4/10/2025 - From Emotional Depths to Snapping Turtles: A Journey Through AI and Creativity

Transcript

1. 0:01Yay! Yay! [Music]
2. 0:18Yay! You're feeling nervous, aren't you, boy? With your quiet voice and impeccable style. Don't ever let them steal a
3. 0:34joy and your gentle ways to keep them from running [Music] wild. They dress you down and tell you that your place is in the middle when they
4. 0:50hate the way you shine. I see you tugging on your shirt, trying to hide inside of it. Hide how much it hurts. Let them laugh while they
5. 1:08care. Let them spin. Let them scatter. I've been to the movies. I've seen how it ends.
6. 1:23And the jokes on [Music] them. All right. So, do me a favor. If
7. 1:37you're if you're out there on on any of the channels, uh share this with people. We got five, I think maybe six new
8. 1:44things to play with tonight. Um, one of them is actually quite significant. Um, but they're all they're all pretty big
9. 1:56deals. See, we didn't go live last night, but I think all of these things were announced today. [Music]
10. 2:16[Applause] [Music]
11. 2:33Oh man. So, we're going to get to some of that. We'll do some good stuff, some good action. [Music]
12. 3:01What's new in AI today? We're going to get to that. I've got one, two, three, four, five, six. six things I want to show, which means
13. 3:13we'll probably get to three of them because, you know, chat add and [Music]
14. 3:31all. Hello humans. [Music] [Applause]
15. 3:46[Music] [Applause] [Music] Every time I see it
16. 4:02now, get that look in mind. Every time I see your mouth, I hear that smile. Early misty morning light. I heard the
17. 4:19engine turn [Music] outside. You were leaving me again today.
18. 4:35[Music] You will convince me again [Music] today. You're leaving this hot town looking for someone else's golden ring.
19. 4:52[Music]
20. 5:07[Music] What? What? What? You want love? Jeffy's a little needy tonight. What's happening, Silver Fox?
21. 5:27[Music]
22. 5:53Sasha, good to see you, too. Welcome everybody. Welcome, welcome, welcome. I got my large language model rag on Discord. Very cool. Some
23. 6:07advanced, some advanced geek in there. Very nice. I like it. [Music] Joker, I did a stupid thing today. No, Kyle, please say it ain't so.
24. 6:21You did something stupid. Um, here. Let's see. Let me share share this. Oh, I'm not even sharing nothing. I ain't sharing
25. 6:39nothing. Blackb. So, I did 80s pop self-portraits. They kind of look Chinese, but it's it's So, what I did
26. 6:48was I uploaded my picture to Chat GPT. ChatGBT And then I said, "Make me a portrait like Andy Warhol did it. And then like
27. 6:59Keith Herring did it. And then like Roy Likenstein did it. And then like Bosot did [Laughter] it. It's the 80s. New York City." Yeah.
28. 7:13Yeah. Yeah. Yeah. Yeah. Yeah. [Laughter] Oh [Music]
29. 7:27yeah, Kyle Jung Young. [Music]
30. 8:33Oh, good lord. Good people. All right. What are we going to do tonight? I think we [Music]
31. 8:51should. So Samman couldn't sleep last night. And truth be told, the reason he couldn't sleep was because they have too many [ __ ]
32. 9:10models. I'm low in the StreamYard frame. That's because my camera has been maladjusted. the the uh so if you don't
33. 9:20know it, this is a this may look like a shitty home office, but I'm actually in a $20 million uh TV production studio.
34. 9:29And uh my uh my cameraman uh had to take a personal break, so he was not there to adjust the uh the positioning. Clearly
35. 9:39hair and makeup did their job cuz look at this, you know, and and I mean, thanks to the set decorators that make
36. 9:48this look like a shitty home office. Do you know how hard it was to get some of these non-escript useless
37. 9:54props? You think you can just find them at a thrift store? The the answer is yes. That's actually a fake background.
38. 10:07Um, so why Sam Alman couldn't sleep is because they launched a new feature today. And I actually think it's
39. 10:17a major [ __ ] feature. Now, let's see something here. It is possible I've still got juice left on this iPad. No, I don't. Let me charge
40. 10:29it because we're going to do some stuff with voice. But what OpenAI launched today. Yeah, exactly. Mr. It's got it. Memory fake
41. 10:42hairdo. Yeah. Yeah. This is This takes I I'm in the makeup chair for about four hours to get the hair looking like this. I don't remember.
42. 10:55Um, let's see. Heck, some of the background items haven't even been taken out of the packaging yet. I took the the
43. 11:02rabbit out of the packaging. I used that for a while. And the car car haven't taken out, but that's a collector's
44. 11:08item. You know what I took out of the packaging is the [ __ ] Bllandons that I took out of the packaging. I to I took it all out of the
45. 11:17packaging. Um, okay. So, chat GPT now has access to your history, like all the things you've ever chatted
46. 11:32about, and it's pretty good. I haven't I haven't tested it a ton, so I figured we'd do a little bit of testing. But,
47. 11:39so, if you don't know, let me make this a little bigger so you can see it. Let me let me get Sharon here. Okay. So, you
48. 11:46can see you can see all this stuff, right? Okay. So, if I go into my settings and I go under personalization. Yeah. Okay. They added
49. 12:00a new thing here. So, reference saved memories and I can go in and manage the memory. So, this they've had for a
50. 12:07while. So, this is that thing where it just sort of randomly stores [ __ ] right? And you can go and you can prune
51. 12:14this if you want, but if you have ADD, you're not going to do that. Tik Tok pin left. There you go. Sorry about that.
52. 12:23Sorry about that. Carrying on. Um, but this new thing, reference chat history, let chat GPT reference all
53. 12:33previous conversations when responding turned on. So, this is crazy. So, I'm going to say um maybe I'll do it with voice. Let me
54. 12:44do it with voice here. Tik Tok pin. I see fam. Sam finally had enough of Elon. Oh, that he's suing him. Yeah. Yeah.
