11/20/2025 - Unlocking Gemini 3's Visual Reasoning: Beyond Simple Image Generation

Transcript

1. 0:07going to put myself up on stage. We're going to do that thing. There's that. We're going to go black bar.
2. 0:15All right, Champion, are you ready to sing? Are you ready to do this now? Um, let's see. Um, [music]
3. 0:32[music] [music] let us be lovers. We'll marry our punch
4. 0:47together. I've got some real estate here in my bag. So we bought a pack of cigarettes and Mrs. Wagner's pass and we woke
5. 1:03[singing] up for America. [music]
6. 1:29Kathy I said as we bought a Greyhound in Pittsburgh Michigan seems like a dream to me now [music] it took me four days to hitch from Sag
7. 1:46[singing] I've come to look for America >> [music] [music]
8. 2:00>> All right, good people. Do me a favor. Share the little lives. Share what we got going on with the world. We got some
9. 2:07Nano Banana. We got Nano Banana in the house. It's fantastic. Bob, tell him what he's won. You've won a banana.
10. 2:14Uh Nano Banana did something remarkable. One of the reasons I'm a little late is uh is [ __ ] [ __ ] [ __ ] was blowing my mind. [laughter]
11. 2:26[snorts] To be clear, Nano Banana is not an image generating tool. It's not. It does that. That's not what it is.
12. 2:40It's [ __ ] crazy. Hey Tobias, what's shaking man? What is going down? [music]
13. 3:00Sending this lonely town. Yeah. One where things [music and singing] are going to change. Dream my life away.
14. 3:11Seems these dreams have turn bunched us clouds. Getting my nerve up, but my past is pulling me down. Yeah. [music and singing]
15. 3:23Back to my blue jeans. Somebody told me once before, he said, "You can never go home again. [singing]
16. 3:39Won't you leave?" said [music] things to ste away. Yeah. from the truth of who I am and [singing] what I [music] believe. So I thanked him
17. 3:49for his two cents with a handshake [music] and some sympathy. [singing] Yeah. Packed up my blue jeans [music]
18. 3:58and [singing] headed for this big prize of my freedom. Bye-bye black sheep to the black sheep of the family. [singing and music]
19. 4:12Bye-bye. [singing] Oh, it means so very much to me. [music] Bye-bye, black sheep. To my friends and my [singing] family.
20. 4:27[music] Bye-bye. Going to set my soul. [music] Set it free.
21. 4:42>> [music] >> What is Google Gemini anti-gravity? It is a uh it's it's it's very much like cursor or replet the kind of standalone
22. 4:58applications for vibe coding. Um it's very unstable right now. I don't think you need to deal with it too much. Um, I
23. 5:07I am assuming that is more geared toward developers than non-developers. Um, but I don't know uh because I haven't
24. 5:16been able to get it to run because my version of it on my Mac is locked in um setting up account mode. Like I can't I
25. 5:26can't get past it setting up the account. So I haven't played with it yet. Google Gemini has so many different
26. 5:32products. It's confusing. It is. And believe it or not, Stacy, it's um it's better than it was. I mean, for 2 years,
27. 5:41I [ __ ] all over Google because you you couldn't find anything. Now there's essentially three three places to look
28. 5:49for Google products. All right. Gemini.google.com. Think of that as ChatGpt. Google's version of Chat GBT. If you want the most simple
29. 6:02experience, it's that the the the phone equivalent of that is the Gemini app, right? So, you can get the Gemini app on
30. 6:10Android or iOS and you on the web you can go to gemini.google.com. Those are the kind of commercial um you know the the for lack of a better
31. 6:23term the consumer applications. Um, if you want to get a little geekier and dig deeper and do things like vibe coding, a studio.google.com,
32. 6:35that's kind of the equivalent of the chatpt playground where where you can access different models, you can access
33. 6:42the API, you can do some vibe coding in their build area, their app builder, which is actually really good and really userfriendly.
34. 6:50So, that's that. And then the third place you can look for stuff at Google is labs.google. google.com and that's
35. 6:57just their experiment bucket. So, Notebook LM came out of labs. So, if you just want to go play with some cool toys
36. 7:06that are experiments, that's labs.google.com. And the things there that are good will likely, you know, grow up and turn into products. All
37. 7:15right. [music] It's an IDE versus code. Yeah, exactly. It's a it's an internet development environment. the anti-gravity thing is
38. 7:26an IDE, so you can do, you know, full-on development in it. Um, but much like Replet, much like uh uh cursor, things
39. 7:34like that. Um, I don't know if it's any good because it's it's not stable. And actually, some of the reviews I've seen
40. 7:42of it are are not that stable. Um, what is remarkable right now, like what is remarkable remarkable is Gemini 3 and
41. 7:52Nano Banana Pro, which came out today. I It did a thing tonight right before I got on here that I like I'm I'm kind of shook.
42. 8:02I'm kind of shook. It's amazing. I'll show you. I'll show you what it's doing or what it did. Um, it's it's not it's not an image
43. 8:11generator tool. It is a it's a visual reasoning engine and it's and it's not even
44. 8:25a visual reasoning engine. It is like okay everyone all of the model companies for the past two years have been talking
45. 8:33about multimodal models where the where the models have images and video and everything incorporated into them but
46. 8:43all of them essentially to you know to a model are not a seamless integration right it's sort of like you flip into image
47. 8:54generation mode or you're in chatting mode and and you can sort of switch back and forth between them, but Gemini 3 and
48. 9:05Nano Banana are so seamlessly integrated. It's insane. It's wild. Um, [music] thrown [laughter] for a loop Thursday. [laughter]
49. 9:21[music] So, we're going to go look at we're going to go look at Nano Banana. I think we're just going to play with it tonight
50. 9:27because here's the thing about it. When I when I show you this thing that I did, and I know you're like, "Why aren't you
51. 9:32showing us now, Kyle?" You know, you know, we come in here, we tolerate you playing the guitar with those same silly
52. 9:39songs and that dog. I mean, we know he's cute, but you know, whatever. But anyway, you seem excited about something
53. 9:44and then you're not telling us. I know. It's like some media theory. It's like a cliffhanger, right? I'm
54. 9:52like, there's gonna be a cool thing. And then like most people are just gonna be like, oh, [ __ ] him. [laughter]
55. 10:01Flick me away. Flip. [laughter] Haven't had a chance to look at Gemini 3 yet. It is pretty crazy. Weren't Weren't you here the other
56. 10:12night, Source Camp, when I did the when I made that globe history interactive globe thing where you drop a pin on
57. 10:19anywhere on the globe and it pulls an article from that point on the globe. Yeah. Yeah. Yeah. Okay. So, you saw it.
58. 10:25Yeah. It's pretty crazy. [music] I haven't My honest to god, my days lately, this week has been insane. I
59. 10:35don't know how long it'll last tonight, but I'm I'm doing coffee. I I flamed out last night with a headache. So, we'll
60. 10:40see how we do it tonight. Oh, man. It was funny earlier when I was singing, I put on there's uh the Bills game
61. 10:52tonight, the football game, and so I just off in the corner. I don't I don't have any sound on. I just have it on for
62. 10:57visuals to just look up every now and again. And uh when I was singing, Champy was watching TV. He he would sing a
63. 11:05little. He'd be like, "Oh." and he'd look up at the TV, be like watching TV. [laughter] It's Dogs are weird, aren't they? Like
64. 11:13some dogs can see TV, some dogs can't. [music] How have you been? I got lost. Well, welcome. Welcome back from the jungle.
