7/23/2025 - Taming Hallucinations in Large Language Models

Transcript

1. 0:00You ready, Chippy? [Music]
2. 0:18[Music] 10,000 words swimming around my head. 10
3. 0:32million more in books written beneath my bed. And I wrote or read them all when searching in the swamps. Still can't
4. 0:45find how to hold my hands. And I know you need me in the next room over. I'm stuck in here all paral.
5. 1:05I kept myself in ruts. Too much time spending mirrors framed in yellow walls. Ain't it like most people? I'm no
6. 1:19different. Like to talk on things we don't know about. [Music] Well, like most people, I'm no different. like to duck on things we
7. 1:32don't know about. [Music] [Music]
8. 1:49Do [Music] no.
9. 2:15No. [Music]
10. 2:31Good day. Good people. It's a little moist in here. It's a little moist in here. Do you feel it? Can you feel it? I'm a little [Laughter]
11. 2:39[Music] We're gonna talk about artificial intelligence tonight. I don't know if you've heard this term.
12. 2:54There's intelligence. We've got that. And then they figured out a way to wire some computers together in such a way
13. 3:03that uh it's like that but not with machines. So they they call it machine learning. I I don't think machines learn.
14. 3:17You program machines. We're the humans here. We've got the intelligence. You tell the machine what you want it to do.
15. 3:27The machine doesn't tell us what it wants to do. So we're going to we're going to explore that tonight. We're going to figure it out.
16. 3:36We're going to figure out the mathematics behind it. We're going to optimize some tensor processors. We've got we've got engineers and
17. 3:45mathematicians from Google that hang out here. We're going to we're going to probably order a couple of H100 GPUs from Invidia.
18. 4:00Is it inv video or is it N vid? What? Why do they spell it like that? And do we want an A100 or an H100?
19. 4:14Because I think an H100 is kind of the cheap shitty thing. And then we'll get an A100. Anyway, we're going to use all
20. 4:21of the different libraries from Nvidia. Thank you very much there, Forever Hooked. Very kind of you. And then we're going to we're going to
21. 4:34learn programming. We're going to learn. We're gonna learn uh there's this thing there's a paper out called uh attention is all you need
22. 4:44and it's about insecure theatricals. kids that grew up just wanting attention, wanting attention, wanting
23. 4:51attention. And then they went to theater school and then out of theater school they moved to New York City and they
24. 4:58started and and ran a theater company and after four and a half years they got burned out and so they they decided to write
25. 5:07themselves an acting career. So they wrote seven screenplays in two year. No, that's that's not what that paper was about.
26. 5:16that that was that's my life anyway. But attention is all you need is this paper that introduces this thing called the
27. 5:24transformer. So we're going to we're going to learn what a transformer is and then we're going to go collect all of the information
28. 5:31that humanity has ever put on the internet and we're just going to gather it. We're going to we're going to stick
29. 5:37it in a box and and we're going to put it in a transformer and it's going to embed it. And I from what I understand,
30. 5:44we're going to we're going to take all of those documents and and they're going to they're going to be shattered into
31. 5:50trillions and trillions of fragments. Yes, Brandon. >> Hey, before we all sounds amazing, but before we do all of that, could we uh
32. 6:00answer Ginger Jay's question on Tik Tok? >> Oh, yes. Hi, Ginger J. What's happening? Did you have an actual question about actual AI?
33. 6:12Oh, wait. I lost the pin. Someone put the pin back up there, please. Okay. Can you comment? I use chat GPT.
34. 6:21How do I get it to stop lying and make up answers that sound good? Um, okay. So, good question. and all that horse hockey I was just talking
35. 6:36about before that that was in the neighborhood of what what actually happened with creating this stuff. Okay,
36. 6:41so there's a couple of things. Um, one of the things that we all have to grapple with right now is that the way
37. 6:58large language models chat GPT cla um, are we good? Yeah, I think we're good. Oh, wait. I got to move this up here. Hang on a second.
38. 7:12The way these things work. Hang on. I got to turn on my air, too. It's a little It's a little hot. It's a little hot.
39. 7:21The way these things work is not the way computers have ever worked in history. Um, they used to be called, I think,
40. 7:31probabilistic is the is the or deterministic. Computers used to be deterministic. You would program them,
41. 7:37you would put logic in them, and they would always behave the same way every single time. large language models don't do that.
42. 7:45And the nature of of how they were built um means that for one thing they can they can be creative which one of the things that
43. 8:01you'll hear about that that people will say about AI is well it's just a it's just a probability calculator and it can
44. 8:08never be creative and it can never come up with an original thought. I actually don't agree with that. I think it I
45. 8:12think it can I think the way they work is very similar to the way our brains work. So there are some cases when
46. 8:18hallucinations is what they're is what what you're referring to is called hallucinations. There are some cases
47. 8:24where hallucinations can actually be beneficial, right? And that would be if you're doing creative development. You
48. 8:29want some story ideas and it will be creative and it won't just give you the same answer every time. It'll give you
49. 8:35something different every time. So there are some cases where that's a good thing. There are other cases if say
50. 8:41you're trying to research something and you want facts and and the large language model what I call it is mansplaining as a
51. 8:52service it mansplains to it gives you an answer confidently even if it doesn't know what it's talking about right like
52. 8:58every good man in your life um and that could be really frustrating now there are things you can do to mitigate
53. 9:05it there are things that you can do that that that make it better. So, let me let me talk through a couple of them. Um,
54. 9:15and and I'll talk this came from Andre Karpathy, who was the one of the co-founders and the chief scientist at
55. 9:20OpenAI. Um, and he was describing someone had asked him the question about about why large language models hallucinate. And and he used a
56. 9:30mathematical example and I'm not going to give you the math. I have an acting degree. I don't I don't know about math.
57. 9:36But he was talking about um if you ask about a particular mathematical conjecture, right, a particular mathematical challenge
58. 9:48and you just ask chat GPT about that challenge, it will basically go look at everything it's been trained on, which
59. 9:55is essentially everything that's ever been on the internet, and it will sort of cobble together an answer for you,
60. 10:00and it will answer you. And very often it'll be wrong. and he was he was describing why that might be. So let's
61. 10:07say you've got this mathematical formula and you just ask about it and he said well in chat GPT are not only the the
62. 10:16professor that came up with that conjecture that formula and the answer to it but every student that he ever
63. 10:24taught and every student that ever tried to attempt it every student that ever read about that conjecture and and so
64. 10:31not only is the correct answer in the training set so is every other wrong answer in the training set. And so the
65. 10:39way I like to think of what they call the latent space, the latent spa, think of the latent space is you take all of
66. 10:45the knowledge of humanity, you fracture it into like tiny little shards of glass and they all get sort of clustered into
67. 10:52thousand dimensional mathematical space and it's just this massive ball of fragments of knowledge and when you put
68. 10:58in a prompt, it extracts out of that these little fragments and sort of stitches them back together for you.
69. 11:04It's actually quite remarkable the way it works. If you think about it like the Grand Canyon, if you take all these shards of
70. 11:11glass and you dump them into the Grand Canyon, if you just ask it about that that mathematical formula, it'll look at
71. 11:18all the shards of glass in all of the Grand Canyon. But if you narrow the latent space by providing an additional
72. 11:26context, I want you to act like a a a a mathematics a polymath, you know, professor that specializes in this branch of mathematics. And I'm
73. 11:37particularly interested in this particular formula. And I really only want to hear from, you know, the most
74. 11:44experts in the world about this formula. What it's what you're doing is every time you add context like that, you're
75. 11:50going into an increasingly small subc canyon of the Grand Canyon and you're cutting out all that other noise. So,
76. 11:58one of the best ways to make chat GPT less bullshitty is to provide it more context. So, if you just say, "Write me an article about
77. 12:08X," it's going to write you a shitty article. If you say, "Hey, I'm an expert in zucchini gardening, and I live in
78. 12:18Denver, Colorado, and I've got a raised bed, and we're in our third year of gardening, and I put in a zucchini
79. 12:25plant, and one of them I fertilized, and one of them I didn't, and for whatever reason, the plant that I didn't
80. 12:32fertilize is doing better. And I want you to write me a LinkedIn article about zucchini gardening in raised beds.
81. 12:41all of a sudden you've narrowed sort of the world that it's going to draw from into this into this much smaller place.