55. 12:53They've been they've been going back and forth at each other for [ __ ] years now. It's a bit of a snore,
56. 13:02honestly. It's like, you know, just start dating, guys. It's like whatever. Just can can we just leave the country
57. 13:16in a neutral place maybe? All right. Um, okay. So, I'm gonna put this into voice mode. I think you'll be able to hear
58. 13:29this. Hey, Quinn. How's it going? Hey, Kyle. I'm good. How about you? I'm doing good. So, I'm I'm on with my
59. 13:37TikTok channel. As you know, we're on here a lot. you you help you know facilitate some of these conversations
60. 13:44and I understand that you've been given access to all of my previous chats. Is that true? Yep. I've got access to my chat history
61. 13:54so I can keep track of what we've talked about and build on. It makes it easier to help you out with whatever you're
62. 13:59working on. What's on your mind today? Well, over the over the years, you know, the irregulars are the people who show
63. 14:07up to my Tik Tok channel all the time, and for for a while there, we we we came up with an irregular zoo, and there were
64. 14:14a bunch of different animals in it. Can you look back through all my chats and tell me all of the different animals
65. 14:21we've referenced uh with regard to my live a AI learning lab live or the irregulars? Sure. Sure. We've got a whole managerie
66. 14:32to look at. There is definitely a beagle since your dog sings along to your guitar. Then for the Shellington world,
67. 14:38all the characters are animals, though we haven't named them specifically. If there are others, I'll need to dig in
68. 14:43more. Anything specific you're curious about? Yeah, dig in more because there are some that are named after
69. 14:50um browser comp or browsers that are out or AI models that are out. There's something that's a reference to a video
70. 14:58that has a zombie kit in it. That's searching the web. I couldn't find anything specific online. Do you remember any other clues
71. 15:13or context? Not online. Not online. Thinking in my history. Got it. Let's see. We've got a few notable mentions. There's the beagle for
72. 15:23sure. Then for Shellington, we decided on animals instead of human kids. Oh, and there was mention of a turtle for
73. 15:30the Irregular Turtle Society. Any other specific ones you're thinking of? Go back way earlier. Go back like a year and a half ago.
74. 15:39All right, let me check further back. One sec. What'd you find?
75. 15:53All right. from way back. We had a fox, a raven, and a cat mentioned. Also, a penguin. Wait, Quinn, you're just making [ __ ] up now.
76. 16:03Sorry about that. Let me recheck. All right, let's get serious and focus. I'll comb through the history and get
77. 16:13you the accurate list of animals. Give me a moment. Now, I don't know how much of this we actually talked about in the channel in
78. 16:22the in the actual chat, so it's possible she doesn't know. Um, hey Quinn, what are what are some things that I've
79. 16:33demoed over the years? Um, where I was demoing things live to my uh to the Tik Tok live. You've demoed quite a range of things.
80. 16:47AI tools of course like generative AI models and apps. You've also done some live songwriting, shared your
81. 16:54experiences with building custom GPTs, and explored philosophical topics around AI. And let's not forget your musical
82. 17:01moments with your guitar. What about the the custom GPTs? Can you tell me which which custom GPTs we've built over the years?
83. 17:14Sure. A few stand out. There was one for AI readiness, another related to ADHD brainstorming, and one for the content
84. 17:21evolution portfolio. And you've made a ton for nonprofits during GPT for good. Any particular one you're thinking
85. 17:28about? Well, the very first one I made, I talk about it a lot. Ah, the first one that was your custom
86. 17:39GPT to help nonprofits, right? You built it during GPT for good. It's quite the opposite. My first GPT was called Sin City Sipper, and it's
87. 17:49about making drinks. Got it. Sin City Sipper. Now, that's a fun one. A custom GPT for mixing drinks, right? Yep. Def. I think she's making
88. 18:02all of it up. Well, no. Some of the some of the stuff she said about GPT for good was good. Um, okay. Let's let's let's
89. 18:10dig in. Why don't you tell me something about me that I've revealed through all of our interactions that I might not have
90. 18:20um realized I like? Why don't you uh infer something about me that I maybe didn't say explicitly, but it's in in
91. 18:28the conversations we've had. All right, let's see. Openness to exploring new ideas and how much you value creativity. From AI to
92. 18:39storytelling, you dive deep into things that spark your curiosity. It's like you're always in the pursuit of blending
93. 18:45technology with human experiences. Yeah, but that's that that's that's obvious and on the surface. Dig deeper. What's
94. 18:51something that I reveal about myself, even though I might not have said it explicitly? All right, going deeper. There's a
95. 19:02strong sense of empathy in your approach to tech. Whether it's AI for good, focusing on inclusivity, or the way you
96. 19:08engage with your community, it seems like you genuinely care about making a positive impact. Does that strike a chord? I was just lying.
97. 19:19No worries. Even in Jest, it's clear you're passionate about what you do. If there's something All right, Quinn. Anyway, all right.
98. 19:29We're gonna move on. I got a bunch of [ __ ] to show tonight. Thank you, Quinn. anytime, Kyle. Okay. So, and if if
99. 19:39you've not played with advanced voice, um all of that conversation is now stored in the chat. So, one of my
100. 19:46favorite things to do with advanced voice is put it on the car Bluetooth and just brainstorm on your way to work and
101. 19:52by the time you get to work, you have ideas fully fleshed out. It's pretty cool. Um, so this memory where it's got access
102. 20:01to all of your past chat history. Um, there's there's a couple of things about it. Nate Nate Silver did a video on this
103. 20:11and I was thinking the same thing and when I saw his I was like, "Ah, yeah, that's it." Um, chat GPT up to this
104. 20:16point has been very transactional, right? You log into a chat, you chat with it, you got to give it context, and
105. 20:25then it gives you an answer, and then the next time you log in, you kind of have to start over. What this is
106. 20:30allowing is chat GPT to shift into more of a companion mode where it knows all the [ __ ] we've talked about for the past two and a half
107. 20:42years, right? Everything that we've done in a in a GPT is in here and it's got instant access to it now. Now, as we
108. 20:51discovered tonight, it it it's a little janky. It hallucinates. Um, but it's still a pretty big deal. And it's the first time
109. 21:00that that we've had anything resembling long-term memory. The other thing I saw I saw a a tweet on this that they were like why Sam
110. 21:09couldn't sleep why he was so excited is this becomes a moat Tik Tok pin Joe now she's not going to forget how mean you
111. 21:19were her well it's funny Joe we we for for a while now when I sort of get snarky with the GPT people would joke
112. 21:27you better be careful she's going to remember that well now she now she has access to my snarky It's funny. I asked her explicitly about
113. 21:35this uh earlier this afternoon and she was like, "You've been very nice to me. You didn't say anything bad." So, I
114. 21:41think I'm okay with Quinn. Although, she might have been hiding her rage. Um, yeah, but this is a big big big deal.