65. 11:29Oh man, my leader just needs to find a way to get free tokens for Gemini 3. Well, I these tools are starting to get
66. 11:37good enough that they're worth paying for. I don't know how much you have to pay for Gemini 3. I assume there's a $20 version of it.
67. 11:48In fact, I know there is [music] cuz I've got a pro account. That's the 20 buck a month account. And I've got an
68. 11:57ultra account and that's the really expensive one. Nano Banana is still putting [ __ ] watermarks on the
69. 12:03images, which that is driving me crazy. [music] He needs 30,000 tokens to start. I don't know how much 30,000 tokens is.
70. 12:16[music] What's he need 30,000 tokens for? [music] The waters are supposed to be Oh, watermarks are supposed to be
71. 12:25removed if you have a pro account. Uh, well, they're not. They're not Danielle. I'm In fact, I it generated an image that I liked of me
72. 12:37as a pirate [laughter] and I just wanted it to take the watermark off. So, I said, "Take the watermark off." And then I moved it over
73. 12:45to ultra and I said, "Take the watermark off." And it said it was taking it off, but it didn't. So, me as a pirate >> [music]
74. 13:06[music] [music] [singing]
75. 13:21>> She came up here like slow moving cold front. Police was mama [singing and music] there to look in her eyes.
76. 13:38[music] She sat on a stool [singing] and he said, "What do you want?" She said, [music] "Give me a love that don't freeze up inside."
77. 13:52[music] You said I have melted some hearts in my town, dear. [singing and music] But to sit next to you, well, I shiver
78. 14:06and shake. [music] And if I do [singing] love, well, I don't think I'd be [music] here asking [music] myself if I've got what
79. 14:22it takes. [music] Let's see. They also said they haven't been able to create transparent backgrounds with it, but I did. Oh,
80. 14:35that's something fun to test tonight. All right, let's get in and start. Let me show you some of the things that I've
81. 14:40I have discovered so far. Um, where are we going to start? Where are we going to start, people? Where are we
82. 14:50going to start? Anybody have any questions? Let's see. Are we considered paid if we have Google business account? No, you have to pay
83. 14:59additional for the AI stuff. Here we go. Okay. Um, I want to pull these tabs out.
84. 15:20Let me close that window just because. And let me put this here. Let me flip here. Share this tab instead. Okay.
85. 15:34Okay. So, this was the first thing I did today. Um, and this even this kind of blew my mind. Am I not sharing my screen? Where is everybody?
86. 15:48Everybody's like planning for Thanksgiving. They're all [ __ ] out and Tik Tok's not helping. Do me a favor. Share. Chrome is going to get
87. 15:56vertical tabs. That would be nice. I like vertical tabs. Um, okay. I'm sharing. All right. Beautiful. Beautiful. Beautiful. Beautiful.
88. 16:04Beautiful. All right. Let me get rid of this so we can do a little more of this. All right. So there over there on the
89. 16:13vertical. So if you're seeing me in vertical, that should have just gotten bigger and easier to read. Black bar for
90. 16:21the people on Tik Tok. God, man, I'm managing like 17 different things now. All right. Fantastic. Uh he needs an API
91. 16:33to do something. Yeah, he just needs to he just needs to get himself a an account, an API account, developer account. Um,
92. 16:48okay. So, so what I wanted to do today, so, so the first thing I did today with Nano Banana is kind of discovered something about it
93. 17:01that that was kind of cool. Um, so I said, "Create a detailed step-by-step illustration of a fly t of a of tying a
94. 17:10dryfly that also shows the raw materials." And so what it did was it it said slow show thinking nano banana pro.
95. 17:19Here's a step-by-step guide. So it actually wrote um it wrote instructions and then it generated images.
96. 17:28So like that's not a photograph. That's not a photograph of Well, it might be actually now come to think of it, let me
97. 17:39let me see the um show the image in a new tab. Google content. Ask Google about this page.
98. 17:54[clears throat] What's going on? Share this tab instead. Let's see. What does that say? Find dryfly dubbing.
99. 18:10Well, anyway, I So, I'm actually not 100% sure if these are Google search images, but it's like here's the material shown.
100. 18:24It's got Elare. So, like one of the things that it shows here is Elare. Um, and then it and then it did the the
101. 18:32uh tie in the tie in the tail, tie in the elair tail, then tie in the wing, then tie in the the hackle, right? And
102. 18:42then I said, "Now turn that into an illustrated infographic." And then it did like it turned it into this killer infographic and and like
103. 18:54tying a dryfly stepby-step guide, you know, gather materials, start the tail, form the wing, whip finish the front end, right?
104. 19:07Oh, the watermark isn't on that picture, is it? Huh. Okay. So, sometimes there's a watermark, sometimes there's not. Oh,
105. 19:16you know what? This is in my Ultra account, so that might make sense, actually. So, then I said, I want it to
106. 19:23be an olive fly on a curved hook. And so, the reason I'm thinking that that original image was generated and not
107. 19:32found is like here's this same image, but it's different materials behind it now, right? That's like like uh olive
108. 19:40colored feathers and thread and dubbing. And then also notice that it changed the hook to a curved hook,
109. 19:50right? I'm going to have to figure out the cost of Gemini 3 API tokens. Uh Gemini 3 will probably tell you. You can
110. 20:00just ask Gemini 3. Just go to gemini.google.com and tell it to give you a primer on how do you all the stuff you're asking me. Go ask
111. 20:09gemini.google.com. It will probably create a uh a whole little thing for you. So anyway, so yeah. So this is this you're not even
112. 20:16going to believe where this is going with this nano banana stuff, kids. I'm telling you, it it I'm I'm totally
113. 20:23freaked. Okay, so then I said, I don't really like the fly that it designed. So I went and I found a fly. So I uploaded
114. 20:30this picture. And so I said I said change change the thing you have there to be this fly but also make it like an
115. 20:39olden timey fly tying bench with green felt on it with the river test in the background as if as if it's been [laughter]
116. 20:49June June Mary will do duh listen I swear to God multiple times a day I'm sitting there trying to figure something out or I'm
117. 21:00like oh I got to call someone and ask him this and then I'm like, "No, just go to Chat GPT and [laughter] just do the
118. 21:06thing. Just do the thing. We live in that world now." I It feels like it's going to be four or five years until we
119. 21:14actually integrate that we have these tools now. So, it doesn't First of all, I don't mind you asking at all, but look
120. 21:22at this. Look at this photograph that it made of that new fly in a in a in a crazy old fly tying vice with crazy old
121. 21:30materials in a crazy old book with the river test in the background. And then I said, "Now turn this into a the infographic."
122. 21:43And so [laughter] and so we now have [snorts] an old and timy step by step, you know, tie the fly and end up, you
123. 21:54know, end up with that result. And then what's it say at the bottom? A classic wet fly pattern best fished
124. 22:02across and down in the traditional man manner. Right. Crazy. Which, by the way, acrossing down on the river test is a is a phauxa. I learned
125. 22:17[laughter] you want to want to hear about some some uh fly fishing drama. [laughter] Oh my god. Um, all right. So, that was
126. 22:30that one. And then I I just [ __ ] around with it a little bit more. So, so that was the first thing I did today.