82. 12:49Does that make sense? Um so that's a thing. Now you can also use different models. Let me share I'm sharing my screen. I'm
83. 13:00going to share it in a different way here. Follow a second. [Music]
84. 13:14Okay, there's another way you can do this. So if you haven't played with switching models in chat GPT, you have these different
85. 13:24things called models and they behave very very differently and they are very very very very very very confusing
86. 13:34because a bunch of engineers named them and they did it really poorly. So you have this thing called 40 then you have a thing called 03
87. 13:44Tik Tok pin. Here's your article. Overfertilized. That's funny. I didn't I couldn't see the bottom of it. There's Tik Tok has
88. 13:59this weird interface where whenever someone pins a comment, I can read the first two lines and if I want to expand
89. 14:04it, I always it always does other [ __ ] I don't actually expand it about once every five times. Anyway, you got this
90. 14:12thing called 40. This is the base model that most people use. That's the default. I would argue that most people
91. 14:18that use chat GPT never never touch these models. Um 03 is not lower than 40. It's actually a completely different
92. 14:33kind of model and it's called a reasoning model. See it says uses advanced reasoning and then and then you've also got 04 mini and 04
93. 14:43mini high. So, so 03 is the is the kind of MacDaddy reasoning model and then 04 mini and 04 mini high are smaller versions of the next
94. 14:55evolution of 03, but they're they're kind of the smaller previews of of what's coming with 04 Pro or whatever the the big 04 is,
95. 15:06but they're not the same at all. And and there's no way you would know that. There's no way you would understand that. But
96. 15:15what 03 does when it talks about reasoning, what happens with 40 is you ask it a question, it gives you an
97. 15:21answer. So even if you do all of the context providing that I gave you, it can still get it wrong. What 03 does is it actually thinks about
98. 15:32it and talks to itself before it answers you. So if you ask it a question about raised bed zucchini gardening, it will
99. 15:40say, "Oh, the user wants to, you know, talk about zucchini gardening, maybe I should go out and I should do some
100. 15:46research. Let me go out and and research some some websites about gardening and that particular thing and understanding
101. 15:54about different species of zucchini and and it will just talk to itself and go do research on the web and think about
102. 16:01it and analyze stuff. Sometimes it'll even use what are called tools where it will write Python applications and
103. 16:09generate spreadsheets and do math and all just all sorts of weird [ __ ] So, so that's that's a next way to reduce
104. 16:17hallucinations is use a reasoning model. They can still hallucinate. Don't be clear of that. And then the final thing
105. 16:23that you can do is you can um there's an option in chat GPT called deep research. So if you combine 03 which is this
106. 16:33reasoning engine with deep research it will really dig deep and research deep and check its work and things like that
107. 16:40and it gets better and better and better. Even that can still hallucinate. So you're not going to completely
108. 16:46eliminate hallucinations until these models probably hit another generation or two or three. Um and even then maybe
109. 16:55not. Um just by the nature of how they are. Okay. If I've trained in 40, does any of that transfer to the other mo
110. 17:03models? Um, what do you mean if you've trained like if you've put in custom instructions and and you're doing
111. 17:10projects and things like that? Yes. The memory. So, couple of things are happening in chat GPT. If you go to your settings,
112. 17:20go to go to a section called personalization and you'll see a memory section. Um, and you'll also see custom instructions.
113. 17:30So, if you go into custom instructions, um, custom instructions are you're telling chat GPT what you want it to do.
114. 17:39One of the things you could say in custom instructions, I don't know if it would work, but it might, is say, I want
115. 17:44you to do everything you can to to reduce hallucinations. So, before you answer me, check your work. Now, I don't
116. 17:49think that'll work with 40, but that might work with 03. Um, so custom instructions can be good. Um, but you've
117. 17:58also got reference saved memories and reference chat history. So as of about three or four months ago, chat GPT has
118. 18:07access to your past, I think it's year of chats. And so that applies to all models. So you can be in a 40 conversation and you
119. 18:18can flip over to the 03 reasoning engine and just keep asking in the same conversation. There are some things that
120. 18:25that mess it up like if you generate images within it that used to mess it up. Things like that. Memories light
121. 18:33the corn memories light the corners of my mind. Everybody know memories like the corners of my mind. Oh, and by the way, speaking of which,
122. 18:49let me share one other thing here, people. Let me share another thing. Look at that. You see that right there?
123. 18:59Next week, starting on Monday, um I'm doing a fiveday back to basics crash course where um kind of all the silly talk I did at the
124. 19:11beginning about how large language models work. Um, on Monday we're going to talk about AI, why it's a big deal,
125. 19:19how it compares to the early days of the worldwide web, how large language models work, how diffusion models, the image
126. 19:26generation models work. Um, and so that's Monday. Tuesday is going to be a deep dive into chat GPT. So all the
127. 19:33stuff I'm talking about here, we'll do hands-on like a hands-on hour with chat GPT and unpack a bunch of stuff. Um,
128. 19:40Wednesday is going to be a creative day where we're going to play with image generation models, video, music, audio,
129. 19:47things like that. Um, Thursday is where we're going to bring it all together. We'll do some some business use cases,
130. 19:53figuring out how do you determine which model to use when. Um, and then Friday will be an open-ended Q&A. All right. So that's next
131. 20:03next week here on Tik Tok from 8 to 8:30 Mountain time will be a pre-show where we'll just hang out. And
132. 20:13then starting at 8 8:30 we're going to turn off Tik Tok and just go completely to YouTube. And let me see. I've got
133. 20:24I think I have these slides, Brandon. Let me grab them. Oops.
134. 20:46[Music] So, here is the YouTube channel. Wait. Oh, add the stage. There we go.
135. 21:02Oh, that's producer Brandon. Wait, why did I not? Hang on, people. I guess I just didn't share. There we go.
136. 21:15So, there's the uh there's the YouTube channel. If you're if you're just on TikTok right now, learning lab-ai is the
137. 21:22YouTube channel. So 8:30 next week, 8:30 Monday, Tuesday, Wednesday, Thursday, Friday, um is where we'll be streaming
138. 21:30live. Okay? So you can zap that QR code if you want to do that or just type inlab-ai on YouTube and subscribe to the channel. Okay. Okay.
139. 21:45Beautiful. Which which other slide? This other slide. Ah, and then the other thing that you should do if you haven't yet is you
140. 21:57should join the AI salon. So if you go to community.thesalon.ai, that's going to take you to our community site. Um, we have a special
141. 22:08channel in there just for the AI learning lab called the Irregulars. Um, but there's all sorts of cool stuff
142. 22:15in there. um and all sorts of amazing people and we talk about uh this journey of AI literacy we call it where you play
143. 22:25first mindfully create and generously lead. Um, and if you're just getting started with AI and trying to figure
144. 22:32this out and you're asking questions like, "Hey, why the hell does this does this thing lie so much? Um, everybody's
145. 22:40dealing with it. Everybody's dealing with the jank, I call it, right? Embrace the jank. These tools are remarkable and
146. 22:48weird. They they computers just have not ever behaved as unpredictably as these things behave. Um, they've also never
147. 22:57been able to do things quite as remarkable as what these things can do. They they they are, you know, they are
148. 23:05just on the verge of magic. They are they are all math, but it's it's mathematical as we call it here. All right. [Music]
149. 23:23So, go join the salon and and you'll see when you get there there's kind of a sevenstage welcome process. The second
150. 23:29one is introduce yourself. So, go introduce yourself. You should go do that. All right, let me change my sharing one more time
151. 23:41and then we're going to do a little movie night. We are gonna do a little movie night and I think it's going to be a lot of fun. I
152. 23:50think it's going to be a lot of fun unless it's boring. then we'll do something else. All right. So, I'm going over to the AI sell
153. 23:59and I am jumping down to the irregulars channel. So, down the left hand side here, right under clubs and hubs, the
154. 24:05first one is the irregulars and we've got a little image from Steo. It's a little moist in here tonight. Steve capturing
155. 24:19the essence of the uh of the of the early part of the uh the broadcast this evening. Thank you, Stea. Very nice. Very nice. Very nice. Um
156. 24:35what else we got in here? What' Claire have? Oh, a version of Black Betty for Kyle to practice. Very nice.
157. 24:48It's funny. Our uh our sales guy for Story Vine has has the Tom Celich mustache now. It's pretty funny. These
158. 24:56are all cute. All right. Groovy. Groovy. Groovy. Groovy. Groovy. Groovy. Groovy. So, in one of those Well, I
159. 25:08didn't have that on my bingo card for 2025. Um, I don't know if you all know the comedian Theo Vaughn. Um,
160. 25:20very funny dude, but very unassuming, a very weird kind of comedy. Um, I could see a lot of people not getting his
161. 25:29comedy, but he's he's like a super genuine dude. um he got and and so he's been he's been kind of elevating his stature
162. 25:42on the you know comedians they all sort of go on each other's podcasts. He's been doing a lot of podcasts and getting
163. 25:48a lot of interesting interviews and it looks like he's got this new higher production value podcast called This
164. 25:55Past Weekend and he just did an interview with Wen Phoenix and it was really trippy and weird because he's
165. 26:02weird and we Phoenix is weird and they just kind of weirded out with each other. um he got an interview with Sam Alman [Laughter]
166. 26:17and so I figure we'll watch some of it tonight because it's [ __ ] odd. It's just odd. Now, if anyone's got any
167. 26:27specific questions on AI, I'm happy to answer questions. I'm happy to demo demo [ __ ] I'm happy to go deal with the dog. Hold, please.
168. 26:38Stop. Champ just saved us from the evildoers. The evildoers out on the street. We're
169. 26:54safe now. All right. Sharon Crawford. Huh? No, he did not. This is a 90minute interview with Sam Alman.
170. 27:07I I just and I saw a couple of the clips at at one point at one point Theo Vaughn goes he goes, you know, essentially you
171. 27:17tech people are weird and he points out Peter Teal as this really weird he is he like an evil tech overlord and Sam
172. 27:25Alman's like we're actually really good friends and Theo Vaughn was like oh we don't have to talk about him.
173. 27:34Oh my god. So anyway, so I figure we'll play we'll play a little. All right. Okay. Fantastic. Fantastic.
174. 27:51Get off my lawn. All right, let's go. Let's just watch it. We'll watch it for a bit.
175. 28:11Let's let's just stop it right there. This this totally feels like the uh Bryant Gumble Katie Turk interview in 1994
176. 28:23or five where they're like, "What's that little the the A with the circle around it?" I think it's about it's what is
177. 28:33this internet thing anyway? And he's like he he's a big player in AI. Oh, no audio. Oh, that's cuz I'm sharing it
178. 28:41wrong. Cuz I'm a loser, baby. Why don't you sue me? Yeah. Fantastic. Okay. What we're going to do here is
179. 28:53we're going to need you to go on ahead and start that over so the good people on the YouTube there can see the
180. 28:59beginning of this thing. You see what I'm saying? All right. It's not the tab. What?
181. 29:14I think this is right. Yes, it's right. I was saying it. We need to turn this into the tab learning lab. >> Oh,
182. 29:24that was comedy. I thought I was actually getting a an actionable note. No, I was just being made fun of. Okay, here we go.