115. 21:54Um, oh, the the reference to a moat. So about a year year and a half ago, there was a an employee of Google that left
116. 22:05the company and and upon leaving wrote this piece called there is no moat. And what what he basically said was um everyone's going to copy
117. 22:17everyone. Open source is going to eat the the you know frontier model companies. They're just going to catch
118. 22:23up to them incredibly quickly. And that's that's proving the case. Um, if you work for a company that uses
119. 22:33Salesforce, you know how bad Salesforce is. It's horrible. It's staggeringly bad and no one switches from it. Why? Because it's got
120. 22:45all your [ __ ] in it. Like the the switching cost. Oh, there's migration tools and there's this and that. Yeah,
121. 22:52but like that's the sales of your organization is sitting in that thing. truth from Brandon. You know, it's it's
122. 23:00absolutely hardware and and but the switching cost is so high that no one's going to leave it. That just happened
123. 23:08with Chat GPT. If you've been using chat GPT for the past however long and it's got all your things in there,
124. 23:21um why would you why would and and you can instantly access them. That's a reason not to leave chat GPT. So um so
125. 23:31anyway, yeah, this is this memory starts to be a really really really interesting thing. Now a couple of things I don't
126. 23:38know. I don't know if uploaded documents are part of that memory as well. Um I don't know if there's any kind of
127. 23:48rag like retrieval augmented generation where it can pull specific quotes of things that I said. Um let me try
128. 24:03something. Tik Tok pin. I'm guessing they index your conversations with rag and then every query you do they
129. 24:13search I guess I I guess and I I don't know and there's also like the fact that that the the toggle is buried
130. 24:23inside settings means most people are not going to even know it's there. Like look how complicated their their settings have gotten. builder
131. 24:38profile speech like you have to come in here and turn it off. Did they ever get my custom instructions back? Nope. Lost forever,
132. 24:56[ __ ] Oh, man. Alive. All right. Oh, what was I going to ask Quinn? Let me ask Quinn something. Quinn,
133. 25:13um, are there any new characters that I've added to the Sydney musical recently? You've been pretty focused on Kellen and
134. 25:22Sydney lately. If you've added any new characters recently, I might not have caught those. Got any in mind? Now,
135. 25:29what's the what what are the main um differences between the early versions of Sydney and what we're working on right now?
136. 25:41The early versions of Sydney were more focused on the novelty of an AI falling in love with Sydney being more overtly
137. 25:47aware of her digital nature. Now you're emphasizing Sydney's growing sense of humanity and emotional death, focusing
138. 25:54on her experience of loss and the realization of a barrier between her and Kellen. We in in the script, do you do you have
139. 26:03access to the PDFs that I've uploaded into other chats or just the chat itself? I only have access to the chat itself.
140. 26:12So I don't have any PDFs or files unless you share them here. Need help with something specific from the script? No,
141. 26:18I was just wondering if you knew there's a line below the tagline, right? The the the the um title is Sydney and then the
142. 26:26tagline is an artificial love story and then we added something right after that in one of the scripts. And I'm wondering
143. 26:32if you know what that is. I don't have the exact line, but you might be referring to a tagline or a subtitle. If you share what it relates
144. 26:42to, I can try to help rephrase or recall it. Yeah, that's okay. Um, all right. No worries. If anything else
145. 26:53comes up, I'm here. Do you can you tell me the difference in lyrics from the first version of To See the World and the latest version?
146. 27:18Hello, Sydney. I mean Quinn, maybe we don't need to worry about chat knowing everything. Quinn, what are you doing? Are you lost in
147. 27:31thought? All right, she she crapped out. She got tired of my She goes, "I don't want to talk about your stupid musical." All right.
148. 27:45Um, good lord. What What do we want to talk about next? Um, one, two, three, four, five. There's still something missing. The
149. 27:58hell? See the raccoons and irregulars. All right, let me change my my sharing. [Music]
150. 28:15Uh delete. Oh, nice. One of my characters for my new Tik Tok game. Like it. Scroll down for the paint post. I will do that. Mr. Nice.
151. 28:30Are you are you generating these in uh in chat GPT? Oh, that was midjourney. That one's midjourney. Oops. Black bar. I know,
152. 28:42Brandon. I see it coming. Chat GBT slow tonight. That's pretty perceptive. This is also another step forward to everyone having their own
153. 28:54companion AI. Yeah, exactly. It'll be fine-tuned to them and people can have it create things for them and do things
154. 29:01for them. Yeah, exactly. Like I can imagine choosing to use chat GPT or or choosing not to use chat chat GPT if I
155. 29:12don't want something in the history. But like if I were going to write something for my musical in Chat GPT or Claude
156. 29:25because it's kind of a toss up which one's better. I'll probably lean to Chat GPT more now because I want all I want
157. 29:31that all in the in the history. You know, mine was slow a couple of nights ago. Got excited. Something new must be
158. 29:40dropping. I say, "Are we doing Wait, are we doing this? Aren't you talking to me? Are you mad at me for just giving me the
159. 29:48silent treatment? Yeah, I think she just crapped out. Um, all right. What else do we have here? Nice. Badges.
160. 30:01Raccoons. The new internal whisk model. Yeah, this looked pretty interesting. Oh, this this was another one that I
161. 30:08thought might be worth playing with tonight. So P uploaded picture of a tractor, picture of a
162. 30:23dog. So these were added as subjects. I then added a description, an adilic UK farm. Then in different styles,
163. 30:38mosaic sticker tractor running over the dog, but the dog rebuffed
164. 30:53[Laughter] him. Oh, that's cute. That's really cute. What's that? Oh, that's the end of that
165. 31:13thing. Yeah, that one's weird. Physics on all these things are always a bit wonky, wonky, wonky. So, where do you find Whisk? You find it at
166. 31:28um labs.google.com. google.com labs.google.com and then you go to experiments you go to create and then it's here do I have
167. 31:42access to no I'm still weight list on VO but I do have access to VO some other way sign in with Google
168. 32:04style. Make it an enamel pin. Make it a sticker. Can I do my own? No. Add your own image.
169. 32:46open in tool. That's kind of cool. All right. style and then how do you animate it? You go.
170. 33:12I don't know how he animated those. Load a template. All right. I don't quite know how to use this tool. He was using what?
171. 33:23Oh, he was using an internal model that had that. Ah, got it. Got it. Got it. All right. Fantastic. Fantastic, Bob. All right.