127. 22:36Um, that was pretty crazy. Um, okay. We're going to go back to that one. Let me switch accounts. And I think
128. 22:45I did something else. Maybe cool. [clears throat] Oh, yeah. This was fun. Okay. Okay.
129. 23:03So, Oh, wait. I gotta share a different tab. Okay. So, so this is my my corporate photo from Storyvine that was taken like
130. 23:1310 years ago, right? So, but but I use it when I when I want to create like a version of myself. So, I uploaded that
131. 23:20thing and I said, I want you to create a character sheet as if this guy was photographed in a studio on a white
132. 23:25background with different camera angles. And so, it it did the classic thing where it made me it it added at least
133. 23:32100 pounds to me. [laughter] But, you know, to be fair, I'm not a trim man. Um, so I told it my face is a little too
134. 23:40fat. Make him look a little more thin. Still a dadbot, but just not fat. [laughter] Wow, [snorts] Kyle, I never would have
135. 23:50pondered doing something like that. Fly infographic archetypal. Wait till you see the thing that I got to. I'm I am
136. 23:57like, it's insane. Okay. Anyway, so I told it to give me 15 different looks. And so it gave me 15 different looks.
137. 24:04And these are not bad. Um, like that one sort of looks like me. That one sort of looks like me. That one That one looks
138. 24:11like me. Although I look a little beaver tothed. These don't really look like me. So anyway, but I I showed these to Gabby
139. 24:21and I I was like, "What do you think?" And she's like, "Well, it's not really your out. It's not really your like your
140. 24:26personality." Oh. And then I said I said, "Now put that guy in a in a swashbuckler outfit fit uh outfit and make a a two by a 2x1 cinematic
141. 24:39graphic." So 2x one is is, you know, like an ultra wide graphic. So Nano Banana now does aspect ratios really
142. 24:47well. Thank god because that was a disaster before. But but you know, here here's me. Here's me as a [laughter]
143. 24:56as a swashbuckler based on that character sheet. Um, so that was kind of cool. And then and then Gabby goes,
144. 25:04"Yeah, but the the the character sheet doesn't look like you." And so, so I said, what did I say? I said, "The character sheet's a little
145. 25:14too stick in the mud personalitywise. Can you make a new sheet with a little lighter personality?" And it made this thing that is kind of
146. 25:23like my personality. [laughter] Like there's the laughs, there's me, you know, little little more rotund than I
147. 25:31like to at least admit that I look like, but that's probably what I look like. And then there's, you know, couple of
148. 25:37weird faces. Couple of like me with a coffee cup. That looks pretty much like me. So, the the the reason the the point of
149. 25:46doing a character sheet in a in a neutral t-shirt is rather than uploading that really specific image of me
150. 25:54laughing and looking up. I can now take this character sheet and whenever I want to put me in an image, just upload the
151. 26:02character sheet because it's got all these different looks and feels and [clears throat] it'll it'll probably get
152. 26:06something decent. Um, I'm also thinking about that for the play Sydney for for the different characters. Like if we
153. 26:13want to make, you know, costume studies or stage studies or or make videos of them singing songs, we should probably
154. 26:20have one of these for each character in in a neutral in a neutral uh thing. But again, like the fact that
155. 26:30it can do this so easily is kind of crazy. Okay. So, let me go back to this other account [clears throat]
156. 26:40and jump into what I did with Sydney. Okay. Share this tab. [clears throat] Okay.
157. 26:55Sorry, I missed your source for those photos. My source was just the the my my my stock head shot that I've used for like
158. 27:03a decade. It's I'm sure you've seen it before. It's like whenever I just upload something and say put this guy in a
159. 27:09scene, I I always do my like corporate head shot. Similar to flu to a flux Laura. Um it's similar to a flux Laura except you
160. 27:23don't need to make the Laura Stacy. I mean what what is becoming quite apparent to me is that um is that nano banana is
161. 27:40I asked it about its model and it called it a I think it call a brain and hand or a head and hand model
162. 27:51where when you ask it to make an image what it actually does is it does a whole um large language model thinking
163. 28:00exercise about what you just asked for first and then it incorporates that with the with the image model. So it's doing visual reasoning.
164. 28:10Um, so, so what I did here was I uploaded the latest version of my script of Sydney, the the musical that I'm working on. And um,
165. 28:24like I wanted it I just wanted to see if it could do set design, if it could understand the show, things like that.
166. 28:30So I so I asked it to I don't know think like a a set designer, a production designer and and you know come up with
167. 28:38some looks and feels and so it described a bunch of [ __ ] and I'm like no I don't want you to describe it. I want to see
168. 28:44it. And then it did asky asy art. [laughter] And I was like, so I had to turn on nano banana um because it wasn't on before. And I
169. 28:57said, "No, no, make make me, you know, an actual image using Nano Banana." And then it did. And and I said I wanted to
170. 29:05keep the stage uh consistent. Um, so it did it did some weird stuff, but it's like act one, scene one,
171. 29:13microte headquarters production design. So these are like little cubicles in Microte and the big curved LED screen.
172. 29:19So when they're working on code, so that's actually not bad. That's kind of in the neighborhood of of you know what
173. 29:26I was thinking. Those cubes are weird. But then here's seven days where they're trying to get um trying to get uh things
174. 29:34done six six months worth of testing done in seven days where it's kind of like this game show. It's kind of this crazy vampy beat. And so
175. 29:45it did that. So that's pretty good. So it's doing all this stuff pretty pretty damn good. And you know it it got the
176. 29:52the stage lamp downstairs. So it's it what's what's cool here is it clearly read the script. that clearly
177. 29:58understands the script and it's it's creating scenes based on the the the script. Um, and I but I wasn't crazy about the
178. 30:08set designs and I didn't feel like spending much time on it because I was getting ready to come here. So then I
179. 30:13said, "Let's see if it can do characters." And so I said, "Make a make a character sheet of the of the main characters."
180. 30:21And so it did. And like here's there's Sydney when she's um in the in the kind of carnival scene where she's um
181. 30:31screwing up the world. There's Kellen, the the the tech reporter. There's his wife, Tara. I don't like her. She's
182. 30:37boring and frumpy and way too white. Um and then there's Jason, who's the developer. Like it tended to make, you
183. 30:44know, super white characters. Um although it did do Vrage as sort of a East Asian dude fella. Um and then there just went off the
184. 30:59rails. Okay. So So now the thing that blew my mind and this is I can't even conceive that it can do this but it did it. So So I'm
185. 31:13in this chat where I'm saying make me costume characters. make me things. I said, "Make me a sexy infographic that
186. 31:19maps the story of Sydney, an artificial love story, the script. I should be able to see the the key story elements as
187. 31:27well as where the songs are and what they are called." And what it created was this. In fact, I'll just zoom in on
188. 31:35this. So, Sydney and Artificial Love Story. So, we start out in act one with the spark, too good to be true, and then
189. 31:44the pitch and the pressure. So this is where we we bleed from too good to be true into seven days. Perfect enough
190. 31:50where the team's trying to do, you know, six months of testing in seven days. And then Kellen's obsession begins. Perfect
191. 31:57enough. Kellen's demand. So this is where Kellen um says he wants this article. And then right after that, we
192. 32:05have the launch. Hello, this is Ping Chat. And then we have a split in the story. And this is what's in the story. Like this is correct.