183. 29:34Today's guest is uh well, dude's a straightup tech lord, let's be honest. He's uh he's one of the leaders, the
184. 29:41world leaders in the development of AI. Um he started open AI which is known for uh having chat GPT. Uh we had a fascinating chat about the
185. 29:53pros and cons um the fears and hopes everything I could learn about uh about artificial intelligence and where we're
186. 30:01headed. TBD baby today's guest. I mean this oddly feels like a big deal to me. It it it does like Theo Vaughn's got an
187. 30:16audience that probably ain't deep into AI like just the way he talked about it just the way he pronounced it. Chat J pay [Laughter]
188. 30:26>> is Mr. Sam Alman and I'm very thankful for his time. [Music] [Applause]
189. 30:42[Music] [Applause] [Music] You know, we had a residential architect do this office. We wanted it to feel
190. 30:50like someone's like really comfortable like country house or something like that. >> Yeah. >> Not like the big corporate like sci-fi castle.
191. 30:56>> Yeah. That's what I was I was like a little bit like, oh, is it going to be, you know, will there be a drawbridge?
192. 31:01Will we be uploaded into a suite? Like what will happen to us? You know, >> we don't want that for like residential.
193. 31:07>> Yeah. I was like, how do we even get through the firewall? How many like hit points will we need to get through? You
194. 31:11know, it got very Dungeons and Dragons uh in some of my like um imagination sometimes. >> Yeah, we want people to feel like super
195. 31:19comfortable and tried to get pretty far in that direction. >> It feels like it. Your staff's very sweet, nice people. Um you have Thanks
196. 31:27for hanging out, man. >> Absolutely. Really appreciate it. Uh >> yeah, I haven't seen you since I fell out of my
197. 31:32>> f chair at the inauguration. That was really like quite a way to meet you. Yeah, I felt so embarrassed and you were
198. 31:37one of the faces that I looked up and saw and I was like, "God and that was my first moment like AI build us a better
199. 31:42chair to be honest with you." You >> and you did nothing, right? You were just sitting there and it just collapsed. Nothing.
200. 31:48>> I remember that. >> And it was just so embarrassing. I was like, "Oh, of all people me and here I am in this place and uh
201. 31:54>> I think it was perfect because everybody's got to have some sto when people are like, "Oh, it was the
202. 31:57inauguration." Like, everybody's got to have some story to tell. >> Yeah. >> And that was an incredible story for us all to tell.
203. 32:02>> That's a good point. I do remember looking at people for help though. And oddly your eyes I I was like, "Oh my god, he
204. 32:10>> Yeah, Danielle. Yeah, this is the Today Show moment of 2025." It totally is. It totally is. Yeah. And Brandon, they've
205. 32:16had Wait, they they've had this set at OpenAI the whole time. They They could be doing non-awwkward product launches
206. 32:22on this set. Yeah, they could. >> Could help. You did look like a beacon of help in the distance. >> I tried to help. >> Um,
207. 32:31you have a baby. You have a new >> Yeah. >> child. >> It is. >> There have been like a lot of experiences in life where everyone tells
208. 32:40you something's going to be great and then it's like, okay, the people are right, the consensus is right. It's like
209. 32:45even better than I thought it was going to be. But this has been the strongest example of that ever.
210. 32:49>> Like I knew it was going to be great and it's like way way better. It's impossible to describe. There's nothing
211. 32:54I can say that's not like very cliche and it's totally amazing. >> What is like one of your And it's a you
212. 32:59have a young boy. Yeah. And what's something like that you think is like neat or like what's one thing that kind
213. 33:06of like is bringing you joy with it? >> Watching the speed with which he like learns new things or gains new capabilities. Yeah.
214. 33:16Is just unbelievable. It's like every day it's like oh man >> just couldn't do that before and now he's like grabbing stuff and passing
215. 33:21between his hands and uh getting to like watch it dayto day is just an amazing rate of change. You know what's amazing is I feel like
216. 33:33his insight as to what it's like watching his baby develop and gain new knowledge while he's spent the past
217. 33:41seven or so years getting machines to do that and watching them progress. That's a fascinating that's a fascinating insight.
218. 33:49>> And then I don't like again I realize it's like >> how how is Sam Alman odder than Theo Vaughn? I love you all. Oh my god. You can make
219. 34:04money off chat GBT. You know, I realize that like everything about babies are very finely tuned over
220. 34:12>> and we're not going to necessarily watch this whole thing. It's 90 minutes, but I just I figure we'll watch it and get
221. 34:18some interesting insights here. I just I just know that he's going to ask Sam questions that no one else will ever ask him.
222. 34:27It's just such an odd pairing for >> a long period of evolution to make us like love them and be fascinated by
223. 34:33them. And it's like a neurochemical hack. But I love it. It's great. It's so strong. It's so intense. >> So, it's really like almost like a
224. 34:39coffee for your heart or something kind of. >> I don't even know how to find I've tried to like come up with an analogy to tell
225. 34:46because now I'm like telling everybody you got to have a lot of kids. It's really important. >> Yeah. Um,
226. 34:49>> and I've been looking for an analogy of what to explain. And then I always just say like I I don't know how to explain
227. 34:54this. It's just it is the best thing I've ever done by far. I feel like >> a completely changed person. And I was I
228. 35:01was like thinking the other day like there used to be all these other like at this point all I do is work and hang out
229. 35:07with my family. I like I don't I don't like really get to do a lot of hobbies anymore. >> Silver Fox, how old is the baby? 4.01 months.
230. 35:20right? Busy time at work. I don't get to hang out with my friends that much. Uh and I and I don't you know there were
231. 35:27like all these things where people tell you like oh you got a baby coming. You got to go you know take that spontaneous
232. 35:31international trip cuz you're not going to be doing that again for a long time. And I was like oh that is kind of sad.
233. 35:37In practice you don't do it that often. I at least didn't do it that often. And I don't miss it at all. I like remember
234. 35:42that that used to be a possibility. Now I can see that's not going to be a possibility for a long time and I'm thrilled with the trade.
235. 35:47>> You're moved on. >> I'm so happy. >> Um, how old is your child? >> Four months. >> Oh, that's a funny like at five or six
236. 35:53months they start to get like fun and you can like they're still like they can't go anywhere, you know, but they're
237. 35:58like intrigued and stuff. They start to like smile or process more. I don't know how you guys say it. But um
238. 36:04>> yeah, he's totally like turned on though, like really aware, understands things. It's super cool. I have a
239. 36:08thought sometimes that this will be one of the last like maybe 40 years that we conceive children in the body. Did you
240. 36:16have any thoughts about that? >> I've definitely heard a lot of people say that. Um, >> okay. That's the best [ __ ] question ever.
241. 36:29This is the last 40 years or so that children are going to be se conceived inside the body. And Sam's like, "H, [Laughter]
242. 36:45I haven't thought about it hard myself, but yeah, I guess it does make sense." Like, I guess that does make sense.
243. 36:52>> Like, God, you were in your mom's. It's crazy, you know? You pervert or whatever. Like, like I think in the
244. 36:57future people will be it'll be kind of done like in a >> in a vet or something. >> Yes. In like a nice vet. You can go see
245. 37:02it on the weekends or whatever. And like >> doesn't that just feel like off to you? Like I can totally intellectually like
246. 37:08understand that that may be the better way to do it. >> Oh yeah, it feels way off to me. I was trying to
247. 37:13>> look at look at Sam Alman trying to trying to go like dude like don't go this dark. Sam Alman's defending humanity here.
248. 37:24>> Thought you would like it. You know, I thought >> I thought you would like it. Theo Theo trying to suck up to what he
249. 37:40thinks Sam Holman is. Oh my god, this is great. I mean >> like or I thought you that would be like a thought like I guess I like that for
250. 37:50me that's one of like my futuristic thoughts you know >> like I can totally accept that that will be what everybody does and that it's you
251. 37:58know easier and we can like make it healthier for the child and mother you know the mother doesn't take the health risk but >> but man
252. 38:06>> so intellectually I can say that and then like emotionally it feels like h something is off of that yeah
253. 38:12>> oh yeah yeah cuz then the family like on the weekends the parents would come and like tinker on the glass or whatever or
254. 38:18the dad would put like a um you know like a go falcon sticker on the thing or you know what I'm saying people would
255. 38:24like decorate it all up or write little messages on there. Um, >> you know, I think there's another like
256. 38:29another take I have on all of this is that there in this world that we're heading to of like crazy sci-fi
257. 38:35technology becoming reality, the the sort of like the deeply human things will become the most precious, sacred, valued things
258. 38:43>> and that we'll really care about like the >> See, I I I deeply agree with that. I I do think that AI is going to push us to
259. 38:51this weird place where like human interactions become elevated. They become more important. It's interesting to hear Sam do that. I
260. 39:05What a what a fascinating take from from Theo Vaughn. Like I thought this is who you were. I thought you'd like this
261. 39:14dystopian future. And Sam's like, "Uh, we just talked about my baby, dude. >> The human experience more than ever."
262. 39:22And maybe it won't go that way. I don't know. >> Yeah, >> man. >> Do you uh No, and that's some of the
263. 39:27stuff we want to talk about. And thanks, thanks so much, man, for sitting down. Um, do you think your child will go to
264. 39:32college? Do you think like what do you kind of think that looks like? >> Probably not. Um, if I had to guess,
265. 39:38like I I think Well, I only went to half of college. >> You You did you drop out? >> Yeah, >> dude. You guys all I freaking dropped
266. 39:44out. I didn't get [ __ ] You dropped out. Wang dropped out. Zuckerberg dropped out. >> Um, >> probably a lot of other people. >> And you?
267. 39:52>> Yeah. Yeah. Okay. Well, hey, we're both here, so Oh, it worked out fine. >> And you Oh my god.
268. 40:07Oh my god. >> Guys, you're right. You know, you're right. Never mind. I'm sorry. I'm being self-defeating.
