8/22/2025 - The Impact of AI on Creative Industries and the Future of Work

Transcript

1. 0:00Are you ready? Yes, I'm ready to fall in love. [Music]
2. 0:20Boys were singing shinglings the summer that we met. [Music] You were 10 and 17. How could I forget?
3. 0:37>> I was shy to kiss you while the whole world could see. So shingling said everything to me. The way we danced was not a dance, but
4. 0:54more a long embrace. held on to each other and floated there in space. I was shy to kiss you while the whole wide world could see.
5. 1:11So shingling said everything to me. And all the poor old folks, they thought we lost our minds.
6. 1:25They could not make heads or tails with the young folks funny rhymes. >> You and I knew all the words and we always sang along
7. 1:38to old jam dong ding jaming dang dong. [Music]
8. 1:53[Applause] [Music] Happy Friday night, date night, people. I don't know, those of you on YouTube, you might notice I got myself a little
9. 2:08microphone. I'm a professional. I'm a professional. Now, look at this thing. Now, if it sounds like crap, that's cuz
10. 2:17I don't know how to set it up. Ah, Cindy [ __ ] is first. Cindy [ __ ] look at my microphone. Hello. We can do ASMR now.
11. 2:31Hi. I got a doggy. I think I'd be pretty good at that. You know what I'm saying? [Music]
12. 3:01Why you Why you getting so ramyammy? Why you so ramy, champ? [Music] If you haven't, uh it's Friday night
13. 3:09date night. Go get your nachos. Put them in the air fryer. Little uh little three or four cheese combo right on top of
14. 3:18there with some jalapenos out of the can. You know, get your hot pocket heating up in that dirty microwave.
15. 3:24Let's get this show on the road. I have I have some Kirkland sparkling water cuz I know how to live. Pinkies out cuz we're classy.
16. 3:37That's fantastic. It's It's going to make me a little gassy. What you got going? Oh, Kirkland. Nice. Producer brand. >> Unofficial sponsor.
17. 3:49>> Yeah, exactly. That's That's awesome. Oh, man. So, yeah. So, I'm a little discombobulated. I'm not used to having
18. 4:01a microphone in my face, but this will be good for comedy. You know, lean in for a comedic effect. You know, I don't
19. 4:08really know about this. This whole AI learning lab thing. Who is this? Kyle Shannon. I mean that, you know, right?
20. 4:15It's going to be good. It's going to be fantastic. All right. [Music]
21. 4:39[Applause] [Music] [Applause] [Music]
22. 5:04[Applause] [Music]
23. 5:27[Music] All right. What do we got going here? Let me just let me get things lined up.
24. 5:42Everything's different. Everything's different. You know, it's funny with ADD. Like when you have ADD, you kind of live
25. 5:53for the chaos, right? Except for [ __ ] like how my setup is. Like for example, let me show you something. This is [ __ ] hilarious.
26. 6:03Okay, so for two years, [Laughter] for two years, this was the stand that I put my my LED
27. 6:18light on. It was a cardboard box and there's nothing in it. So, it's all light. It's completely unstable,
28. 6:26right? And my little LED light sat here. Now, sitting right next to this was a little microphone tripod I've had for like 10 years.
29. 6:39So, for two and a half years, this was my light stand. And now look, I'm professional. [Laughter] ADD is [ __ ] weird, man. You got some
30. 6:55I must have some OCD mixed in there. Oh my god. It's all comedy all the time here at the AI Learning Lab.
31. 7:10Um all right so I was playing around with GPT5 how do you share the output you can so there's so in chat GPT
32. 7:27with any chat that you have go grab a chat I did casual conversation marketing email for far this was actually really impressive I I'll show
33. 7:40you this in a second here. Um, so so right up here, this this upper right hand corner. Oh, wait. Share your tabs. Share your
34. 7:52tabs, Kyle. Be a professional. Uhoh. What did I just do? Uh oh. What did I do? Hang on. Hang on, people. We just had a malfunction.
35. 8:08Calm down. Everybody calm down. Everybody calm down. Can you still hear me in the little microphone here?
36. 8:17Okay. I just lost my I just lost my monitor. We're going to be okay, people. It's going to be okay. Oh my god. I I just I should not be left
37. 8:35to my own devices for any kind of professional situation. Okay. Anyway, what I was saying before I was so rudely
38. 8:42interrupted with technology in the upper right hand corner of Yan here. Let me move this a little bit.