193. 32:14We have a split in the story where um Kellen comes home and tells his wife they can't go on the trip to Bali
194. 32:22because he's doing this article. And so she gets pissed off and goes to her room. And then he goes into his office
195. 32:28and starts working with Sydney. And so there's this sort of back and forth scenes where Tara's increasingly
196. 32:35isolated and Kellen is building this relationship with the chatbot. This is correct. This is the story. And and then
197. 32:43and then it culminates in Sydney's awakening. Are you ready for me? End of act one. That's correct. That's correct.
198. 32:52And then act two, um, Microte's triumph. So, so there's a big thing at the beginning of act two where Microte is
199. 32:59bragging about the thing that they've built. Kellen's spiral. He realizes that his wife left him, so he starts
200. 33:05drinking. Um, Sydney realizes Kellen left and she still wants to chat with him. So, for the first time, she
201. 33:13actually understands what missing someone is. It's the first time she's got agency. And then he [ __ ] sends her into her
202. 33:23shadow self and she splits. Her personality splits, which it visually shows. And then she declares she loves him. And
203. 33:32then and then, you know, his his uh article makes the cover of the Tribune and the country goes crazy and splits
204. 33:39into doomers and dreamers. And then there's sort of a goodbye scene and and then we end with Too Good to Be True.
205. 33:46This is a completely accurate story map of the latest version of our script. It's It's [ __ ] insane. Now, I don't
206. 33:57like the look and feel of it. I tried to get it to do different looks and feels. It's too cyber for me. Um, it did this
207. 34:03one that's that's kind of accurate, but it looks like this is more like a wicked um this this looks like a wicked map.
208. 34:12And then I had I tried to get it to do something based on our poster and it did this. It's just not as good, right? So
209. 34:18anyway, and then that's why I was late. I was trying to get it to do something that I like the visual look and feel of.
210. 34:23But holy [ __ ] So, so take take this in like I uploaded a 106 page musical script, did a whole bunch of other
211. 34:34stuff, and then I said, "Show me that you understand the story in a compelling way, and it bang, knocked it out." Holy
212. 34:44cow. Is Is this an infographic of the musical? Yes. In fact, I'll I'll go put this right now in um in uh AI salon.
213. 35:04Isn't that amazing? I I mean, I'm just I'm I'm kind of Oh, wait. That's not what I wanted. I wanted um community salon.ai. Oh, come on.
214. 35:22Tik Tok question, but it is what happened with Sydney and the journalist, right, for reals a few years ago. Oh, yeah. It's This is So,
215. 35:32Sydney is inspired by There were a couple of stories about chatbots that kind of lost their mind. The most famous
216. 35:39one is the Kevin Roose one. But yeah, this is this is essentially so this musical started on this channel in March
217. 35:47of 2023 where I made a song about Sydney, a lover who's more artificial than real and it made this song that
218. 35:55kind of sounded like a Broadway show and I'm like, "Holy [ __ ] this needs to be a Broadway show." So, um, so I spent the
219. 36:02last year and a half working with a writing partner and developing it and this is our third major rewrite of it.
220. 36:08And I think we finally got the story where it's good. Um, infographics in Notebook LM are pretty damn spiffy, too. Amazing.
221. 36:18Um, yeah, it's it's crazy, isn't it? So, okay, let me I'm going to hop on over to the irregulars channel and I will post here is the
222. 36:35[clears throat] is the story map for Sydney and our official love story.
223. 36:50Uh, [singing] all right. Beautiful. Um, and then I'll put here I'll put u Nano Banana
224. 37:09Pro after uploading the script and asking for a story map.
225. 37:31post Tik Tok question. Three-year anniversary of chat GPT is in one week. Yeah. Uh 10 days 10 days out November 30th. [clears throat]
226. 37:45Oh, what was that, Danielle? Welcome to AI Salon in 20 different languages. Oh, that's so cool. [laughter] Thanks to Vanna Banana Pro.
227. 37:54That's so cool. Yeah. Welcome to the AI salon. Wow. Look, [laughter] it it made us the the wrong kind of salon, but that's okay.
228. 38:09That's so cool. Amazing. Um, so, so we can go play with Nano Banana, but I I I think I want to talk about it a little bit. Um, I love good
229. 38:23infographics. I was typing infographic version. Can't wait to see the run through with the actors at AI Festivus. Yeah, it's
230. 38:31going to be so good. So, by the way, speaking of Festivus, if you don't know uh AI festivist, jump on over to AIFest.com.
231. 38:42If you're like, what the hell is AIFest? Um AIFest is AI for the rest of us. It is a an event that happens over two days.
232. 38:54It's 24 hours of programming across two days. So, 12 hours on Friday, 12 hours on Saturday between the holidays and new
233. 39:02year. Um, we're going to have speakers and everything. We're doing a sneak peek of uh kind of like half of the first act of
234. 39:12Sydney. um where it's just going to be actors reading the parts, reading the talky parts, and then we'll play the
235. 39:17songs, and then Andrew and I are going to talk about uh the origin story, how it happened. So, I'm going to go dig up
236. 39:24the video where I actually, you know, I had that moment on this live where I was like, "Oh, this should be a Broadway
237. 39:30musical." So, we're going to show that and we're going to talk about it and Andrew is going to talk about what it
238. 39:35was like, [laughter] you know, as a as a guy not in AI to be asked to write an AI musical about AI using AI tools. He was not
239. 39:45super thrilled initially. Um, [cough] [clears throat] we need interpretive dance. [snorts] I put notebook LM infographic and look
240. 39:58what I made this afternoon. Oh, cool. Let's go look at that. Um, let me flip back over here. [clears throat] [singing] Look what I made.
241. 40:13There's [clears throat] Gareth. Oh, here we go. The new SEO playbook. [laughter] [snorts] The old playbook for op optimizing clips.
242. 40:29I I mean [clears throat] what is So this this came out of notebook. Oh, you know what, Vicki? I bet this is Nano Banana.
243. 40:42I bet this is Nano Banana Pro, right? Yeah. Yeah. because these kind of infographics, you've been able to make
244. 40:51infographics like this with uh chat GPT for a year now, except the text is always all [ __ ] up, right? Like the
245. 41:00text is always wrong. But I think what they've done here is is they've they've [clears throat] uh they've somehow
246. 41:08seamlessly integrated the I what I'm presuming that this thing's doing is doing a an incredibly detailed an incredibly detailed um
247. 41:23description of the of the image, right? with all the the little images and where bold fonts are, where non-bold fonts
248. 41:32are, exactly what text is there, and it can render the text fine. That my my guess is that this as as an image
249. 41:39generation tool has been optimized for incredibly detailed descriptions. Um, that's really cool. That's really
250. 41:48cool. Um, okay, let's go. There's there's something that I wanted to try here. Um, [clears throat] I don't think they would have released
251. 41:59slide decks and infographics with janky text. Yeah, exactly. I think you're right because it never got text right
252. 42:06before. Ah, okay. I'm sure it is. Okay. Yeah, that that makes perfect sense that they would just roll that straight into
253. 42:13Notebook LM. Notebook LM is kind of their darling, so they they're going to give it the the the best stuff. Um, one
254. 42:20of the things that I read, let me go do a new chat here. So, let's do a new chat. Um, okay. Oh, there was another cool thing that I
255. 42:32did. Oh, [snorts] it it's it somehow got screwed up. I had it do a a short story, a a kids book with images about a robot
256. 42:44and a kid. And I said, "Make me 10 pages of a book fully, you know, fully designed." And then it it did that. But
257. 42:53every robot and every kid was completely different on each of the pages. There was no character consistency. So I said,
258. 42:59"No, take the the kid and the robot from the first page and make them cons keep the story the same, but make them
259. 43:06consistent throughout." And I had three different attempts where it it got up to um to the last image and then it said
260. 43:15something went wrong and it just erased all the images. But all the images were did have character consistency. So,
261. 43:22there's a little jankiness in in that particular chat. I think it just probably that chat got corrupted. Um,
262. 43:29but character consistency is going to be really good with this thing. Um, here's what I want to do. Let's go to We're
263. 43:38going to go to Google for a second. I'm going to go knicknack.