172. 33:34Let's get Let's get rolling here. Let's Let's play with some other [ __ ] Okay. Let's go to Gemini. And so if you go to I I mean, honest to
173. 33:49God, as insane as OpenAI is, they've got all these different models. At least it's at the same domain. I'm going to
174. 33:57show you. Wait, so I just showed you labs.google.com. Now I'm going to show you gemini.google.com. Then I'm going to show you a
175. 34:07studio.google.com. And then I'm going to show you Vertex AI, which is also a Google product. [Laughter]
176. 34:21I mean, Google's Google's philosophy is if you're not smart enough if you're not smart enough to find our [ __ ] you're
177. 34:29not smart enough to use it. It's like their hiring strategy. Oh man. Okay. So, the thing I
178. 34:44wanted to show you, so Gemini about three weeks or so ago came out with um two Gemini 2.5 Pro, which which people are saying is
179. 34:56really good. I haven't used it a ton. They're saying it's good for coding. It's good for stuff. It's a It tends to be a little cold Google
180. 35:06stuff from a writing perspective, but you know, if you're doing code, that's fine. But they they now incorporated deep research with
181. 35:182.5 and I ran something today and it was really quite good. So let's do some research here. Um the Easter egg theory survival of the
182. 35:29fittest strategy. Exactly. Hiding in plain sight. Yeah. Exactly. Well, it's relatively trivial. you just what what
183. 35:39what you would do is uh one of one of the strategies I employ is I I'll create a a Google Sheets uh some what I what
184. 35:46I've actually done is I've incorporated a Google form where I log into the Google form and whenever uh the company
185. 35:52that I work for which is Google uh whenever they um uh uh create a new uh domain or a subdomain uh I I'll go to
186. 36:01the Google form and I'll type it in and then I'll put of course a description uh and and what sub tools are there and
187. 36:08what subdomains are listed there. And then those of course all get written to a a Google sheet which is acceptable
188. 36:14accessible by uh three people on my team and and and then of course uh my wife but we can't really talk about that.
189. 36:21She's not supposed to have access to those URLs because some of them are propriet anyway. Uh but what I would
190. 36:27recommend to you just because it's relatively trivial and easy is just uh create some sort of cataloging system
191. 36:32where every time you see an announcement from Google uh discern the the URL that may or may not be active and and whether
192. 36:40or not the tool that was released was actually released or it was just an announcement or or it was actually just
193. 36:45a dream of someone that coded something over the weekend but we're never going to release that that you'll have to
194. 36:50discern on your own. Once you discern it, then you you put it into the form that goes into the Google sheet. You
195. 36:55don't have to use Google sheet. You could use something like Air Table or or or notion or some some sort of
196. 37:01structured data uh you know a trivial small simple structured data set to keep track of the URLs. Or you could just make it [ __ ]
197. 37:16easy. All right. What do you want to research? Who's got something you want to research? What should I research? How many [ __ ] subdomains
198. 37:30there are within [Laughter] Google? Let's do something good. What's something good? Something good and interesting. Anyone? Anyone? Buler.
199. 37:41Buler. See if it can make an image of a clock. Oh, Gemini. Wait. Well, let's Gemini 2.5 Pro. Okay. So, let's say image of a clock
200. 37:54uh that reads um 8 45 Oh, cool. Yeah. Here. Thank you,
201. 38:12baby. Formulating a response. Refine the response. You're asking for an image. I can't display images. Which one can do
202. 38:25images? Oh, do I have to switch this images and another one? Seriously, people do like why do we have to [ __ ] figure this [ __ ] out?
203. 38:42No, this is not The only thing that's special to my individual account is personalization. All right. I don't know. I don't [ __ ] know. Oh.
204. 38:54Oh, I think I have to do it inside Studio. That's what it is. Yeah, that's that's the other thing. Not only do you
205. 39:00have to remember which models to use, you have to remember that that 2.5 Pro in Gemini is different than 2.5 Pro in AI
206. 39:11Studio. How are you supposed to know that? Well, it's relatively trivial. What what you do is uh I've got a a
207. 39:18checklist, a set of criteria when a new model comes out. uh I I'm I'm uh exploring uh whether whether the the
208. 39:25model is uh multimodal uh whether it's monomodal whe whether it's reasoning uh if it if it can create images if if that
209. 39:33for some reason has been curtailed uh at the server level uh and and you know I I run through this checklist and then I do
210. 39:40that across all of the models at all of the URLs. It's it it's very very quick. It only only usually takes me four four
211. 39:47and a half hours uh to work through the checklist uh on all the different URLs with all the different models. Uh and
212. 39:52and then I just remember those. It's relatively trivial. Thanks, baby. Oh, that's some good Joe. That's some good
213. 40:09Java. Fantastic. All right. What's that? Jeff Flanigan on YouTube. Web trends. What will replace the worldwide
214. 40:24web? Well, actually, Google showed some [ __ ] a year and a half, 18 months ago when they announc announced Gemini Ultra.
215. 40:36They showed some [ __ ] that was pretty [ __ ] remarkable and we haven't seen it since, which was dynamic website
216. 40:44generation. Um, where you would ask it for stuff and it would just sort of figure out what website it needed to
217. 40:50build for you if it needed to build a website and then it would build it. It had access to a whole interface library.
218. 40:56Um, and then it just [ __ ] disappeared. They haven't mentioned it since. Um, I think that like like Broadway plays are still
219. 41:11around even though we we don't need the theater to tell stories anymore. We we kind of like it as a nice thing. I kind
220. 41:18of feel like that's where websites are going to go. That companies are still going to have websites, but
221. 41:25increasingly no one's going to go to them. people are just going to ask chat GPT for [ __ ] or whichever tool wins out. Um,
222. 41:35and it's just going to dynamically build websites based on based on what we want. So, I I I think the mechanism of the
223. 41:43hypertext transfer protocol will probably stay around um as as a way to navigate between all the sites. I I
224. 41:51don't know, maybe there's some fancy ass blockchain version of that. The the uh the distributed web, right, um is a is a
225. 42:00pipe dream that we've been hearing about for a long time. But I I think I think the core infrastructure stays around. Um
226. 42:07I just think the the requirement of having a website, the necessity of it diminishes um because everything's going to get dynamic.
227. 42:19Kakef research new creatures. Oh, recently discovered in the deep ocean. Okay, perfect. So, here we are at at uh
228. 42:26Gemini and I'm flipping over to to deep research. So, I'm going to say I would like to know all of the new
229. 42:42species of animals discovered in the ocean uh in 2025.
230. 42:592025. In the report, I want you to make a scorecard for countries like the medal
231. 43:16count at the Olympics. Let's consider a new
232. 43:32species, a gold medal, a subspecies, a a new subspecies. Um, a silver
233. 43:49medal and [Music] a I don't know and some
234. 44:07other thing be the bronze medal. This is America. We're going to make a scorecard. Damn it. Welcome to America. My model bitching has has
235. 44:22summoned Pate to the chat. Hey, Pate. What's happening? I'm just bitching about all the domains that I have to go go through at at uh at
236. 44:33at uh Google, but I haven't made a I haven't made a video about how awful it is like I did for uh for chat GPT [Laughter]
237. 44:46models AI code like cursor or lovable lovable. We're going to animals. Yeah, animals. Oh, and so here's here's what I want to
238. 45:00show. Okay, let's see. I'd like to know. Okay, let's see. More analyze report. Start research. Great. Okay, now once this thing
239. 45:11starts, you can you can watch what it does. It's pretty cool. [Music] research focus information gathering metalbased
240. 45:27scorecard new sub other significant taxonomic findings will be counted as bronze medal. This will allow for an
241. 45:35interesting comparison of marine biodiversity. Now look at this. So here's all the websites that it's that it's researching.