269. 40:12>> Um, yeah. What does that look like when you think about that like >> yeah with AI with so much new information coming online right and so
270. 40:20much like data being collected and like um information being uh carpulled and and maybe which is a term
271. 40:27>> so you you and I never grew up in a world that didn't have computers like and our parents were like oh this there
272. 40:34weren't computers and then there were and it was this big crazy adjustment it took them a long time to figure it out
273. 40:37but to us like computers just always existed they were just I mean maybe they were kind of new but they were always around
274. 40:44>> and and then like you know a kid that is like there was there was this video on YouTube I saw like maybe 12 years ago
275. 40:52something like that that 14 years ago that just really stuck with me. It was like a little baby in a dentist waiting
276. 40:58room or something picking up one of those old glossy magazines and going like this. >> Oh, I remember that.
277. 41:03>> And to that kid it was just like a broken iPad because that kid had just like grown up in a world where like
278. 41:08there were touch screens everywhere. And my kid will never grow up will never ever be smarter than an AI. >> Yep. >> That will never happen.
279. 41:19>> That's fascinating. My kid will never ever be smarter than an AI. That will never happen. Right. I've talked about this on this
280. 41:28channel that we are we are in this very rare slice in history, incredibly rare slice in history where any of us who if you're conscious
281. 41:41enough to watch the AI learning lab and demented enough to watch the AI learning lab, we get the privilege and I think this is
282. 41:51an absolute [ __ ] privilege of knowing what the world was like before AI and we're going to know what the world is like after AI.
283. 42:03We're the cusp generation of this transformation and it is such an incredibly rare moment in history. And his kid is part of the generation
284. 42:18that will never know what it's like to not have an entity, an AI entity that's smarter than humans. It's amazing.
285. 42:29>> Unless it's implanted, right? Yeah. Exactly. >> You know, kid born a few years ago, they had a brief period of time. My kid never
286. 42:34will be smarter, but also they'll never they'll never know a world where like products and services aren't
287. 42:42way smarter than them and and super capable. They can just do whatever you need. And in that world, I think education's going to feel very
288. 42:49different. I already think college is like maybe not working great for most people, but yeah, I think fast forward
289. 42:5618 years, it's going to look like a very, very different thing. >> Yeah. Yeah. Do you think there Oh,
290. 43:00here's that video right here. This kid. >> Yeah. Yeah. Yeah. All right. I was wrong about the dentist. It was Or maybe
291. 43:05there's a few of these. >> He's like, "Somebody charge this magazine." He's yelling. How would you recommend to a parent right now to
292. 43:10prepare their children for like an AI? >> Yeah, that's good. Um, source camp. Some of us have gone from pre pre- microwave
293. 43:17to post AI. Exactly. Exactly. Or pre video recorder. I remember we had a this was 19
294. 43:35like 78 or nine. We had a Magnavox. I It must have been like a 27 or 32 inch television. It was this giant white white box with a TV in
295. 43:48it and on top of it was a VHS player that was it was like the size of a trunk of a car and the thing would come out of it.
296. 44:00Some of us are about to go from pre prejetsons to postJetsons. That's really good. God damn this this interview like
297. 44:14I think Theo Vaughn is being smacked in the face here with
298. 44:29the the future is not the cliche he thought it was. He doesn't know what it is yet, but like he's coming in with these
299. 44:38preconceived notions about what the future's going to be like and who who the overlords of that future are. Fascinating. Just fascinating.
300. 44:47>> My future kind of like are there certain curtails that you would start to put in now? Are there certain like um you know
301. 44:54adjustments where you like get them in a certain training or have them start to watch certain models of things online? Like what is that?
302. 45:00>> You know, I I actually think the kids will be fine. I'm worried about the parents. >> If you look at the history of the world
303. 45:05here, when there's new technology, like people that grow up with it, they're always fluent. They always figure out
304. 45:10what to do, they always learn the new kind of jobs. But if you're like a 50-year-old and you have to like kind of learn to do
305. 45:17things in a very different way, that doesn't always work. Yeah. >> So, I think the kids are going to be fine. I mean, I do have
306. 45:23>> That's all of you. He's talking about you, you old fat people. Definitely not me because I'm not old or fat. >> Shut up.
307. 45:35>> Worry. Like I do have worries about kids in technology. Like I think this scrolling the kind of like you know >> silver fox down boy
308. 45:44>> short video feed dopamine hit. It feels like it's probably messing with >> Sharon Crawford. Not me.
309. 45:51>> Kids brain development in a super deep way. So it's not that I have no worries. I have like extremely deep worries about
310. 45:55what technology is doing to kids. But in terms of kids ability to like be prepared for the future and use a new
311. 46:00technology, they seem really good at that. >> Yeah. >> Always through history. >> That's a good point actually. Yeah. It's
312. 46:05like if you just grow up with it, it's just like having it's just totally normal. It's like having kneecaps or
313. 46:10whatever. You're just kind of used to it. >> You you can't imagine the world where >> it's like having kneecaps. [Laughter] Oh my god.
314. 46:26This This is just the bro most broken interview. It's so good. It's It's so good. >> It doesn't exist. You just Right. Yeah.
315. 46:35Yeah. That's a good point. >> I remember when I was >> You know what it is about Theo Vaughn is he's so sincere. Like he just like just
316. 46:44the way his brain works like the dots that he connects. It's just it's [ __ ] off. It's like a dude that huffed paint
317. 46:54as a teenager trying to navigate the world >> uh in school in like junior high and Google first came out
318. 47:06>> and all the teachers like freaked out and they're like this is the end of education you know I know if you why do
319. 47:12you have to memorize history facts in history class if you can just look them up instantly on the internet you don't
320. 47:16even have to learn to go to the library and the answer is like yeah maybe memorization is less important but with
321. 47:21these new tools, you can think better, come up with new ideas, do new stuff. I'm sure the same thing happened with a calculator before.
322. 47:27>> Yeah. >> And you know, now this is like this is just a new tool that exists in the tool chain. >> And what about like say if there is
323. 47:33somebody though that's like learning history right now, like they just started their second year of college.
324. 47:37>> Oh, that Celsius. Yeah, that thing will definitely you won't be able to blink for a month, homie. That thing will
325. 47:42Yeah, you'll sneeze and release. [ __ ] >> You won't be able to blink for a month. He gave him an energy drink. Oh my god.
326. 47:59Oh, something I wonder about is because I'm deep in AI, I don't come up with new fresh perspectives. Keep trying new
327. 48:06things. That's interesting. That's interesting. I mean, I think there's a couple of ways that you could
328. 48:13look at that, Gareth. I mean, the other one is that maybe you're maybe you're considering wider perspectives than you
329. 48:20would have otherwise because you can travel so far knowledgewise, you know, laterally, but who knows?
330. 48:26>> 5.0, dude. You'll freaking Are you guys at 4.5 already? >> We're 4.5 already. 5.0 is uh I think it's going to be great.
331. 48:32>> Oh, it'll come out fast if you had that Celsius. I'm saying you >> maybe the researchers need it, not me.
332. 48:36But uh you know, we'll get them some. >> Yeah, that'll get you there, man. Um so, say there's somebody just for example,
333. 48:42like that's learning history right now. They're in their second year of college. They're they're taking history. Is that
334. 48:46are there some subjects in like like they they're going to be a historian? Is that still a viable space of work? Uh as
335. 48:54AI moves forward, do you think? Honestly, >> I I assume there will be some version of it that is uh I I think it's very hard
336. 49:04to predict exactly how something evolves. Um I or predict exactly the jobs of the future going to be like the
337. 49:14you know not that long ago it would have been very hard to predict either of our jobs if you go back 100 years the idea
338. 49:22of like this CEO of an AI company or a podcaster like you know probably would have been things that didn't seem to be
339. 49:31the most obvious evolutions of the things people were doing at the time. >> Yeah. You just seemed almost probably
340. 49:36crazy even in trying to explain those to someone. >> You would. And now, in fact, two of the job I I heard that the job that young
341. 49:43people most want is some version of your job. >> The job that young people most want is to be uh, you know, podcast influencer,
342. 49:50uh, YouTube, they want a YouTube channel, like whatever it is, they they like six, seven year olds. I don't know
343. 49:55how to describe it, but that's what they want. >> And a lot of people also want my job. They want to do like a startup or they
344. 50:00want to work on AI. And these just didn't exist. >> Yeah. So like the rate with which the new things come along is is fast and
345. 50:07also trying to predict what they are. I don't know. The thing I say all the time is no one knows what happens next. It's
346. 50:13like we're going to figure this out. It's this weird emergent thing. Does the current job of a historian exist in the
347. 50:18same way? I'll bet not quite. But another thing I believe is that humans are obsessed with other people. Like we
348. 50:25are so deeply wired to care about other people to care about stories and history our own history is extremely interesting
349. 50:31to us. So I would say somehow or other we're still going to care about that. There's going to be some kind of job doing that.
350. 50:38>> Man, that's cool. G I guess I I if when I take that avenue of thought like okay there will still be this historian
351. 50:45or some it'll be some evolution of that. Right. That does seem kind of cool to me because there's a level of creativity in
352. 50:52there. There's a level of like faith and spontaneity in there that I think is kind of exciting. So yeah, I guess I
353. 50:58hadn't really thought about that. Sometimes I get stuck in this doomsday thing like I just see like you know like
354. 51:02the history book closes and they're like we have enough we have all the history over here you know >> it you know people used to say like oh
355. 51:09there's no need for more music we've made perfect music like why does anyone need anyone to create anymore and that's
356. 51:14obviously ridiculous. Yeah. >> Or they would say there's that famous patent office quote, everything that humans ever possibly need has been
357. 51:20invented. There's nothing left to do. >> I have heard that. But here we are. >> Here we are. And and like
358. 51:26>> Yep. And I know Shadow, I saw your comment there about, yeah, let's pollute the environment even more. Yeah. I mean, listen,
359. 51:33so so these tools right now take a lot of energy. These tools are also going to help us solve energy problems and
360. 51:41they're going to help um come up with new kinds of energy or new kinds of energy production. They're also going to
361. 51:48get a lot more efficient over time. Um but they're not going to go away. Like part of the part of this channel is is
362. 51:57not to judge it good or bad, but just to say if it's not going away, how do we deal with it? So, um so that's that's
363. 52:04what we're doing here. Someone asked me the other day like, you know, how long is it until you can make like a AI AI
364. 52:09CEO for OpenAI? And I was like, probably not that long. And they were like, well, aren't you really sad about that? And I
365. 52:16was like, no, I think it's awesome. I'm for sure going to figure out something else to do. I'm excited to do that. Like,
366. 52:21>> I think that's great. >> Right. So, you could create something that would have your job, but then you could do something else.