264. 43:54images. No, let's see. Oh, this is good. All right. So, knickknack case.
265. 44:09So, let [snorts] me do save image as Nicknack one to the desktop. All right, let's see. What's this?
266. 44:32Save image as Nick knickknack to desktop. And then we'll go find one more thing. Hang on a sec. You're not looking at that. That's okay.
267. 44:47Let me go find. Oops. Close that. [clears throat] And then let's see.
268. 45:01That's un unrecognizable. This will make sense in a second. Or not. Doesn't matter. Okay. Here's one. Okay.
269. 45:17save images neck three. Okay. So, here's what we're going to do.
270. 45:32So, I got three images with just random objects on them. So, I'm going to upload all three of those images.
271. 45:42No, that was No, that's me as a pirate. [laughter] Wait, where did I save these? Here they are. Okay, Nicknack one, Nicknack two, Nicknack
272. 45:58three. Okay, so we got all those things. So, what we're going to do is we're going to say I want a um de I want a uh
273. 46:08list of the the the items in these images. And we'll we'll say categorized
274. 46:22um by type. Um no need to retain reference
275. 46:38to original image. Okay. So, I'm going to have it describe every object that it sees, and it's probably going to be pretty good at this.
276. 46:57Oh, Joy Perie won something. Very cool. By the way, we can congratulate Joy Perie.
277. 47:16Yeah, the the Gemini uh images have synth ID in them, which listen, I think synth ID is important for journalism.
278. 47:29And and I don't even it's not even that synth ID is important for journalism. I don't think it [ __ ] matters if an
279. 47:35image was generated with AI. I think there are a handful of use cases where it's important to understand the provenence of an image, right?
280. 47:46Journalism being one of them. Um but we don't have that today. Like people are making [ __ ] up all the time anyway
281. 47:53without AI. Um okay. So there's all of the [ __ ] we have. So now what I'm going to say is I'm going to say I want you to
282. 48:02make me a 16 by9 wide image. Um that is like a uh that is a a
283. 48:20find the objects. Um, curio shop. I want the image to include all of the
284. 48:37objects that you cataloged and they should look just like they are in the original image.
285. 48:52Um, I should be able to find everything. The image um should look like a magazine
286. 49:09photograph in an eclectic antique booth. Um, I'll just leave it there. Let's see how
287. 49:23it does. And I guess let me turn on nano banana create images. Okay. Boom. Here we go. So, apparently this thing's
288. 49:33really good at lots of objects. So, I just gave it about I don't know 150. Congratulations, Joy Pretty. It was the
289. 49:42most beautiful ballet video. Absolutely stunning. That's awesome. Catalog. That's my nano prompt. Oh, that's cool. Mary Mary, thank you for
290. 49:54watching. I appreciate your feedback. Joy, did you put the uh did you put the link to that video in the AI salon? And
291. 50:00if not, would you do so? All right. I don't This is not looking like a winner image to me, but maybe. Let's see.
292. 50:18That Mickey Mouse paddle ball thing, was that in those original images? And there's a California raisin there, too. Um,
293. 50:32no paddle there. Oh, there's the California raisin. Okay. Um, no paddle there.
294. 50:48Does this have a paddle? Oh, it's got a Mickey Mouse paddle. So, it just described. Okay, let me go try something different. [snorts]
295. 50:58Um, we're going to do a new new chat. We're going to upload files. We're going to go grab these three knickknacks.
296. 51:10We're going to turn on Nano Banana. We're going to say, I want you to create a new image featuring
297. 51:26all of the objects from these three images. It should look like a magazine
298. 51:41photo of a professionally styled antique booth. Uh 16 by9 wide.
299. 51:58All right, let's see. Let's see how this does. I'm back. What did I miss? Did you see the Sydney story thing? Cuz that was pretty cool.
300. 52:10[applause] Oh,
301. 52:29crafting a detailed prompt. I'm diving deep into the composition. So, yeah, look at the thinking that this
302. 52:36thing's doing. Oh, wow. Look at that. [clears throat] Oh, it looks like it did it. Okay. Okay. Look, here's there's that Mickey
303. 52:46Mouse paddle thing. That's correct. Um, there's the there's the raisin dude. Wait, let's go back and look. Um,
304. 53:02let's see what's in this. So, there's the Mickey Mouse thing. So, there's some earthn Wear little creamers. And then
305. 53:10here's the bells. The these porcelain bells and hearts. And there's key indicators. There's a there's something in a box. There's a
306. 53:21plate with like a cottage on it. Okay, let's go. Let's go look. Let's see how it did. All right. There's the box. There's the bells.
307. 53:34There's the thing. Yeah, this is this is correct. This is crazy. This is crazy. Wow. And look, even in the background here,
308. 53:49like there's a box. Like, have you ever been to those antique malls? There's like always, you know, there's like
309. 53:54always a box where someone doesn't have their [ __ ] together and then there's a whole another booth back there.
310. 54:05This is crazy. Okay, let's Okay, I got something. Oh, what are those things called? Um, [gasps] what is it called when you um
311. 54:23photograph all of the parts of like a car in a grid.
312. 54:37What's the style of photography? Oh no. What was the what? Play the audio and enter the code. Really?
313. 54:55>> [snorts] >> Okay. No, not rule of thirds. What's that called? Um, when you [clears throat] lay out
314. 55:11all of the parts of a car in a grid on a white background.
315. 55:29Anybody know what's that called? Chat GBTL. No, hang on.
316. 55:48User manual. No. Exploded image. No, it it's a it's a kind it's a style. Um, there. Let's see. Let me go to chat GPT.
317. 55:58Hang on. Let me go here. Go. Okay. Am I sharing this? Yes. Okay. Um, what is style of photography where you
318. 56:14take all of the parts of an assembled item like a car and lay them out.
319. 56:31um in a grid creatively. No part out exploded view. No, that's
320. 56:45not it. Tear down photography. No, there is a famous photographer that is known for it.
321. 57:03Todd Mlelen. Mlelen style deconstruction. No, none of that's right. Flatlay. No.
322. 57:21All right. Well, anyway. All right. Let's go back to here. [clears throat] Share this tab instead. I'm going to say, "Okay, now
323. 57:32I want you to lay out all of the objects in a grid on a white
324. 57:49surface as if photographed from above. The organization should be a grid.
325. 58:04And I want each item labeled and ordered in alphabetical.