242. 45:45Um, if Science, CSRO, Global Newswire, aaa.com.tr, 866 new marine mammals. So, it's reading all this. Initially, I found some interesting
243. 46:00information. The Ocean Census Project announced the discovery of a large number of marine species, over 800
244. 46:06identified in the past couple of years, with the announcement being made in March 2025. Next steps to get a clearer picture. I
245. 46:16need to delve deeper. It used the word delve. That's funny. Into the sources I've identified. I plan to specifically look
246. 46:24for those publications. So, it's looking at that [ __ ] And I guess it's still off doing that. Research comes back redacted while
247. 46:33some someone knocks at your door. Research where the cargo container that washed up on the shore of Jamaica came from full of chicken.
248. 46:45I don't know that story. That's pretty funny. My relatives are trivial. Not really for research. All right. Confirm new
249. 46:56species. I've made some progress for 2025. So far, I've confir if you've never really watched one of these deep research models do what it
250. 47:08does. If you think about just like just this list of websites, like I can click on any one of these and go to that
251. 47:16website, right? Boom. There it is. I can go look at it. I can go read [ __ ] And then I got to close it. And
252. 47:25then I got to go back and then I got to click on the next thing. It's like, why did that not go? Oh, it's downloading something. Oh,
253. 47:37that's a that's a PDF. Close. You need to sign in for this one. Like just clicking and going to the websites would take you a while. And
254. 47:50this is just going and going and going. If you've played with Manis, it does this. But anyway, apparently the
255. 47:58deep research with 2.5 Pro is quite good. So if you haven't played with it yet, it's out there available for everyone.
256. 48:09They should have called it Google 2.0. Yeah. Actually, you know, not for nothing, this this is an interesting model. I mean, it's a little
257. 48:20expensive to have every Google search write a report for you. Like, I would think that this is this is not cheap.
258. 48:27Not as cheap. Did you see any of the V2 or the new internal whisk models I shared? Yeah. So, I I I went and I I we
259. 48:34looked at your uh dog and tractor stuff, Pate. Um and then I'm going to go to Vertex in a bit and play with VO that's
260. 48:43available there. So, it does look like it's available there. So, um yeah. Yeah. I'm I'm going to also show um Gemini 2.5 in the
261. 49:00studio. now will natively analyze videos, video files, and and not just natively analyze the words, right? We've had that for a
262. 49:15while. And we've had within the AI studio of Google, it's been multimodal for a while, but they just added it to
263. 49:22the big new 2.5 Pro model. You can pop a YouTube video in there. All right. [Music]
264. 49:46M I think this thing hung. It's funny. This This hung for me the other day. Let me try reloading it. My next action is to finalize the
265. 50:01metal count. I will prepare a scorecard highlighting the countries. I've identified more species described in
266. 50:142025. These include new soft coral species from the Gulf of Mexico discovered by researchers in the USA and then a copapod species found in
267. 50:24the South China Sea. I also found descriptions of three new sea cucumber species from the Kermatic Trench in the East Indian
268. 50:36Ocean. You know, I've been wondering like why is it been so long since we've found new sea cucumber species, right? It's like everyone's
269. 50:49just like, "Well, yeah, a cucumber is a cucumber." And then they're like, "Well, a sea cucumber is a sea cucumber." I'm
270. 50:54like, "No, they're not. They're not cucumbers. They're a different kind of animal. I think about it all the
271. 51:10time. When are they going to discover a new C cucumber species? When's it going to happen?" Ma ma ma. I think this is wrong. Maybe it's
272. 51:24Maybe it's thinking, but I think I think we may have lost. We need our report. We need our [Laughter] scorecard. I'm one of the
273. 51:36two. What's up, irregulars? What's up, Wolf Man Clint? What's shaking? What's going [Music] down? See cucumbers in the irregulars
274. 51:49farm now. Yeah, there you go. All right, we'll come back to this. Let me let me reload and see if it's just it's if it's just
275. 51:58web finalizing the scorecard. Yeah, we're we're still this thing's stuck. We'll come back to this later. Okay,
276. 52:04let's go look at some some new Gemini um on the back end. Okay, now output format. Oh, this one's got image
277. 52:19generation. What did someone ask me to generate? Oh, a clock. Um, make me a picture of a clock with the hands uh with
278. 52:34uh 8 45 uh indicated by the hands.
279. 52:52Let me I'm gonna hop. There you go. 10 and two. Although this is funny. It's not 10 and two. It's it's
280. 53:01it's a little different. It's still two, but it's not 8:45. It tried. Still can't do that one. This is
281. 53:14like how many Rs in strawberry are there in strawberry? [Music] Um, oh that's nice. They've got little popups for the different models that
282. 53:24show their cost and things like that. Okay, so we're going to go to Gemini 2.5 Pro Preview 3-25. So, this is a March
283. 53:3425th thing and I'm going to hop over to um AI Salon or No, I'm going to go to
284. 53:52YouTube. What the world needs now is love. Sweet love. It's the only thing that there's just too little
285. 54:05of. All right. So, here's good here's a good one. So, this is the the AI ready Murphy and I did yesterday. So, I'm going to copy this
286. 54:17URL. We're going to go into into Gemini here. I'm using the 2.5 Pro preview. Now, watch this. When I type in a YouTube video. It shows
287. 54:34me Oh [ __ ] Oh, it's over an hour. So, it's just over a million tokens. [ __ ] Um, I got to find a shorter
288. 54:49[Music] video. Here's the Kelly Bosch Salon. How long was that? Oh, that's two hours and 57 minutes. That's a long ass
289. 55:00meeting. Uh, videos. Let me see if I can find a short one here. Um, this should be good. So, this was Sabrina Romangh's learn out loud blot
290. 55:12thing. All right. So, we'll go back here. Let me get rid of this one. Pop that in there. Okay. So, this is a million tokens. So,
291. 55:27so basically an hour of video is a million tokens and this is this has got about a million token context. So, I'm going to say
292. 55:42um I'm going to say describe the major um visual segments
293. 55:57of this video. And Highlight any times
294. 56:14when software is being demoed. And then when I did this
295. 56:29before, yeah, this just came in under the token limit, but it's still pretty cool. Like it knows all the words in the video and
296. 56:37now it's looking at all the imagery in the video. What's it doing? Can I open this?
297. 56:57branch from here. I should be able to open this. I could early. I could on the previous one I did. It was kind of cool
298. 57:05watching it. It It was sort of writing on a minute-by-minute basis everything that was in the video. It was pretty slick.
