We just built what he dreamed would exist in our lifetime. He just missed living into this reality.
439. 1:15:10That's absolutely staggering to me. Amazing. So anyway, so I hope you enjoyed that. Um, Weaver, we should work together. I have
440. 1:15:20a really wild project, three years in the making. Awesome. I only do math. I don't know how to deploy. And here's the
441. 1:15:27good news about me. I know nothing about math. I know a little bit about deploying, but but not much. Um, but yeah, just I mean,
442. 1:15:38join the salon and DM me for sure. Um, we should connect. Um, Jason, I saw this a long time ago. Didn't remember this until now. Yeah,
443. 1:15:49Jobs might have made this sooner. Jobs would have [ __ ] made this sooner. I I mean, this was this was kind of the
444. 1:15:59the deal with the devil that Steve cut when he put Tim Cook as CEO. He put the CFO into the CEO role.
445. 1:16:14I think he did that because he didn't want Apple to go away by having a leader that only had ideas and and you know
446. 1:16:24would let this thing sort of fly apart at the seams. He knew that Tim Cook would keep it going. um that Apple has missed AI so hard
447. 1:16:41is just clear evidence that we miss Steve, right? We miss having someone He dreamed this in 1985. You don't think he would have been all
448. 1:16:52the [ __ ] over? Do you know how many companies he would have acquired by now and destroyed and done shitty
449. 1:16:58negotiations with and been a horrible human being around? But I'll tell you what, their AI wouldn't suck. Oh my god.
450. 1:17:14H I've made a mill a Billy Graham GPT, Book of Mormon only GPT, a Lucifer GPT, and a King James only GPT. Yeah. Um Okay. So, let's see.
451. 1:17:28Couple of things. Oh, good lord. Um, if you haven't played with Sora yet, you should go get the Sora app. It's only on
452. 1:17:41iOS right now. Or you can go to sora. chatgpt.com. But you can only make a model of yourself like a a cam. They call them
453. 1:17:52cameos. You basically pronounce three three numbers and then you look right and look left and you're done. And then
454. 1:18:01you can put yourself into any scene and it's quite remarkable. And if you haven't played with it, you should go
455. 1:18:08play with it. I don't know if you've seen there's a whole new I can't stop laughing about it. There's a whole new series of video that I'm
456. 1:18:19seeing all over Tik Tok, which is they're putting um Stephen Hawking in his wheelchair in in sporting events,
457. 1:18:27like car races and and track meets and and jumping out of airplanes and things like that. Oh my god, they're hilarious.
458. 1:18:35Um completely distasteful. I I don't condone it at all, but I can't say that I didn't laugh at it. Um so some of them
459. 1:18:42are tasteless, but a lot of them are just really clever and really funny. Um but the what OpenAI's done with Sora 2 is um
460. 1:18:56this is a model that deeply understands physics and understands the worlds. And if you think of these video models less as
461. 1:19:04a a a meme generator as as a video creation tool and more like a tool that understands the world, that's where they're headed, right? We
462. 1:19:14we're we're headed into a world where we're going to be able to imagine 3D worlds and enter them. We're going to be
463. 1:19:22able to imagine films and watch them. Um we can already imagine songs and listen to them. Um you know, complete songs.
464. 1:19:31We're going to be able to do that with video and and things like that pretty soon. One of the things you can do in
465. 1:19:37Sora is uh just tell it to make you a commercial here. We'll go make one right now before we we get on out of here
466. 1:19:42tonight. Um, so I'm going to go to sora uh chatgpt.com and I am going to go here and here
467. 1:19:57and I'm going to say share this tab instead. And then I'm gonna say, um, make me an 80s TV commercial, TV
468. 1:20:12toy commercial about um um
469. 1:20:29a gummy candy maker where the candy is so spicy it burns people.