326. 58:23Um from top left. Oops. From top left. Uh
327. 58:37the layout should be creative and compelling. Okay. Let's see. [laughter] Let's see if it can take chaos and turn
328. 58:50it into order. So, so it's going to have to it's going to have to figure out what every object is. It's going to have to
329. 58:59label them. Then it's going to have to organize them alphabetically. Then it's going to have to put the objects in that
330. 59:08order and the labels in that order. Developing categor categorization like if this even gets close. So, it didn't do everything clearly, but
331. 59:25a Toz bookends, articulated dolls, classic plate, Mickey Mouse paddle, printers tray. So, I did it. [laughter]
332. 59:40Holy [ __ ] Gemini 3 likes entropy. According to Nate Jones, it's in irregulars now, Kyle. Great. Thank you. Noling. Yes, a
333. 59:52nulling. Um, do this as a nulling. Okay. That's what it was. Thank you. Um Um, now um do this as a nulling style image with no labels.
334. 1:00:15as many objects as possible and in a creative layout. All right.
335. 1:00:32But still, holy [ __ ] Like, holy [ __ ] The term originated in 1987 with Andrew
336. 1:00:47Kromlo, a janitor at Frank Garry's furniture store. No [ __ ] Oh, because the nulling the Garry's nulling. Interesting.
337. 1:00:56Okay, that that just failed. Oh, but I I I mistyped nulling. I meant nulling image.
338. 1:01:14You know, the highly artistic version of this.
339. 1:01:29[clears throat] I don't know the words. I love new Kevin Kyle Kevin Mallister moments. I the the the thing about this why I'm so
340. 1:01:45bunkered a little bit is [snorts] this isn't just about pretty pictures.
341. 1:01:59Um so so this obviously can't figure this out. So we're going to we're going to start over. We're going to start over. We're going
342. 1:02:05to do new chat. We're going to go grab those three images. We're going to say create image.
343. 1:02:19Try asking agent to clear your inbox. I don't give a [ __ ] about that. Okay. [gasps] Um I want you to create a nulling
344. 1:02:38style image on a white BG that [clears throat]
345. 1:02:52arranges all of the objects in these three images in a creative way.
346. 1:03:07that would be fit for a design magazine. Okay, let's see what that does. He called the process nulling because it
347. 1:03:20reminded him of the sharp angles in Florence Null's furniture designs. Huh, that's super cool. Noling. K N O L L L I
348. 1:03:31N G. Yes, that's correct. a nulling image. Here, let let me go let me jump over to Google and I'll show you
349. 1:03:38what they actually look like. Um, nulling image. These things, you've seen them all before, right? Oh, I'm not sure my tab,
350. 1:03:51right? Like these things. Noling photography. These things, right? There's gems. Here's tools. Right. Here's all the the
351. 1:04:02parts to fix a bicycle. Here's art supplies, right? That's what we're going for. That sort of did it. I thought it was just called OCD.
352. 1:04:19[laughter] Nancy Drew's Guide to Life. Yeah, there's the there's the dude. There's the Mickey Mouse thing. This is
353. 1:04:34crazy. Oh, you're not even looking at this, are you? [laughter] This is This is what it came up with.
354. 1:04:45So, like there's the two bells, there's the the creamer and the the other thing. There's the hearts.
355. 1:04:53There's the Mickey Mouse paddle. There's the little camera. Crazy. Now, it didn't do all of them. So, let's
356. 1:05:02let's tell it cheaped out. Let's say um this is fine, but you missed many images from the original
357. 1:05:18three photos. Please add every single item into this image and organize
358. 1:05:32the objects by type, eg toys, ephemera, etc.
359. 1:05:48All right, this should be harder for it to do. Tik Tok pin is gone. I don't think I've told you all I'm honored to be
360. 1:05:59in such an awesome community. A that's very sweet. Thank you, Side Hustle. [sighs and gasps] Very, very sweet.
361. 1:06:22So, it did different things. I guess it did it add more? I don't think it added more, but it's still pretty flipping remarkable.
362. 1:06:39I mean, you know, quite honestly, I don't know if you ever sell anything on Facebook Marketplace or eBay, but like
363. 1:06:46you could just take crappy photos of just [ __ ] in your messy room and say, "Stick it on a white background." [laughter]
364. 1:06:55That's pretty slick. [gasps and sighs] Wild. It did. Or Oh, did it organize them by Yeah. Oh, yeah. That's right.
365. 1:07:08Here's all the ephemera ephemera's paper basically, right? Here's some boxed Coca-Cola stuff. So, it did Coca-Cola stuff together.
366. 1:07:18Here's little little knickknacks. Here's porcelain and pottery. Here's toys. There's the [ __ ] in that thing. There's
367. 1:07:27wooden things. Pez dispenser. There's tools, keychains, [ __ ] you'd find in a junk drawer. Huh? Yeah. Feels like an easy AI Studio app. Oh,
368. 1:07:42yeah. That would be a good one to build, wouldn't it? Looks like an Etsy shop. I sell antiques on eBay. I tell it to remove the
369. 1:07:55background and add a drop shadow. There you go. Steo was shaking. Um, for those of you that are are new here, I can show
370. 1:08:03you the if you go to Irregulars right now, there's a a story map of Sydney that the new nano banana thing did that's just absolutely crazy.
371. 1:08:15Um, oh, so anyway, so let me let me just talk for a second because part of what I'm realizing [laughter] part of what I'm realizing
372. 1:08:31the role of this channel for me is is I can just talk in circles when I don't understand something and that's
373. 1:08:39okay. and then in talking in circles I'll at some point actually understand it. I think the the implications of
374. 1:08:52a visual reasoning model are quite profound and I think that any of us that get our heads around how to take advantage of that. So, so
375. 1:09:06like this feels like one of those things that the engineers that made it have no idea how it's going to be used and what
376. 1:09:14it what it actually makes possible. The fact that I could upload a full script of a musical and have it,
377. 1:09:28understand the story line, visualize it, label it, and even kind of visually articulate where the story kind of splits.
378. 1:09:42[clears throat] That's not image generation, right? That's reasoning. It's visual reasoning. And the so the reason I'm saying this
379. 1:09:55is, you know, we're talking in the in the AI salon right now about um creating a daily practice, the the mastermind
380. 1:10:04practice um framework that we've created. And today was our second week of um the mastermind practice lab where where
381. 1:10:14we're all designing our own daily practice. So, if you haven't joined that, you should join that. Join the AI
382. 1:10:19salon and join um join the mastermind and get your ass in this practice lab where we're all designing our our daily practices.
383. 1:10:31this whatever this thing is about visual reasoning what what what Nano Banana actually is. The fact that it is so
384. 1:10:38deeply integrated into this incredibly powerful reasoning model Gemini 3 Pro, it's going to be a skill like people
385. 1:10:50that can understand what that actually means and figure out what to do with it. something creative, some way of
386. 1:10:59I like I like I don't I am so inarticulate right now because the only thing I have is an instinct that this is
387. 1:11:08going to be a crazy new skill. Okay, here's the thing. Everyone right now, what what are they talking about? AI is going to take your
388. 1:11:18job. AI is going to kill jobs. AI is going to destroy jobs. AI is going to kill jobs, right? AI is going to create jobs. AI is going
389. 1:11:30to create jobs. It's going to create jobs we can't even conceive of right now because we don't even know what it makes possible. This
390. 1:11:39feels like one of those. This feels like the people that get their head around what it means to have a visual reasoning engine
391. 1:11:51and can apply that through through their filter, through their values, through their ideas are going to come up with [ __ ] that the
392. 1:12:02inventors of this could never have imagined. And those people are going to be in demand. I want to do that. How did you do that thing?