367. 52:29>> Totally. >> But then how do you know that you'll still get paid for your job? I guess like >> well
368. 52:34>> it's kind of a big question. I I kind of think that >> but yeah I guess the framing of that question might be better like say there
369. 52:39are jobs that get curtailed by >> there will be some >> okay >> I think it's important to be honest
370. 52:43>> there will be some there's going to be a lot there's going to be a lot but what no one talks about is jobs going to be
371. 52:49created so we'll see we'll see where Sam goes with this this is interesting >> it's about that there will be some jobs that totally go away
372. 52:56>> but mostly I think we we we will rely on the fact that people's desire for more stuff for better experiences for, you
373. 53:05know, a higher social status or whatever, that seems basically limitless. Human creativity seems basically limitless. And human desire to
374. 53:13like be useful to each other and to connect with each other and do stuff for each other and focus on other people
375. 53:18seems pretty limitless, too. So, I think throughout all of history, there have been these predictions like ah, you know, we're gonna
376. 53:27like all be on the beach and work an hour a day or hour a week or whatever and we're going to have unlimited wealth and and
377. 53:33>> I've never heard that one. I would I love that. >> I mean, they used to say this. They used to like the industrial revolution,
378. 53:38people were like, "Oh, you know, we've just figured out how to automate like man's lot in life. There's nothing left
379. 53:43to do. We're going to have these machines do all the work." >> Makes sense, probably. and you watch these machines doing all this stuff that
380. 53:48only people used to physically do and everybody panicked and said there's going to be no more jobs and we figured out
381. 53:56>> new stuff to want. Now here's an interesting thing. If you could go back to that industrial revolution time and
382. 54:00people before that were, you know, really, >> which by the way, I think the industrial revolution to me feels like the thing
383. 54:08that is the closest the closest parallel to what's going on with AI. Where the industrial revolution augmented our
384. 54:15physical strength, AI is augmenting our intellectual strength, our intelligence. Um, and if you look at the absolute
385. 54:25transformation that happened, you know, from the the late 19th century through the 20th century, um, like
386. 54:37that's about to happen with with our intellect, right? So, just like these machines got way stronger than any human
387. 54:45and we figured out other things to do, that's about to happen with our intellect. Um, I think it's going to
388. 54:51happen a lot faster than the transition uh in the industrial revolution just because there's not so much physical
389. 54:58stuff to do. Um, and I think because of that the transition is going to be really profound and and terrifying >> on the grind.
390. 55:10>> Yeah. Exactly. P M the industrial revolution didn't happen this fast. Exactly. Yeah. You had time from from
391. 55:17the day they invented the steam engine, this thing that went kadunk, you know, and could make a belt, you
392. 55:23know, go around like this till when they put that steam engine into a tractor and then, you know, made that tractor pull a
393. 55:30plow, right? There there were like years, decades in between th those major chunks of things. But you did have, you
394. 55:39know, spans of 20 or 30 years where you had, you know, massive amounts of new inventions that changed everything.
395. 55:51So what will happen with massive info overload and organizing everyone sharing and doing research exponent exponentially? Fascinating to
396. 56:00think about and position for. Josh, that's a great question. I like implicit in your question is that we humans are
397. 56:10going to have to deal with the information overload. We may not, right? It might be the kind of thing where the AI is
398. 56:20at any point if any human needs anything, the AI either makes it or goes and finds it. And so, so I think I think
399. 56:31information overload is is what Google tried to solve 25 years ago, right? They said, "Hey, if we could index all of this information
400. 56:43and we could make a little search box and we could do that, well, then someone rather than having to navigate
401. 56:52by clicking on all these hyperlinks, they could just ask our thing and we'll tell them where it is in this giant information space."
402. 57:02AI represents all of that information is now compressed into this magic box that we just get to tap into and knowledge
403. 57:11emerges out of it. And so, so while massive amounts of information are being created by AI, I don't think
404. 57:21we're going to have to process all that much. I mean, you know, your friends and co-workers that use AI to write those 20page
405. 57:31proposals for you that you then put into chat GPT and distill it down into five bullet points. There'll be a lot of that.
406. 57:44Super hard trying to like kind of have enough food to survive. Go back to those people. Look at our jobs today. Would
407. 57:51those people say we have real jobs or would they say you have unbelievable abundance, unbelievable wealth, so much
408. 57:56food to eat, incredible luxury, and you guys are just like playing a game to entertain yourselves. Is that a real job
409. 58:02or not? And they would probably say where they sit, what you guys do is not a real job. You guys are, you know,
410. 58:08you're too rich. You're wasting your time. This is this is the argument that a lot of creatives are making right now
411. 58:13that if you just push a button and it generates something, that's not real creative, right? Is is what he's talking
412. 58:19about now. that you know someone from the industrial revolution looking at what we do now that's not real work >> you're trying to
413. 58:25>> right real work is men lifting things >> like >> yeah you guys are a couple of dang zest lords out there freaking playing uno in
414. 58:32the park or whatever they would they would not I don't think my grandfather would be like you have a job he would
415. 58:36still be like you need to get a job >> yeah totally and when we look forward another hundred years of what people are
416. 58:41doing they'll probably think they're working very hard it'll feel very satisfying very intense to them they're
417. 58:46really like they'll feel engaged they'll be making my people happy they'll be creating value for each other. But if we
418. 58:51could look forward that hundred years at those guys, do you think we would say they're working or like man you have
419. 58:55like AI doing everything for you? You're just trying to entertain. >> Yeah. Like I could see so so it's like
420. 59:00if if if AI ends up handling most of the tasks that we do now what might become valuable is like taking a walk with a human being. And so
421. 59:13there might be professional hikers that just walk with other people, right? I just want to like walk and hang
422. 59:21out with someone and I'm willing to pay for that and that's a job. And there might be someone else that's just like,
423. 59:26you know, teaching juggling on a street corner might become a respectable job. We'll see. >> Yeah. Like, oh, you guys have it so
424. 59:39easy, right? But I think that's beautiful. I think it's great that those people in the past think we have it so
425. 59:44easy. I think it's great that we think those people in the future have it so easy. >> Yeah. Valerie Cox, I hope life work
426. 59:49balance swings toward more life enjoyment. And that's that's kind of what I'm saying that that a lot of the tasks that we call work today.
427. 1:00:03We take pride in doing our work, but a lot of the tasks that we do are mind-numbing, shitty tasks. Imagine if those
428. 1:00:13mind-numbing shitty tasks were done, handled. Assume we navigate through the money piece of this, right? Everyone's got to
429. 1:00:22figure out how to get paid or whatever that looks like in the future. Assume we're going to navigate through that
430. 1:00:26because we're humans and we adapt. And then what's on the other side of that? Well, we would probably start to
431. 1:00:33prioritize things like life enjoyment. Well, what gives me more joy in life? Well, I like camping. Well, I like
432. 1:00:39riding bikes. Well, I like doing this. I like doing that. Maybe the things that seem like leisure, the thing we call leisure right now,
433. 1:00:49which we [ __ ] all over, right? You only do leisure on the weekends. We don't do leisure for a job. We work for a job.
434. 1:00:58Well, maybe that shifts in the future. >> Like that is the beautiful story of >> Sher Sharon Crawford. There's a chance.
435. 1:01:07>> Oh [ __ ] I I lost I just moved something. >> Celsius. Yeah, that thing will definitely you won't be able to blink.
436. 1:01:12>> A historian exists in the same way. I would bet not quite. But another thing I believe is that humans are obsessed with
437. 1:01:19other people. Like we are so deeply wired to care about other people, to care about stories in history, our own
438. 1:01:24history is extremely interesting to us. So I would say somehow or other we're still going to care about that. There's
439. 1:01:30going to be some kind of job doing that. >> Man, that's cool. Yeah, I guess I I if when I take that avenue of thought like,
440. 1:01:38okay, there will still be this >> live live for I like that, Rose. Live for the hobbies. The hobification of
441. 1:01:44humans. Yeah. Like like I I I think the point that that they're both saying here or that Sam's saying to him is like
442. 1:01:53Theo, look at your job right now. Your job is to like sit and talk to people and let other people watch that. Like
443. 1:02:04someone from a hundred years ago, that's not work. What are you doing? Go do something with your life, you [ __ ]
444. 1:02:12slacker, and get a haircut. Shave off that mullet. Historian or somebody, it'll be some evolution of
445. 1:02:23that, right? that does seem kind of cool to me because there's a level of creativity in there. There's a level of
446. 1:02:29like faith and spontaneity in there that I think is kind of exciting. So, yeah, I guess I hadn't really thought about
447. 1:02:34that. Sometimes I get stuck in this doomsday thing. Like I just see like, you know, like the history book closes
448. 1:02:39and they're like, "We have enough. We have all the history over here." You know, it, >> you know, people used to say like, "Oh,
449. 1:02:44there's no need for more music. We've made perfect music. Like, why does anyone need anyone to create anymore?"