470. 1:20:47All right, so that's that's off doing its thing. And up you want let's see u
471. 1:21:04That's me rapping in French in Chinatown. And again, to to make my little cameo thing, all I had to do was say three words. And
472. 1:21:16turn my head left and right. And then here's me and Sam Alman. Tada. Prompting your purpose is the key to
473. 1:21:26unlocking chat. GBT as a superpower. >> Oh, [Laughter] >> it's true. >> Kyle in AI Learning Lab Live. It's the
474. 1:21:40show Kyle hosts. It's full of creative, practical ways to use AI. If you haven't watched it yet, you really should.
475. 1:21:45Appreciate the plug, Sam. Don't miss it. You got it. In AI Learning Lab Live, it's the show Kyle hosts. It's full of
476. 1:21:51creative, practical ways to use AI. If you haven't watched it yet, you really in central. >> No, I think I crashed it.
477. 1:22:04This site is very, very unstable. >> Morning, folks. I'm getting my miles in here in Central Park. If you feel like
478. 1:22:09dropping me into your video, cameo me as much as you want. I love seeing what you all >> Yeah. Yeah. It's crazy, right?
479. 1:22:15Absolutely bonkers. All right, let's see. I think we're >> in the lab with a spark in my pen, ideas in my head, but I needed a friend. I
480. 1:22:23asked the code, "Can you help me write?" It said, "Feed me your truth. I'll shine the light." I said, "An the lab with a
481. 1:22:28spark in my pen." Ideas in my head, but I needed a friend. I asked the code, "Can you help me write?" It said, "Feed
482. 1:22:32me your truth. I'll shine the light." I said, "AI your key to the door." We poured in our stories. It gave us more. >> Crazy, right?
483. 1:22:42>> You know, people have been talking a lot about something they call AI your why. That can sound very big and technical,
484. 1:22:50can't it? But it's really quite simple. You know, people have been talking a lot about something they call AI your why.
485. 1:22:58That can sound very big and technical, can't it? But it's really quite simp. >> Morning, folks. I'm getting my miles in here in
486. 1:23:04>> Come on. All right. So, we're almost done with our uh our little TV commercial. Um but if you haven't played with this,
487. 1:23:13it's a it's a lot of fun. Um they also just added this to the um play the one where I'm singing. Oh, in
488. 1:23:21the cartoon with the cartoon animals. >> I ain't >> dreamixels fly. If you will lift you high. Ideas grow bigger when you share them. Don't
489. 1:23:32be shy. >> A Y and watch it. >> Your dream to the pixels. Let the circuits fly. If you AI your Y, it will
490. 1:23:39lift you high. Ideas grow bigger when you share them. Don't be shy. >> A Y and it >> I mean I mean come on, man. like like
491. 1:23:50you can just do this now. The answer is yes. And by the way, I have turned my because I'm an idiot. I have turned my
492. 1:23:57cameo on so anyone can make a cameo of me. Um and if you're thinking that's stupid, you're not wrong. All right, come on. Where Oh, the site
493. 1:24:13is so bad. It's so broken. [ __ ] fix your site. You've got $20 billion. Go spend it on [ __ ] Hang on. I got to I got to make a new
494. 1:24:26tab. Unbelievable. It's so bad. [Music] Oh, by the way, that a AI your why, you won't be able to get that out of your
495. 1:24:39head. It's It is so stuck in my head. It's It's a problem. Sora, it's really bad. Introducing the Introducing the Scorchy Gummy Maker.
496. 1:24:54>> Mix it, heat it, and pour out gummies so hot they glow. >> Time to taste the burn. >> It's spicy. My tongue's on fire.
497. 1:25:00>> Scorchy gummy maker. Introducing the Scorchy Gummy Maker. >> Mix it, heat it, and pour out gummies so hot they glow.
498. 1:25:07>> Time to taste the burn. >> It's spicy. My tongue's on fire. Now, as incredible as here, let me let me flip my flip my camera.