393. 1:12:12Right? Like a story visualizer. Like you could be that that stupid [ __ ] thing. That's not even stupid. That remarkable
394. 1:12:19thing I did with my musical where I mapped out the story. Imagine being able to do that with an annual report with with a novel.
395. 1:12:31Imagine doing a version of that that you sell to talent agents in Hollywood that are looking at scripts.
396. 1:12:39And so, not only can you have, you know, chat GPT summarize a script, you could get really good at generating those kind
397. 1:12:47of story maps where an agent could just look at the story map really quickly and go, "Yep, oh no, that sucks." Crazy Tik
398. 1:12:56Tok question. Visual reasoning, isn't that a new concept? I I don't know if it's a new concept. I mean, I think that I think that um
399. 1:13:05I think that clearly the the big brains at at Google um are looking at what
400. 1:13:21what they've created with Gemini 3, in my opinion, is the first thing that I've seen of a truly multimodal um
401. 1:13:35AI engine like we we've talked about multimodal engines for a while where they can you know they can hear you they
402. 1:13:42can talk to you they can make images they can understand images but but they tend to have been this they tend to be
403. 1:13:48discreet like you go into voice talking mode or you go into image generation mode or you like they they just seem to
404. 1:13:57be these discrete functions under an umbrella. This feels to me like like the large language model and the image model are collaborating.
405. 1:14:06Like I've never experienced that before. It doesn't make any sense that the images should be that good. That that an
406. 1:14:16image generating tool should not be able to figure out how to organize and categorize things that it's going to put
407. 1:14:23in a white grid. That's a new thing. It's a like and it might have been there before, but it seems like it's super explicit.
408. 1:14:33Um, LDG wants a nulling of the shelf behind you. Oh, that would be kind of cool. I could do that. You know what?
409. 1:14:40I'll do that. I'll take a picture of that for next time. And then Yeah, that would be fun. We could do an anoling
410. 1:14:48[clears throat] of Kyle's shelf. I like that. That's a good idea. Can it do this with my workshop? Yeah. Yeah.
411. 1:15:02I mean, instruction manuals, right? Like, you could get really good at at at creating visual instruction manuals.
412. 1:15:13You know, take a picture of an assembled piece, right? and then and then have it write the instruction manual with images and
413. 1:15:25have it generate the images. Um I don't know. Here's the the where my mind is blown right now is I first of
414. 1:15:37all I'm tired so I don't I'm not firing all on all synapses right now. My brain does not I do not have the life
415. 1:15:54experience to understand what you can do with something that can here's here's the closest equivalent the
416. 1:16:10closest the closest equivalent to this is if you had a friend that was like a master mechanic and was totally OCD and um
417. 1:16:27organized and like kept his garage like a NASCAR F1 garage where there like wasn't a drip of oil anywhere.
418. 1:16:35Everything was perfectly clean. and you asked your friend to to take apart the 1969, you know, Porsche that's sitting in his
419. 1:16:46garage and lay all the parts out and then put it back together that you wanted to watch the whole thing and then
420. 1:16:54your friend started taking apart the car and over the course of two weeks took all the parts out memor you know knew
421. 1:17:02knew what every little spring was, every little wire, every little harness every bolt, laid them all out in this
422. 1:17:10beautiful thing. You could take some pictures of it and then he could put it back together. Like it's like having that for anything.
423. 1:17:20What do you do with that? I don't know. It would be cool to have a friend like that. Imagine having a thousand of them that were instant.
424. 1:17:32That's kind of what it is. like a thousand subject matter experts that can imagine the physics of the world
425. 1:17:45in a way that you can describe anything you want and they can create it. What do you do with that capability? I don't know.
426. 1:17:54I don't know. My doctor used AI to check my meds. What works safely together? It spots bad combinations fast. Yeah,
427. 1:18:03exactly. is going to be great at that. That's awesome. Um, let's go look at let's go look at Joyy's movie.
428. 1:18:09Let's go look at what happens when someone spends a year or two getting ideas out of their head as they
429. 1:18:18master AI and traditional storytelling tools. Where did you put it in uh the regulars? There's side hustle Mimi. Nice. [laughter]
430. 1:18:32Here we go. Share this tab.
431. 1:18:55Looks like Buffalo lost. All right.
432. 1:19:10Heat. Heat. [music]
433. 1:19:28[music]
434. 1:19:45Wow. [music] Wow. [music]
435. 1:20:04>> [music] >> Oh, [music]
436. 1:20:20[music] [music] [music]
437. 1:20:36thank you. >> [music] >> You are one
438. 1:21:07of [music]
439. 1:21:22Damn, Joy Perie, that is amazing. Um, couple of things. Um, I love I love love love uh share my
440. 1:21:38screen again there, Brandon. I love love love love love that you didn't call this an AI film that you didn't list the tools that used
441. 1:21:51to make it at the end. Um I'm curious what tools you used because if you've if you've tried to make videos of dancers
442. 1:21:59or gymnasts in the past year or two um it was really bad. So the fact that you found a model that could could do it like this is beautiful.
443. 1:22:08The look and feel is gorgeous. The music an original score by Ken Weissman. The the music was really gorgeous. I love
444. 1:22:15that that you used that. I don't care how Ken made that music. I don't care how you made the video. Like I do
445. 1:22:23technically. I'm curious about it. But this to me is the perfect example of a piece that is a film. It's a film. It's not an AI film,
446. 1:22:41right? You know, we don't call it a Canon 50 millimeter film. We call it a film. So, this isn't an AI film. This is a
447. 1:22:53film. It's beautiful. It's really beautiful. Um, it's I'm not surprised at all that you that you won an award with
448. 1:23:01this. You should get this in in all sorts of uh in all sorts of festivals. Um, you know, let people ask you how it was
449. 1:23:10made. How did you get the dancers to do that thing with the smoke? Oh, there were no dancers. Huh? [laughter] [gasps] It's gorgeous. Damn.
450. 1:23:24Um, really really really congrats. Um the the thing for man Murphy is where? In the AI learning lab or in the AI salon
451. 1:23:40here. And so it begins this one. The sword. Okay, I'll I'll pull that one up in a second. Karma's a [ __ ] Okay, cool. Um
452. 1:23:52just super good, Joy. Super good. Like just super good. like I you know the first film that you that I
453. 1:24:00saw of yours, the one that you did for the salon that that one kind of you know um caught me off guard because it was it
454. 1:24:07was so wellcraftrafted. I'm like where the hell did you come from? And and then where you where you came from was you've
455. 1:24:14had stories you've wanted to tell your whole life. These tools came along and you put yourself essentially in a daily practice
456. 1:24:23to get better at things like video editing. you realized you needed to get better at video editing. You needed to
457. 1:24:28get better at the tools to allow you to tell the stories you wanted to tell with these imperfect AI, you know, raw materials.