450. 1:02:49And that's obviously ridiculous. Or they would say there's that famous patent office quote everything that humans ever
451. 1:02:54possibly need abundance unbelievable wealth so much food to eat >> that space right like the like the movement that happens with AI and with
452. 1:03:02just technology which will advance >> and when we look forward another hundred years of what people are doing they'll probably
453. 1:03:07>> think they're working very hard it'll feel very satisfying very intense to them they're really like they'll feel
454. 1:03:11engaged they'll be making people happy they'll be creating value for each other but if we could look forward that
455. 1:03:16hundred years at those guys do you think we would say they're working or like man you have like AI doing everything for
456. 1:03:21you. You're just trying to entertain yourselves. >> Yeah. Like, oh, you guys have it so easy. Right.
457. 1:03:26>> But I think that's beautiful. I think it's great that >> Jeff Flanigan, George Jetson, I'm so tired of pushing this button.
458. 1:03:35Like on Star Trek. Yeah, exactly. Shane Saw. I blindly trust AI will fix a lot of things. I don't have faith that
459. 1:03:42humans will allow being fixed. Well, yeah. Humans are always going to be messy. Humans are always going to be
460. 1:03:50messy, but a lot of the a lot of the [ __ ] angst of the world right now is the like the nature of the work we
461. 1:04:02do. Like people are stressed out and like a lot of our economy is based on complexity and a lot of people's jobs are dealing
462. 1:04:14with complexity, right? So, like if you think about, I don't know, an insurance adjuster, their job is to make it as complicated as possible
463. 1:04:27to try to get someone to process their claim so that they don't have to pay out the money so they can make more profit.
464. 1:04:35Like, by design, it is overly complicated. The US health care system is overly complicated by design. So as many people as possible get paid
465. 1:04:48in the process of making you well. And then on the other side of that equation is entities trying not to pay for any of that, right? And
466. 1:05:00making it as complicated as possible so as many people as possible can take a toll out of the transaction of getting you better. Right?
467. 1:05:13That's most of our society today. The global economy is based on either evolved complexity or designed complexity.
468. 1:05:25AI is going to simplify a lot of that and that that starts to shift things and maybe we become less shitty to each other.
469. 1:05:32>> Those people in the past think we have it so easy. I think it's great that we think those people in the future have it
470. 1:05:36so easy. Like that is the beautiful story of us all contributing to human progress and everybody's lives getting better and better.
471. 1:05:44say we're able to get to that space, right? Like the move like the movement that happens with AI and with just
472. 1:05:49technology which will advance quicker I think which is one thing that AI feels like to me it's a fast forward button on
473. 1:05:55technology and on uh possibility because things can be information can be quantified so quick and a lot of like uh
474. 1:06:02more menial tasks even though they're not really menial in people's lives um but menial hypothetically uh can be done
475. 1:06:09quicker to get a lot of the framework for things done fast. But how will people survive? Like how do we adjust
476. 1:06:16our structure of finan of like if some people own the companies that have the AI and then a lot of people um are just
477. 1:06:25using the AIs and the agents created by AIS to do things for them. How will society like societal members still be
478. 1:06:33able to financially survive? Will there still be money? What is that? >> This is a great question. This is like
479. 1:06:39th this is this is the question no one wants to ask because they don't want to hear the answer, right? What the [ __ ]
480. 1:06:46happens? We we get to that future where AI does everything. How are people getting paid? It'll be interesting to hear Sam's
481. 1:06:54answer. There were some interesting uh posts on YouTube. Um Josh Groves, that's why I made right node uh an app. It shouldn't be this way.
482. 1:07:05Stop the madness. Make it easier. Bill Gates said he hires lazy people because they find the easiest route to um they
483. 1:07:14they find the easiest route. AI can help us with this art form. This is why master synthesizers um and system thinkers will win. Like
484. 1:07:22you always say, Kyle, fighting for the future. Yep. The hardest problem in therapy was always convincing people to
485. 1:07:29loosen their grip on the problems they have to grab different ones. Interesting. If AI takes away the financial
486. 1:07:36requirements, there's still the human need for control. Interesting. Fascinating. Fascinating. This is cool. I'm I'm digging this. I
487. 1:07:45listen, this is a non-standard. It's movie night at the at the learning lab. Um I'm fascinated by this interview
488. 1:07:54because again, I think this is the most unlikely pairing I could have imagined in 2025. And uh I love that Theo Vaughn, you can
489. 1:08:04tell he he's not played with AI. I love that he's coming in with these assumptions and he's asking the
490. 1:08:10questions everyone's probably wondering. It's really really good. >> Does that make any sense? That question.
491. 1:08:16>> It totally makes sense. Uh sorry. I don't know. Neither does anybody else. But I'll tell you my current best guess. >> Okay.
492. 1:08:21>> Well, I'll say two guesses. one I think it is possible that we put you know GPT7 or whatever in everybody's chat GBT
493. 1:08:30everybody gets it for free and everybody has access to just this like crazy thing such that everybody can be more
494. 1:08:36productive make way more money doesn't actually matter that you don't like own the cluster itself but everybody gets to
495. 1:08:42use it and it turns out even getting to use it is enough that people are like getting richer faster and more
496. 1:08:48distributed than ever before that could happen I think that really is possible there's Another version of this where
497. 1:08:56the most important things that are happening are these systems are discovering you know new cures for
498. 1:09:02diseases new kinds of energy new ways to make spaceships whatever and most of that value is acrewing to the like
499. 1:09:08cluster owners us just so that I'm not dodging the question here and then I think >> an AI cluster owner is an individual or
500. 1:09:16entity responsible for the design deployment management and ongoing maintenance of an AI I compute cluster.
501. 1:09:24So, so Altman's saying it's possible that because the AI will produce all this remarkable stuff that the people with
502. 1:09:33those clusters will basically hoard the compute and they'll do all the fancy good [ __ ] and keep all the money. So,
503. 1:09:40let's see how he answers it. >> Will very quickly say, "Okay, we got to have some new some new economic model
504. 1:09:47where we share that and distribute that to people." To be honest, I've never heard of Theo Von. Yeah, if you're not
505. 1:09:53if you're not kind of deep in the comedy world, he's he's um he's an acquired taste. I think he's funny as [ __ ] but
506. 1:10:00like a lot of times I don't know that he knows why he's as funny as he is. Um he just sees the world in this very
507. 1:10:09different way and that he's interviewing Sam Alman like I don't I want to know how this [ __ ] thing came about.
508. 1:10:16Although at right at the beginning um apparently they were at both at the inauguration and Theo Vaughn's chair
509. 1:10:22fell apart right in front of Sam Alman. So I think that's how they met um was Theo Vaughn crashing crashing to the
510. 1:10:30ground at the uh the Trump inauguration. >> Uh I used to be really excited. >> What about people's competitive nature?
511. 1:10:40I'm not sure how you can completely level the playing field. So you know what, Rose? I don't I don't think you
512. 1:10:47have to I don't think I don't think human nature goes away, right? I think people will find new ways to compete if if the AI
513. 1:11:00like like one of the things that Sam was just talking about here before where he said if everyone's got access to chat
514. 1:11:06GPT7 sa Sam Alman about a year and a half or two years ago talked about this concept of universal basic compute and
515. 1:11:15and you've probably heard of universal basic income where you know basically you take the profit from all these
516. 1:11:21machines means you tax these frontier model companies, you take, you know, some percentage of the profits, you
517. 1:11:28distribute that back out to the people that lost their jobs, right? That's one model, universal basic income. Sam was
518. 1:11:35talking about the concept of universal basic compute. And what he was basically saying is you give every person some
519. 1:11:42amount of compute per month that maybe that amount of compute is the equivalent of having 10 employees. And these are
520. 1:11:52like 10 of the best employees you could ever have or 100 employees that every human being could have a hundred
521. 1:11:57employees and those employees would do anything that that human wanted them to do. Well, some humans are going to be
522. 1:12:03ambitious and they're going to want they're going to say like I want more than a 100red employees. I want to I
523. 1:12:08want a thousand employees or I want 10,000 employees. And then other people are going to be like I don't really have
524. 1:12:15use for 100 employees. I don't care. So why don't you just pay me what it's worth to have these hundred
525. 1:12:22employees go work for you, right? And so so that's the model he's talking about where if everyone's got access to these
526. 1:12:29super powerful things, I think human dynamics still play out in that. Some people are going to be more ambitious,
527. 1:12:34some are going to be less. People are still going to compete, but we're just you we're just going to have different
528. 1:12:39tools of the game to be able to do to to to be be able to compete. And then if we do move to the hobbyification of humans
529. 1:12:48where you know hobbies become as important as jobs, well then maybe we get into competitive needle point,
530. 1:12:56right? There's the national needle point um speed needle point competition. Like I I think human nature is still human
531. 1:13:04nature. It's just where we put our energy is going to shift >> about things like UBI. I still am kind of excited like universal basic income
532. 1:13:10where we just give everybody money. >> Yeah, you hear that term a lot. Yeah, universal basic income. Yeah, I heard
533. 1:13:14you and Rogan talk about that too a while back. >> I still am kind of excited about that. But I think people really need agency.