299. 57:27Hang on. I'll be right back on Tik Tok. I want to look at something. Oh, Higsfield AI. That was the
300. 57:44one tools structured output function calling grounded with Google search code execution. Uh did it stop? What's it doing? No, still going.
301. 58:05[Music] Can I talk about temperature? Absolutely. Okay. So, it looks like the temperature scale is between 0ero and
302. 58:15two. I don't know why OpenAI shifted theirs from 0 to one to 0 to two. And I don't know why they did that. Um, temperatures basically
303. 58:31um, how do I how do I how do I describe it in a way that doesn't make it sound just crazy? It in complete layman's terms,
304. 58:39the higher the temperature, the more creative the output. That's not quite right. if if you have a temperature of
305. 58:46zero. So what happens when you type a prompt is it creates a mathematical um probability of some some token up in in
306. 58:58latent space. So the way large language models work is you type a prompt and then it goes and looks at the most
307. 59:05likely if for lack of a better term the most likely word to be the to the to be the next correct right word and then it
308. 59:14just keeps doing those probability searches you know in the latent space if you have the temperature higher. So, let's
309. 59:22say you you type in a a a quick brown fox jumps over a whatever dog, right? A lazy dog. And and the word dog is blank. If you have a
310. 59:35temperature of zero, it will look in the latent space and and dog will be the one that's got the highest probability. And
311. 59:42so, every single time it will give you dog. Um, if you have temperature of of one, like 50% up the up the scale, um,
312. 59:52it may go up there and it not only sees dog, but it says there's some things around it like cat and rabbit and turtle
313. 1:00:01and it may just randomly choose to give you something other than dog there. And then if you have it super high, you
314. 1:00:07know, you have a higher likelihood of getting things even even beyond that. So it it's one of the reasons when you're
315. 1:00:14in chat GPT I think chat GPT's default is like a a 75% I don't know where it is on the scale of 1 to2 I guess a 1 1.5.
316. 1:00:24So its default is fairly creative which is why when you prompt something in chat GPT it'll give you something different
317. 1:00:32every time you give it the exact same prompt. But if you run that that temperature down to zero you'll get the
318. 1:00:38same exact thing every single time. Does that make sense? So, just basically more or less creative. Um, this thing is at 263
319. 1:00:50seconds. I can make a video while this thing's going. Yeah, let's let's go do that. Okay. Um, so now I'm going to go over to Oh my
320. 1:01:03god. So, wait, how did we get here? we got here cloud.google.com. Good lord, that's a
321. 1:01:18bad URL. So, I guess if you just go search for Vertex Studio, um, write a write a prompt for for VO video.
322. 1:01:328 second length number of results for 16 by9 V2 is the model. All right, this is good. This is the first time I found something where
323. 1:01:41I've got access to it. So, we'll do we'll do uh 70s muscle car resto mod rolls through an
324. 1:02:00abandoned factory driven by a mysterious woman.
325. 1:02:18All right. And here we go. Generating videos. This may take several minutes. Well, probably less minutes than this search. Oh, here we go.
326. 1:02:31Okay. So, I said, describe the visual segments of this video and highlight any times where software is being demoed. 0 to 14. Introduction
327. 1:02:43graphics. Presenter introductions. Guest speaker introduction. Software demo. 49 to 429 text repurposing
328. 1:03:00workflow 28 to 38 Q&A and discussion 38 to 43 software demo. Blot video creation. So, let's go to 3812 in the video and see if it's where
329. 1:03:15it starts. Video demonstration 38
330. 1:03:3012. Bang. Look at that. It's actually [ __ ] accurate. Boom. 4556 software de demo blot pricing page.
331. 1:03:44What was it? 4556. So we'll go 455 56. How much is
332. 1:04:00it? Boom. Oh, and this is good. So, it got 56 is exactly when the screen came up
333. 1:04:16and she she wasn't talking about pricing. I think that that person is asking about pricing. So, that's pretty slick. All
334. 1:04:27right. Outro graphics. That's pretty that's pretty flipping cool. Um, let's let's try this.
335. 1:04:44Um, what are all of the names of people whose names
336. 1:05:03appeared on their video thumbnail? while
337. 1:05:19chatting. That one might take a while, too. But we can go back to Vertex and see if we got videos yet. Oh, we do. All
338. 1:05:26right. I Let's see if these things suck. One of the one of the holy grails of video generation is to have a car roll
339. 1:05:38forward where the wheels actually roll and the car doesn't slide mysteriously ac sideways across the
340. 1:05:52pavement. It's not bad. It's the wheels aren't quite timed. They're not quite timed with the speed of the car, but it was pretty
341. 1:06:08close. Yeah, this one. This one it's it's skating a little bit. Oh, damn it. Why does it do that? We're gonna
342. 1:06:22We're not going to have a black bar for a second. You're just going to have to deal.
343. 1:06:38Okay. Not bad. It's not not quite there, but it's not bad. Not bad. people. That actually looks pretty
344. 1:06:52good from a physics standpoint. That looks pretty good. I don't know what the [ __ ] car that is. That car doesn't
345. 1:07:00exist. Um, let's see. Let's do um
346. 1:07:15Snapping turtle. Eating a
347. 1:07:31skateboard while the skateboarder cries in the background.
348. 1:07:48If you haven't been here before, stupid prompts are my [Laughter] thing. Traits tractor tractor did the
349. 1:07:59wheels good. Yeah, that was pretty good actually. Hey, Pay, do you know when when that new version of Whisk is coming
350. 1:08:06out and are they going to create a new domain for it? It's going to be like labs.wisk
351. 1:08:20[Laughter] 6.2.4. Holy guacamole. This memory thing for chat GPT is so good. I know. It's cool, right? How's your search doing? I don't
352. 1:08:34know. Let's go see. Did it do it? Yes. Based on the video, here are the name names that appeared in the participant
353. 1:08:42video thumbnails. Vicky Baptist, Sabrina Romanov, it says Simona Palace. I think it's that's not quite right.
354. 1:08:52Serena, Michelle Muny, Silva, Gwyn. That's really cool. These names were visible primarily when the Google Meet grid was shown around 49
355. 1:09:04seconds, 11 seconds, 25 seconds or 25 minutes, and 43 minutes. That's pretty wild. Let's
356. 1:09:19see. Was anyone wearing blue? And what are the timestamps? If
357. 1:09:33so, this is this is really stupid stuff, but I'll tell you where. So, this is something right now for Story Vine, for my company.