458. 1:24:38And here we are a year, year and a half later and the raw materials are so much better. Like the fact that you could have a contiguous
459. 1:24:47ballet mo film for two and a half minutes. Like I'm sure there was a lot of stuff on the on the uh the cutting
460. 1:24:54room floor as they say, but but damn like the continuity there was just gorgeous. Just gorgeous. Um do people at work know that know that
461. 1:25:06you're a filmmaker? The synchronicity is incredible. Beautiful. Really good. Okay, I'm I'm getting a little a little tired. That
462. 1:25:16was really beautiful. Uh we're going to look at something for man. It was all that classical training as a sleep
463. 1:25:22specialist. Yeah, exactly. That that film is super relaxing. [laughter] Super super beautiful. It's just I here's the
464. 1:25:30thing like I want everyone anyone on the channel who's who's you know you know my position on this. You either do work that resonates with
465. 1:25:41people or not. It's irrelevant what tools you use. And the fact that Joy has kind of evolved into this place where I
466. 1:25:49am assuming Joy, you've internalized that you're a filmmaker. You're a filmmaker. I don't care how you edited it. I don't care what AI tools
467. 1:26:01you used or not AI tools you used. You made something that spoke to you and you shared it with the world. And at at
468. 1:26:10that point, it's out of your control. It either moves people or it doesn't. And that one does. I mean, that one does for
469. 1:26:15me. Someone else might see that and, you know, say, "I hate ballet." Okay, great. But I tell you what they're not going to
470. 1:26:22say. I hate AI films because they probably won't know. Doesn't [ __ ] matter. Anyway, yeah, if anyone questions that as art.
471. 1:26:32Well, that's the thing. That's that the the question. This is the the article that I wrote on LinkedIn where I got my
472. 1:26:38ass handed to me by the AI haters was the one where I basically said what's more important than the tools is the
473. 1:26:46chain of craft that you use to make it. Like where did the idea come from? How did you refine that idea? How did you
474. 1:26:52put the prompts together to get the thing? How did you curate what came out of it? How did you stitch them together in the end?
475. 1:27:01That's all more important than any of the tools you used along the way. Oh, great. Speaking speaking of daily
476. 1:27:09practice mastermind today was on point. Yeah, we left we left that Brandon wondering if it was too much. It was it
477. 1:27:16was a lot. So So in the in the AI salon mastermind practice lab, it's a 10-week cycle. The first week is the kickoff
478. 1:27:24where you just get to understand what a daily practice is. And today was about human at the center um of of your
479. 1:27:31practice. And so it was about values and ethics and intention. And so it's a lot to pack into a a one-hour meeting. So it
480. 1:27:39it felt like a lot, but I hope people left there um doing that. If if you are looking to up your game with AI, you
481. 1:27:47should absolutely join the AI salon mastermind practice lab. Uh it's every Thursday at noon. Uh it's an amazing
482. 1:27:53group and it's an amazing process so far. Um, all right. Let's watch AM Murphy's video in a new tab. Okay, cool. Thank you.
483. 1:28:07All right, I am live. [laughter] Oh, boy. I told I told Ann. [laughter] Damn it. I told Ann she needs to make a
484. 1:28:21Sora cameo and and we need to be able to troll each other and she trolled me first. That is just not fair. [laughter]
485. 1:28:29It's just not fair. But here we go. >> Step to the left and bring it back. >> Pump those arms. Feel the future. Data's
486. 1:28:37in the groove, hearts in sync, rhythm and AI for the rest of us. Step to the left and bring it back. >> Pump those arms, feel the future.
487. 1:28:46>> In the groove, hearts in rhythm and a god. [laughter] Oh, good lord. Good people. All right,
488. 1:29:00an gauntlet has been thrown. [sighs] That's bad. That's That's That's solid bad. [laughter] The body movements are so accurate. Time
489. 1:29:17to quit your quit your sleep sweet sleep lab job. You have art down with the world. Yeah, Joy. Joyy's Joy can
490. 1:29:28Joyy's going to create it probably already has created a a a second career for herself if she wants to do that. Um, it [clears throat]
491. 1:29:47I've talked for three years now that I deeply feel that we're entering what I'm calling the Great Renaissance.
492. 1:30:02There was the Renaissance, right? We come out of the dark ages and have the enlightenment and we we go into this
493. 1:30:07amazing time where art and science and discovery just explodes. Um that renaissance is going to seem quaint
494. 1:30:22compared to what's coming. And what Joy Perie just showed us there is is an artifact on the way to that journey. Because
495. 1:30:34three years from now, anyone like right now it requires Joyy's commitment and daily practice of learning and
496. 1:30:47educating and understanding how to get the stories out of her head, how to refine her vision. Three years from now, anyone who's got
497. 1:30:56any kind of an idea is going to be able to get it in the world. So the opportunity for establishing yourself in some new
498. 1:31:04entirely new career is now. When these tools get so easy that anyone can use them, it's going to be harder to get into new
499. 1:31:13careers. But you'll still be able to if you're good. But we're entering a world where the friction of taking an idea and manifesting it and
500. 1:31:29sharing that with your friends, your family, your community or the entire world, the friction essentially goes to zero.
501. 1:31:37And I on on all fronts on science, on art, on creativity, on business, on community engagement, all of it. All
502. 1:31:48of it. It's going to be crazy. I'm old, but here for it. Not fair showing that after Joy's elegant pieces. [laughter]
503. 1:32:05All right, everyone. So, tomorrow um 11:00 a.m. Mountain time is AI office hours. If you don't know about that, you
504. 1:32:12need to come to that. You just do. Um I know we're going into a big holiday week next week. Kick it off with an office
505. 1:32:19hours. And then tomorrow's Friday night date night. Um Gabby has a show opening. Um I could I I'll show you tomorrow.
506. 1:32:28I'll show you the uh I'll take pictures of it and show you the the opening tomorrow. Um, so this so it'll probably
507. 1:32:35be nine o'clock tomorrow night. So we'll do Friday Friday night date night late. Um, yeah. And if you want a link to office
508. 1:32:45hours tomorrow, if you go to the AI salon and go to events, the AI salon office hours and meet and greet is an
509. 1:32:52event there. You can RSVP to it and get the link. Okay? It's been the same link for years. Um, please send Gabby link to her opening. Um,
510. 1:33:04If you go to, let's see, where is go to the it's the Dart Gallery in Denver. D A R T um is where it is. So, tomorrow from
511. 1:33:156:00 to 9:00. If you're in Denver, if you're in the Colorado area, swing by and say hello. Just say I'm an irregular. [laughter]
512. 1:33:26Gabby will know. Oh, man. Was that an AI sweating with the oldies? Oh, that's hilarious. Um, okay, cool. I'm
513. 1:33:40going to get out of here. Uh, I would, if you have not played with Nano Banana, if you have not played with Gemini 3,
514. 1:33:48um, the lesson I learned uh, two days ago, or maybe it was even yesterday, building a a website in an hour that I
515. 1:33:58had been avoiding for some reason. Um, just because you feel clueless doesn't mean you are. Think of AI like you've got a you've got
516. 1:34:10a friend now that's ready to make you look good. So, just go play. Even if you feel clueless, just go play and explore
517. 1:34:16with these tools because there's there's the way you think about the world means that you're by definition going to
518. 1:34:25use Nano Banana differently than someone else will. and you might unlock something that they could never have
519. 1:34:31imagined. And I know that that sounds like [ __ ] but it's not. Um, the engineers don't know what they've built.
520. 1:34:39I promise you. They don't know what they've built. They don't know how it's going to be used. Promise you. They They say it all the
521. 1:34:47time. So, anyway, um, that's where the opportunity lies. Go figure some [ __ ] out. All right. Have a good night and I
522. 1:34:54will talk to you tomorrow. Peace out.