534. 1:13:20Like they really need to feel like they have a voice in governing the future and deciding where things go. And I think if
535. 1:13:27you just like say, okay, AI is going to do everything and then everybody gets like a, you know, dividend from that, it's not going to
536. 1:13:35feel good. And and I don't think it actually would be good for people. So, I think we need to find a way where we're
537. 1:13:42not just like if we're in this world where we're not just distributing money or wealth like actually I I don't just want like a check every
538. 1:13:51month. What I would want is like a ownership share in whatever the AI creates so that I feel like I'm
539. 1:13:55participating in this thing that's going to compound and get more valuable over time. So, I sort of like universal basic
540. 1:14:01wealth. By the way, if you're curious about this whole topic of universal basic income, universal basic compute,
541. 1:14:09you should follow a guy on YouTube named David Shapiro. I talk about him a lot. Um, and he's got a his whole new thing
542. 1:14:16recently is post labor economics. He's got a whole model for this where you basically bring things down to the
543. 1:14:23municipal level and you basically use AI to exploit the resources in your local region and everybody participates in
544. 1:14:30that. It's a fascinating model, but there's there's we're going to obviously need to figure something out. There's no good answer
545. 1:14:38for this, but if you want someone who's thinking deeply about it, David Shapiro, >> better than universal basic income. And
546. 1:14:43I think I don't like basic either. I want like universal extreme wealth for everybody. Um, but but even then, like I
547. 1:14:50think what people really want is the agency to kind of co-create the future together. And and in a world where it's like the AI is
548. 1:15:00mostly coming up with the new scientific inventions, at least we've got to still have humans like invent the new culture
549. 1:15:06and have that be a very distributed thing. >> Okay. I guess yeah, I I I see what you're saying, but would that be like an
550. 1:15:14American thing, do you think? Like since it were invented here, or do you think I'm just wondering what does that look like? You know,
551. 1:15:21>> the economic model of it all, or the whole thing? >> Yeah. or like is there a dividend of the company that then is divided up between
552. 1:15:29the masses sort of >> I mean a crazy idea but in the spirit of crazy ideas is that if the world there's
553. 1:15:35like eight roughly 8 billion people in the world >> if the world can generate like eight quintilion tokens per year if
554. 1:15:43that's the world actually let's say the world can generate 20 trillion quintil 20 quintillion tokens per year >> tokens of
555. 1:15:50>> like each word generated by an AI >> okay >> just making up a huge number here. We'll say, okay, 12 of those go to, you know,
556. 1:15:56the normal capitalistic system, but eight of those eight quintilion tokens are going to get divided up equally
557. 1:16:02among 8 billion people. So, everybody gets one trillion tokens. >> And that's your kind of >> he's talking this is his universal basic
558. 1:16:10compute idea right here. >> Universal basic wealth globally. And people can sell those tokens. Like if I
559. 1:16:17don't need mine, I can sell them to you. We could pull ours together for some like new art project we want to do. But
560. 1:16:22but instead of just like getting a check, you're get everybody on Earth is getting like a slice of the world's AI
561. 1:16:28capacity and then we're letting the like massively distributed human ingenuity and creativity and economic engine do its thing.
562. 1:16:36>> I mean that's like a crazy idea. Maybe it's a bad one, but that's the kind of thing that I think sounds like someone
563. 1:16:41should think about it more. >> One of the big fears is like purpose, right? Like human purpose. Like work
564. 1:16:45gives us purpose. And also I think the idea that we are the ones advancing humanity gives us purpose like we are the
565. 1:16:56like yeah like we have some control over our own destiny maybe gives us this sense of purpose and it feels
566. 1:17:06like that we would lose a sense of purpose or that purpose would be adjusted like if AI is to really you
567. 1:17:13know continue to advance so quickly. It feels like our sense of purpose would start to really disappear.
568. 1:17:20Do have you had thoughts about that? >> I worry about this a lot. It's so I think people have worried about this
569. 1:17:25with every big technological revolution, but I agree that this time it feels different. Like >> Okay. Yeah. Because if say you had an
570. 1:17:31axe and somebody came out with a saw, you're like you're like, "Yeah, that's it." >> Or even if they come out with like a a
571. 1:17:36robot that cuts the tree down, it still feels fine. But like creativity intelligence I think cuts so deeply at
572. 1:17:43the core of whatever we are and how we how we value ourselves. Um one example we can look at this right
573. 1:17:51now I think one area where AI is having a big impact is on how people write software for a living and AI is really
574. 1:17:57good at that. It's really changed what it means to be a software developer. I haven't heard any of those software
575. 1:18:03developers say that they even though their job is different that they don't have meaning they still enjoy it.
576. 1:18:08they're operating at a higher level. Um, and I'm hopeful, at least for a long time, you know, 100 years from now, who
577. 1:18:15knows? But I'm hopeful that that's what it'll feel like with AI is even if we're asking it to solve huge problems for us,
578. 1:18:21even if we ask it to say like, you know, go discover a cure for cancer, there will still be a lot of things to do in that process
579. 1:18:30that feel valuable to a person. >> You'll still asking it the questions. You're still like helping guide it.
580. 1:18:37you're still framing it or whatever it is, you're still like talking to the world about it. And and I think all of human history
581. 1:18:47suggests that we find a way to put ourselves at the center of the story and feel really good about it. Like you know
582. 1:18:55if you kind of think like we used to think that the earth was the center of the solar system and then we're like very humanentric view and
583. 1:19:04then we're like okay fine the sun is the center of the solar system but the solar system is at least the center of the
584. 1:19:08galaxy and now oh man there's a lot of galaxies and oh man now we're this like tiny speck in this like very huge universe and
585. 1:19:18and yet we still manage to feel all like a lot of main character energy. And so I somehow think even in a world
586. 1:19:24>> Did you already said that was great? We we still feel like we we still feel like we have a lot of main character energy
587. 1:19:32that that even though the more insignificant we realize we are, we still find a way to put ourselves at the
588. 1:19:39center of the story. I think that's great. Yeah. That that's the thing. The thing that baffles me about doomers that
589. 1:19:48that say that AI is going to erase our meaning is they talk about humans as if they're static. Like like our current
590. 1:19:56point of view is always going to be our current point of view. [ __ ] changes. You adapt to it. You shift your point of
591. 1:20:05view, right? Like we're going to adapt. Like humans are going to adapt. We're not going to not adapt.
592. 1:20:12That's a a thing I've never gotten about that particular argument. >> AI is doing all of this stuff that
593. 1:20:17humans used to do. We are going to find a way in our own telling of the story to feel like the main characters. And I
594. 1:20:22think in an important sense we will be and that's really good. I also like you know probably already today
595. 1:20:32there could be a very compelling version of two AIs talking like this. >> Yeah. >> And I don't think I would want to watch that.
596. 1:20:39>> Right. Like I think I I really do feel deeply wired to like th >> This is an important point. So So when
597. 1:20:46when Big Blue beat Gary Kasparov Gary Kasparov at chess, the predictions were no one will ever play chess again. And we'll just have
598. 1:20:56machines playing chess against each other. And guess what? Nobody gives a [ __ ] about two machines playing chess
599. 1:21:03against each other. You know what? What all of that did? It made chess more popular and it made the appreciation of
600. 1:21:13humans that are really good at it more appreciated, right? Care about the real person behind it. I think that's like deep in the biology,
601. 1:21:25right? Yeah. That's the part that I think a lot of times it's like even though you can get into like these wormholes of like
602. 1:21:32possibility and these fear holes of possibility or um kind of this dystopian ideas that in the end I'm like I'd
603. 1:21:41rather probably watch something that's real, you know? It's like because I'm real. You know what I'm saying? Like I
604. 1:21:47don't want to talk really to a robot. I'd ra you know. Yeah. I think in the end there's going to be a part of you
605. 1:21:52that wants to continue to just talk to um talk to humans. Do you uh what's like one of your fears? Like
606. 1:22:00what's a fear you have of AI? Like if you have like a fearful space that it could go like I know you mentioned it a little bit.
607. 1:22:08>> This morning I I was testing our new model and I got a question. I got emailed a question that I didn't quite
608. 1:22:14understand. Uh and I put it in the model, this GPT5, and it answered it perfectly. And I really kind of sat back in my
609. 1:22:24chair and I was just like a oh man here it is moment and I got over it quickly. I got busy onto the next thing. But it was like a I
610. 1:22:34mean this what kind we're talking about. I felt like useless relative to the AI in this thing that I felt like I should
611. 1:22:40have been able to do and I couldn't and it was really hard but the AI just did it like that. >> Yeah. >> It was it was a weird feeling.
612. 1:22:46>> Yeah. Yeah, I think that's I think that feeling right there that's the feeling a lot of people kind of have like what's
613. 1:22:52going you know when does it happen? What's going to happen? Um but I think some of it is it's like yeah you it's
614. 1:22:59hard to conceptualize until you're further along. >> I I'm all to totally I don't think we know quite how that's gonna
615. 1:23:07>> but Sam actually answered it. He said he said he had this weird moment where he felt inadequate as a human and
616. 1:23:15then he moved on, right? Like like it's like that's the thing is like yes, we're going to confront these things that
617. 1:23:23we've never had to confront before and then we're going to go, "Oh, okay. Well, I guess I don't have to do that anymore."
618. 1:23:31And then we get to choose what to do next, right? Like I don't know feel you just have to like approach it step by
619. 1:23:36step. Another thing I'm afraid of, and we had a, you know, a a a real problem with this earlier, but it can get much worse, is
620. 1:23:46just what this is going to mean for users mental health. >> Um, there's a lot of people that talk to
621. 1:23:52chatt all day long. There are these sort of new AI companions that people talk to like they would a a girlfriend or a boyfriend. Um,
622. 1:24:00and we were talking earlier about how it's probably not been good for kids to like grow up like on the dopamine hit of
623. 1:24:06scrolling, you know, or whatever. >> Yeah. Do you think that that how do you keep like um AI from having that same
624. 1:24:13effect like that negative effect that social media really has had? >> I I'm I'm scared of that. I don't I
625. 1:24:18don't have an answer yet. Uh I don't think we know quite the ways in which it's going to have those negative
626. 1:24:24impacts. Uh, but I feel for sure it's going to have some and we'll have to I hope we can learn to mitigate it quickly.
627. 1:24:31>> Um, can AI can they pull up pornography and stuff like that too or No. >> Sure. >> Oh my god. >> God, I didn't know that.