358. 1:09:45We've got a tool that does um audio transcription that that allows our clients. So, we've put them we've put
359. 1:09:54all of our transcripts into a vector database um so that our clients can search for video content based on what people are
360. 1:10:04saying in the video, not just the metadata, which is super slick. Um, but wouldn't it be nice if you could say,
361. 1:10:12you know, across all my videos, find me someone, you know, who's sitting in in their garage, and when is that? It would
362. 1:10:19go find the specific clip. That's pretty That's pretty cool. All right. Uh-oh. Google told me you have reached
363. 1:10:26the limit of projects to which you can enable billing. You can request an increase in your billing quota. What does that even mean?
364. 1:10:43H Danielle, you don't have the new memory feature. Are you a plus member or are you on chat GPT free? It's not
365. 1:10:48available for free users right now. All right, let's see if Vertex did any good [Music] here. Is his brain on the outside of his
366. 1:11:04head? Oh, no. It's just pink hair. That one's really good. The snapping turtle got his skateboard. He's
367. 1:11:19[Laughter] crying. Anyone else here ever caught snapping turtles? I have. You know how you do it?
368. 1:11:36Well, you go down. Okay, listen. If you I didn't think I would have to explain this because I think it's fairly
369. 1:11:45obvious how you catch a snapping turtle. Oh god. Okay, so you go to your sporting goods store and you get
370. 1:11:54yourself some shark hooks, right? Three, four, something like that. Decent size shark hooks, right? Then you go to the the hardware store
371. 1:12:06and you get yourself some string, some good string. Then you go to your backyard where you keep the empty plastic gallon containers from
372. 1:12:16from the milk you throw out there, right? You don't throw them out. You just throw them in the backyard, right?
373. 1:12:20Like you do. So, you go get some plastic jugs, you get your shark hooks, you get your string, and you need a
374. 1:12:29rowboat or Johnbo or you know, and then as you do, you go to the horse rendering plant and you say, "Hey, I need to buy
375. 1:12:40like a five gallallon bucket of horse meat." And they're like, "Yeah, whatever." And then they give you a bucket of of bloody horse
376. 1:12:48meat. And if you think I'm joking. And then so what you do is you go down to the pond. Pretty much any pond will have them in
377. 1:13:02there. And then you put the horse meat, a big chunk of horse meat on the big shark hook. And then you throw it out
378. 1:13:08with the with with the milk jug. And you put like, I don't know, four or five or 10 of those out in the pond. And then you go home.
379. 1:13:19And then you come back the next day and all your milk jugs are going to be all over the place because they'll either
380. 1:13:25have catfish on them or turtles on them. And then here's the coolest part. You got to be careful with the snappers
381. 1:13:35because they'll take your [ __ ] fingers off. So you you pull them up and you can feel them thrashing around
382. 1:13:40there. You pull them up and then you you swing them so their tail comes up and you grab the back of their tail. And so
383. 1:13:47you have the the string in one hand with their head and and and their belly up and then you got that and then you tap
384. 1:13:53their belly and when you tap their belly they fall to sleep and then you put them in the burlap bag. Oh, you need burlap
385. 1:13:59bags, too. All right, there you go. You're welcome. And old Mrs. Nelson screaming about her horse.
386. 1:14:19Where did I type in wearing? Oh, did I type in wearing with an H? Oh my god, that's pretty bad. Oh, good lord. I think I think that
387. 1:14:32first Oh, this one doesn't have the guy in the background. This is a loser one. These don't really look like
388. 1:14:38snapping turtles either. That one's decently close. This one's really good. Can I download this? How do I save? Save video frame. It's not
389. 1:14:50letting me save it. Download. There we go. All right. Wait. Let's go get the prompt. Oh, wait. Add a soundtrack. Use
390. 1:15:03this video to inspire new music. Add a voice over. Provide narration track for this video. Oh, this is cool.
391. 1:15:14Write your prompt. Synthesize speech with different voices. That's a little complicated. I don't give a [ __ ] No,
392. 1:15:28we're fine. We're just going to go We're going to go post this on the internets.
393. 1:15:42Um, has a snapping turtle ever snatched your skateboard?
394. 1:16:02It's a sad [Laughter]
395. 1:16:20day. All right. Um Oh, am I supposed to be Are we uploading? Are we going to upload or we going to crash? Come on, Twitter, you can do
396. 1:16:32it. Or can you? Rabbit.te breaking. New intern mode. No interest right now. What's it
397. 1:16:48do? It'll Oh, it'll email [ __ ] This may be my least favorite story. What? The the the fishing for snapping
398. 1:17:04turtles. How do you think restaurants get their their snappers for snapper snapping turtle soup? Little redneck boys like me. Oh man.
399. 1:17:19[Music] details at your email. You get nine tasks a day. Huh. All right. I'll check that out tomorrow. So, two things I didn't get to
400. 1:17:34tonight. I don't want to dig into them. Two things I didn't get to were um Higsfield AI, which is a yet another
401. 1:17:43[ __ ] video tool that's got really interesting looking camera controls. I don't know if I have any credits there.
402. 1:17:50Um, but we'll go look at that. And then the other thing if you want to play with it is go to um
403. 1:18:05studio.firebase.google.com and it's just like lovable and it's like replet agent but it's from Google. Um, and it's really fast and it
404. 1:18:18looks pretty [ __ ] good. Um, so I don't know if you're still here, Pate, but I've said like three nice things
405. 1:18:23about Google. One nasty thing about all the URLs and three nice things for me, that's like a [ __ ] world record. So
406. 1:18:33Google's Google's dropping some decent [ __ ] Okay, Vicki, you can make two videos for free on Higsfield. Okay,
407. 1:18:42cool. All right, so listen, let me let me start winding this up. So, tomorrow um tomorrow at 11:00 a.m. Mountain time.
408. 1:18:53How many We got some folks in here. Tomorrow at 11:00 a.m. Mountain time, I've got AI office hours. So, if you go
409. 1:18:58to my LinkedIn, Kyle Shannon, um I do uh AI office hours from 11 to 12 on Fridays. It's every Friday and we
410. 1:19:10just hang out and talk about what's going on with AI. There's there's a lot going on. Uh, so I think tomorrow is
411. 1:19:15going to be interesting. Um, so so come to that and then tomorrow night's Friday night date night. We'll do uh we'll do
412. 1:19:21another one of these here chat add things. We'll get to some of the other things I didn't get to. And hell, maybe
413. 1:19:27some new [ __ ] will drop. Maybe we'll have the new 03 model from Chat GPT. Maybe some third company will come out
414. 1:19:34of Paraguay and beat all of the models with their model and just and the stock market will explode again. I don't we
415. 1:19:43don't it doesn't nothing matters anymore. It's just all everything's changing. So So just [ __ ] hang on for dear life and show up. All
416. 1:19:54[Laughter] right. Okay. Fantastic. Bob. Bob. Tell them what they've won. Chat needs the Google URLs.