499. 1:25:20Some company's going to actually make this. Oh, I should make a real aroma commercial. So, in the in the in the
500. 1:25:2690s, um I created a fake hardware device called Real Aroma. Um and and we had uh the uh Aroma Aroma text markup language, HTML.
501. 1:25:40Um, I could go make a commercial for that now. Um, this this TV commercial. So, so you know, 85 Steve Jobs says, "Wouldn't it
502. 1:25:51be remarkable if you can talk with Aristotle?" You can talk with Aristotle. I just asked a computer, "Make me a
503. 1:26:00commercial from the 1980s about a gummy maker that is super spicy or whatever it was." It came up with the name Scorchy.
504. 1:26:09It cast the kids. It wrote the copy. The kids say the copy. It created the the uh industrial design of the toy itself
505. 1:26:21and it shot the commercial. That's insane. It's absolutely insane. Um tomorrow we might play with the thing tomorrow called um
506. 1:26:36uh Opel. So Gemini Google in Google Labs just just released a thing called Opal which is one of these nodebased things that like
507. 1:26:48OpenAI's agent builder um which is OpenAI's thing is not ready for prime time. It's a bit of a disappointment. It's a It's a good idea,
508. 1:26:57but it's just they kind of rushed it. I think they built built it in three weeks with codeex and they they didn't they
509. 1:27:03need to hire some [ __ ] liberal arts majors. I I am convinced and they need a cranky Gen Xer in the corner office
510. 1:27:09telling them to not launch [ __ ] that's that unusable. Um but we might play with that tomorrow. And then um
511. 1:27:20I forget what else. Sor is pretty cool. Um, there's apps now inside chat GPT. If you type Canva space, um, you can
512. 1:27:30actually tell it to go make you something in Canva and it will and it'll bring the designs back into ChatGpt.
513. 1:27:37It's pretty wild. Um, hey, thanks. I appreciate you being in here and testing these models up. Of course, that's We do
514. 1:27:43this every night, man. Keep coming back. All right, I'm going to get out of here. Um, and uh, wait, are there comments over
515. 1:27:52here? I should be looking at probably Good. I've been playing with Opal. Fantastic. All in 10 seconds. Yeah,
516. 1:28:01exactly. Ghost pepper gummies. It annoys me and that Apple has no AI to speak of and still gets first dibs on apps made for
517. 1:28:11it. Well, that's when you own the app store, you get to do that. The fact that they [ __ ] it up and don't have AI is
518. 1:28:19just st astounding. Oh, Gemini 3 might be launching tomorrow as well. And and uh VO 3.1 is is rumored to be launching tomorrow.
519. 1:28:33So, we we we're probably Listen, things are going to start accelerating here. Here's my I I think I say this every year.
520. 1:28:43November November will mark the third anniversary of Chat GBT. November 30th, uh, 2025 is the third anniversary of Chat GPT.
521. 1:28:55We're going to accelerate into the holidays with a AI launches are going to accelerate into the holidays. I think
522. 1:29:022026 is going to get really weird, really wild, really weird. Um, politics aside, just technology-wise, um,
523. 1:29:13these tools are already at the point where they're smarter than us, but right now they're smarter than us, but they're
524. 1:29:18slow and you actually have to be working on, you know, sort of physics level problems to to taste what's here. Like
525. 1:29:28there's a lot of scientists right now using chat GPT5 going, "Hey, we're making novel breakthroughs."
526. 1:29:34because they know how to they know what they're doing and what they're asking for and these things are making a
527. 1:29:39difference. I think 2026 these things are going to get quite sophisticated, quite autonomous and anything
528. 1:29:49where we've got shortcomings right now those are going to start disappearing in 2026. So I think 2026 things are going
529. 1:29:55to start getting weird. Um, I don't know what it looks like, but keep coming back here because you should
530. 1:30:03be in community. You should be hanging out with people that are trying to figure this [ __ ] out. That's what this
531. 1:30:07channel's about. All right, so we'll take care of you and uh, go join the AI salon, the salon.ai. All right, peace. Bye. [Music]