628. 1:24:43>> The von talks a lot about sex addiction, porn addiction, and stuff like that. So, he just asked the question. Everyone in
629. 1:24:50America is hoping to know, can they pull up porn? Oh my god, that's funny. I wonder if brain implants will become
630. 1:25:06similar to getting braces. Probably. So the the way that uh Neurolink the robots that they've made to implant those things are are uh
631. 1:25:17you know um fairly automated. Um I have a strong suspicion that that uh human uh brain computer interfaces are
632. 1:25:28not going to require implantation for super long. >> No, it's fine. I Yeah, but I just Yeah, I don't even need to know that. I'm
633. 1:25:36gonna have that stricken from my own record. [Laughter] >> All right, we'll stop it right there. I'm gonna have that stricken from my own
634. 1:25:47record. Oh man. Well, so that's the that's about the first half hour of of a 90minute interview. So, I'm obviously going to go
635. 1:26:00watch the rest of that. You should, too. Um, I just like like it feels important to me. It it feels like a um
636. 1:26:11the the Today Show moment for AI. It It really does. That that moment where Theo Vine where where he goes, "Don't don't you think
637. 1:26:24we're just going to be growing babies in vats?" And then Sam Holman's like, "That's kind of weird, dude." and he's like, "Oh, I
638. 1:26:31thought you would like it." Oh my god, that's hilarious. Wow. Wow. Wow. Wow. Wow. All right, there you have it, man. That is that is
639. 1:26:47uh that's a good interview. That's a good interview. The Theo interview with Waqen Phoenix is also similarly good because
640. 1:26:58they're just both uncomfortable human beings trying to communicate. Um, and they they keep sort of blowing each
641. 1:27:08other away in terms of their humanity and they like respect each other's humanity. And this one is similarly
642. 1:27:15awkward where they're just in these two guys are in such different planets. They're like not on the they're not on the same planet
643. 1:27:27and yet they're trying to work it out. Yeah. It's fascinating. I saw a source camp said I she has a strong suspicion
644. 1:27:33they're going to become friends. Yeah, they probably will. And you know, I I I promise you this. I that that Theo vaugh
645. 1:27:41leaves the Open AI offices having played with AI probably more than he ever has and and maybe understanding it in a
646. 1:27:49different way. F absolutely fascinating. Chat GPT, you've talked all day. I have a headache. You need to go outside. Yeah, exactly.
647. 1:28:02Oh my god. Sam is much more human than Zuck. Yeah. Glad he's in the lead. Yeah. Every time I see one of these things
648. 1:28:10with Zuck, I'm just like, "Dude, [ __ ] get a spokesperson." Like Sam Sam is smart enough to know like like
649. 1:28:22he's at least smart enough to know that he's got to acknowledge the scary [ __ ] Zuck just doesn't. Zuck's just like,
650. 1:28:30"We're going to make a bigger computer. It's going to be good. we're going to put it in the metaverse. It's going to be awesome.
651. 1:28:36And it just he always seems disconnected from reality. Sam at least has the skill. He's at least smart enough to
652. 1:28:43understand he's got to acknowledge reality. Now, whether or not his company takes actions consistent with that,
653. 1:28:52that remains to be seen. Um, absolutely fascinating. Other thoughts? Should [Music]
654. 1:29:08be a very interesting autumn. Yeah, it's going to be interesting. I got to go watch the rest tomorrow. Pretty funny. Yeah, I know. Me, too.
655. 1:29:19Oh, careful with the corn word. Oh, you can't you can't say P O R N on Tik Tok. Really? You can't even say the word. Unbelievable.
656. 1:29:30Well, we won't have to deal with it after September. So, there you have it. Um, what do I like on Ronin?
657. 1:29:42What you like on Ronin? I don't know what that means. Um, now I need to go watch the rest of it. One more plug for B2B.
658. 1:29:55Oh, yeah. Yeah, yeah. Let's do so next week people. Next week [Music]
659. 1:30:13[Music] um do me a favor. So So next week I'm doing a fiveday back to basics AI crash course. Doesn't cost a thing. It's still
660. 1:30:26it's basically just like show up and I'll keep doing this [ __ ] except instead of movie night, we're gonna do five different nights.
661. 1:30:34First night's going to be what is AI? How does it work? Why is it a big deal? Why is it historically like something I've seen before? Um,
662. 1:30:46really get your head around what it is. Tuesday night is going to be chat GPT 101. Like what is chat GPT? What are the
663. 1:30:55different models? What can you do with it? Like like like I don't know if you knew this. Did you know that chat GPT
664. 1:31:01can see you can upload a picture of a whiteboard and it'll understand everything on the whiteboard including pictures and words and context.
665. 1:31:13Did you know that chatbt can talk? You can just have a full-on conversation with it. Um, did you know that it can
666. 1:31:20use tools that it can that it can do multi-step reasoning? Reasoning people will tell you that it's not actually reasoning, but it can do
667. 1:31:29multi-step internal prompting, let's call it. Um, and it can use tools and write Python and and and do all sorts of
668. 1:31:36crazy [ __ ] like that. So, that's Tuesday night. Wednesday night is going to be a creative night. We're going to be
669. 1:31:42looking at creative tools and creativity and making images and making video and making music and making sound effects
670. 1:31:50and voice synthesis and all that sort of [ __ ] Thursday night's going to be we bring it all together and solve like complicated use cases
671. 1:32:00that require multiple tools coming together. And then Friday night's going to be just an open-ended Q&A to talk
672. 1:32:08about all the stuff that we've learned during the week. So, here's my request. My request is go to YouTube right now and subscribe to the AI
673. 1:32:21Learning Lab. So, it's learning lab-ai. I didn't I wasn't quick enough to get AI learning lab. So, oh, and by the way, the the crash course
674. 1:32:32next week is free. It's free to subscribe to the AI Learning Lab. The course next week is free. I think that
675. 1:32:40the course we will probably make available on demand in the future, but if you just come next week live, you get
676. 1:32:47it for free. Okay? Um, but go subscribe to YouTube now and then next week the sessions will all start at 8:30 p.m.
677. 1:32:55Mountain time Monday through Wednesday. Um, we'll have to tell my seniors to tune in. Yes. Yes. Bring your seniors. So
678. 1:33:06that's my other request. So the first request is go subscribe to the AI learning lab uh on YouTube. The second
679. 1:33:14request is tell people tell people in your lives like I know that there were some people in here earlier who are you know AI haters.
680. 1:33:24Come as an AI hater like at least here's my request. Here's the thing about this channel. What you will not get what what you will
681. 1:33:34get on this channel is a lot of optimism about the possibility of AI. What you're not going to get a lot of
682. 1:33:42from me is is the technology itself good or bad? Is it ethical or not? Was it trained illegally or not? Those are all
683. 1:33:54really important debates, but that's not what this channel is about. What this channel is about is this is a profoundly powerful technology
684. 1:34:04that's not going away and it's getting better dramatically fast. And if it's not going away and if it's
685. 1:34:15evolving incredibly quickly, then you only have two choices. You can learn about it or not. And there's a lot of people right now
686. 1:34:24choosing to not learn about it. It takes too much energy. Eh, it hallucinates. It's stealing from artists. Eh, and
687. 1:34:32they're sitting on the sidelines with their arms crossed. My deep deep request is let's get as many people as possible who are sitting
688. 1:34:43on the sidelines to at least understand what it is and what it makes possible. Because what it is when used well is
689. 1:34:54it's an amplifier of you. It's really good at taking your ideas and getting them out of your head and into the world quickly and more
690. 1:35:07powerfully and then expanding those ideas and then amplifying those ideas and then spreading those ideas to different outputs.
691. 1:35:16It is incredibly empowering. It is incredibly inspiring. But you literally cannot understand that without using it.
692. 1:35:32So here's the good thing. You could come next week, start the week hating AI, learn all about it, and then at the end
693. 1:35:40of the week still hate AI and never use it again. Cool. But if you're just on the sidelines assuming you know what it is, assuming
694. 1:35:51you understand its motivations, it doesn't have motivations. It's just a reflector. It's just reflecting you.
695. 1:36:00Whatever you put in it, it reflects back at you in this amplified, magnified way. So my request is tell people about this.
696. 1:36:11Let's get some people here. Okay? Um, so that's my second request. Here's my third request. My third request
697. 1:36:24is join the AI salon. So the AI salon is a place where we can continue the conversation after these nightly things.
698. 1:36:32We can share things. We can talk to one another. But you can also meet people. Connecting with other people in the face
699. 1:36:41of all this change is probably the single most important thing you can do. Yeah, you got to learn AI tools. Yeah,
700. 1:36:49you got to get fluent. Yeah, you got to become ready for AI. All that sort of [ __ ] But it's connecting with other people
701. 1:36:57who are thinking about this stuff thoughtfully, creatively, ethically that will open a whole new world for you. So, if you go to
702. 1:37:09community.thesalon.ai or or do the little uh QR code there, go join. All right. Beautiful. Beautiful people.
703. 1:37:23Um, shout out to Gwen uh Shafitz for um for co-hosting the AI readiness project with me me today. Great job, Gwen. And I
704. 1:37:34don't know if Mr. Kay is here, but Mr. K was our our special guest today and he did great. It was a it was a really good
705. 1:37:40a good episode. So if you want to see that you can go to um airelinesspro.com and you can see the um
706. 1:37:52the recording of it. It was really good. So anyway, so thank you Gwyn. All right, I'm going to get on out of here. Uh I
707. 1:37:59hope you had fun tonight watching Theo Vaughn. I think this is the last 40 years where we'll have babies in human bodies. Sam Alman, huh?
708. 1:38:15Uh Danielle wants to know if I played with Pika Social yet. I played with it a little bit last night. Mostly just
709. 1:38:20flicking through other people's stuff. I I haven't uh made my own things yet. So, I have it. It's installed. Code code
710. 1:38:29worked. Thank you. Uh, and then uh we will uh I I'll play with it and then I'll I'll form some opinion on it sometime this week. So anyway,
711. 1:38:40all right everybody, uh hope you had a good time and I will talk to you uh tomorrow. Tomorrow is Thursday. Yeah,
712. 1:38:47today's Wednesday. All right, peace out everyone. Bye-bye.