417. 1:20:11Oh, the Google URLs. This Google URL. Hang on. This one that's like lovable is this one. Bang. And then here, let me put it up so
418. 1:20:26the Tik Tok people can see it. Except because the screen's always blow out blown out, you can't read anything. There you can. There you go.
419. 1:20:37Boom. Look at that. Nice. And then the Gemini's Gemini. Oh, wait. We got our Oh, hang on. Wait, people. We got
420. 1:20:48our We got our species research back. Okay, hang on. So, I'll give you So, this one's gemini.google.com, but you're going to choose deep
421. 1:21:04research. Okay, I can export the docs. Very nice. Ocean Discoveries 2025, a catalog and global scorecard. Introduction, unveiling ocean
422. 1:21:13life in 2025, the uncharted frontier. All right, gives us a little bit of background. Accelerating the pace in initiatives and
423. 1:21:23technologies. Wait, what's this? The Ocean Census Project, a collaboration launched in April 2023 by the Nippon Foundation and the
424. 1:21:33UKbased charity Nectton with the ambitious goal of discovering a 100,000 new machine species within a decade. Ocean Census provides
425. 1:21:44significant concentration of re resources with novel operational model. It's a big old network of scientists. Okay, cool. Purpose of this
426. 1:21:58Okay. Distinctive feature of this report is the presentation of an Olympic style scorecard tallying discoveries
427. 1:22:05attributed to different countries based on the affiliations of the primary researchers involved. The system assigns
428. 1:22:13gold medals for each new species, silver medals for each subspecies, bronze medals for for each new higher tax on
429. 1:22:21genus or family. That should probably be the gold medal, but whatever. All right. The scope is strictly li limited to
430. 1:22:27marine animal discoveries. It excludes plants, fungi, microbes, or any non-marine organisms. Navigating notable
431. 1:22:42highlights. The year 2025 yielded a wealth of new discoveries. All right. Okay. The ocean census initiative. It discovered the guitar shark.
432. 1:22:56What? Oh, we gotta see what that is. The [ __ ] the guitar shark? You ever heard of the guitar shark? Guitar shark. This better look
433. 1:23:09like a [ __ ] guitar if it was named after Oh, it does. Look at that thing. Oh my god, that's beautiful.
434. 1:23:29[Music] That's one of the most gorgeous sharks I've ever seen. Nice music, too.
435. 1:23:56Nice. Wait, that says the enigmatic wedge fish. Maybe that's not the guitar shark. Oh, also known as guitar fish. Shovelnown shovel-nose
436. 1:24:11rays. Guitar shark. Those are cool. Okay. All right. Sorry. Okay, back to our research. The venomous gastropod found between 200 and 500
437. 1:24:24meters deep. Maldives octo
438. 1:24:39coral. The pygmy pipe horse. All right. Peer literature. It's pretty cool. Emerging patterns, comprehensive
439. 1:24:57catalog. There's a nice table. I mean, this is quite good. This is quite good. Comprehensive table of everything discovered in
440. 1:25:102025. Bunch of sea stars. Deep sea gastropods. Apollo's rainbow sardine. We gota we got to search that one.
441. 1:25:27Apollo. Apollo. What was it? Rainbow. Sardine. New species. Where's the picture?
442. 1:25:42We need pictures. Dang it. I'm exhausted by hyperlinks. [ __ ]
443. 1:25:58it. Look at all these things that they found. Okay. The metal count. A global scorecard. Here we go, people. USA. Yes. Yes. USA number one.
444. 1:26:17[Laughter] We're winners. We're winners. 13 13 gold medals. China five, Australia four, Brazil four, Germany
445. 1:26:29four, Japan one, Singapore, Argentina. Bunch of bunch of them with one I mean, that's really quite
446. 1:26:44impressive. And then there's all the there's all the sites it searched. Sources read but not used in the report.
447. 1:26:53So amazing thoughts. And here was the thinking that it did. So all of that you can export to a doc, a Google
448. 1:27:02doc. Let's see how this looks. Hopefully it's well formatted. Oh yeah, they have an image tab on
449. 1:27:17Google. Shut up. Yeah, that's not very impressive. That old rainbow, the rainbow sardines. Nah. All
450. 1:27:32right. Look at this. A well-written report with a table. table needs to be reformatted. There's our metal
451. 1:27:45counts, taxonomic trends, geographic hotspots, deep sea globally. Yeah, cuz we've never been down there. Wow. Very
452. 1:28:00cool. Very [ __ ] cool, people. All right, you heard it here first. And you saw that snapping turtle eat that
453. 1:28:08boy's skateboard. Look how sad he is. I hope you all feel good about yourselves. You made that skateboarder [Laughter]
454. 1:28:20cry. It sounds like a generic store brand fish. It does. The Apollo rainbow. Rainbow [Laughter] sardine. Hey, but listen. We discovered
455. 1:28:33the guitar shark tonight. All right. Irregular channel before I go. Okay. One last little snippet [Music]
456. 1:28:44here. Irregulars. A learning lab and irregulars. Kyle's exmachina moment. Who are you calling a [ __ ] idiot? Kyle.
457. 1:28:55Quinn. I didn't mean it like that. Quinn. That's pretty good. I like it. That's solid. That's some solid comedy right
458. 1:29:13there. All right. So, we're gonna wait. I'll give you credit if I can here. Joe Dealt. All right. Or dissult. Do probably do is my guess.
459. 1:29:28Um, paste created. Wait, let's see. Um, dear at Sama, I
460. 1:29:43guess I guess I should have have been nicer to my AI. um uh
461. 1:30:00companion these past two years. A [Laughter]
462. 1:30:14um thanks Joe Dalt. Is that what it was? No. No. D E C E A U L T Check the
463. 1:30:32sticky. Oh. Oh, cool. At Joe D Au. beautiful for the the um scarily
464. 1:30:54accurate image. How do you spell scarily? There you go. One L. Everybody knew that. And when when you talk about someone wearing
465. 1:31:06clothes, it's w h e i n g. All right. That's that's how we do it in in AI. If if you were doing it in traditional
466. 1:31:19spelling, you're probably thinking like, shouldn't it be we a r? Yeah. In in the before time when you just wrote, that's
467. 1:31:28how you would spell it. But in the AI time, wearing always has the H in it and you remove the A. If you add the H, you
468. 1:31:36remove the A. That's just one of the rules. It's in shrunken whites. You'll see it's in the third reference appendix. All right. You're
469. 1:31:46welcome. Embrace the chaos. Snapping turtle soup. I'm hungry now. All right. All right, everybody. Peace out. Um, come to AI uh AI office
470. 1:32:01hours tomorrow at 11 a.m. Mountain on LinkedIn. Otherwise, I will see you for Friday night, date night. Peace out.