39. 8:51We'll do it like right there. Right here it says share. You can share that. Now you the thing to be careful with is if you share your chat
40. 9:01um it is not at all clear to me if you're just sharing it with the people you're sharing it with or if it's actually
41. 9:09available for everyone to go discover it. So just be careful with it. All right. This is off to a good start.
42. 9:18Tik Tok pin. 370 chats were published to a public URL. Yeah. Oh, 370,000 370,000 chats were published to a public
43. 9:27URL today by XAI. Yeah. So that share button with XAI today. Yeah. Shared shar shared them all publicly. And then I heard a similar
44. 9:39sort of thing with with chat GPT. So So there you go. Um All right. [Music] M.
45. 9:54Oh, lordy lordy lordy. Since we're here, so I discovered some all sorts of weird things today. The difference between
46. 10:06teams. So, I've got a team's account for my content evolution role. And in the Teams account, you've got instant chat GPT5,
47. 10:24you've got thinking, you've got Pro, and then you have your legacy models. Now, the legacy models don't show
48. 10:32anything beyond GPT40. Let me just make sure that I don't have them that I have it turned on. Yeah, I've got show
49. 10:40additional models turned on and it only shows GPT40, but it does give me pro. Okay, if I go to my personal account,
50. 10:55I also up here have Thinking Mini, which is actually my favorite model lately. I do not have Pro, which makes sense. But
51. 11:04then my legacy models, I have four of them. I've got 4041, 03, and 04 mini. I don't care about those other ones, but it's weird,
52. 11:14you know. I wonder actually now that I think about it. Sometimes chat GPT turning things on and off. Show additional models. Turn it off.
53. 11:33I go here. All right. Then I got 40. Let me come back in here. Turn it back on. I bet they show up now.
54. 11:47Nope. So, the legacy models don't show up in Teams and and Thinking Mini does not show up in Teams, but Pro does.
55. 11:55So, I don't know what the hell. All right. So, go figure. Um, one of the things that I talked about today on office hours, by the way, we
56. 12:07had a really fun office hours today because Jim Ross, who's got his own community of self- storage entrepreneurs
57. 12:15who he's teaching how to use AI, he said, "Hey, my meeting ends right before yours starts. Do you mind if I bring a
58. 12:21bunch of people over?" So, he brought over a bunch of people and we had a a really good a really good uh office hours today.
59. 12:31But in all three of my my I have three AI meetings in a row on Fridays. I've got my content evolution collab, I've
60. 12:38got office hours, and then I've got the AI mastermind, the AI salon mastermind. In all three of those conversations,
61. 12:47I kind of, you know, tested the waters for like who's got GPT5 figured out? And no one. No one. Everyone. Everyone is
62. 12:59sort of feeling a little discombobulated. Um I I have not recombobulated yet around chat GBT5. It just feels awkward and I
63. 13:09don't quite know how it works and I don't quite know which model to use when. Um and so if you're feeling if
64. 13:16you're feeling a little out of sorts because of chat GPT5, don't forget we've had GPT4 since March of 2023.
65. 13:29And so it's been more than two years where we've had we've gotten acclimated to how that model worked. So if this new
66. 13:35one feels weird, that should be expected. Now GPT6, I think, is going to come out before the end of the year.
67. 13:42Grock 5 is going to come out before the end of the year. Gro 6 might come out before the end of the year. So depending
68. 13:50on which tool you're using, you know, Gemini keeps releasing new stuff. You're gonna have We're gonna have a lot of
69. 13:57adapting to get used to. A lot of adapting to do. Vicki, fun office hours indeed. Indeed. Oh, yours went even more off the rails
70. 14:07today. Your chat GBT. Yeah, it's just weird. It's It's weird. The other thing is um one of the things that Altman said
71. 14:17I don't know a week or two ago is that their GPU usage is like they do not have enough GPUs to handle they have 700 million
72. 14:29users on a weekly basis and the stuff that we're playing with if you've noticed um image generation has
73. 14:36gotten really slow. They don't show the preview anymore. Um, I I think they have things quite throttled down right now.
74. 14:43So, if you're seeing inconsistencies and weirdness, um, it's because they're they're they're overloaded.
75. 14:51Um, Vicki, it was awful. It kept pulling conversations that had nothing to do with the project folder it was in. Oh,
76. 14:59yeah. Yeah. Projects are weird right now, too. The if if you don't know about projects, they're incredibly powerful.
77. 15:06You can basically group a bunch of chats into a project and then it'll just talk about those chats. But chat GPT also has
78. 15:14access to all of your other conversations. So it it doesn't seem to have good boundary control. You you know like that ex-girlfriend you
79. 15:24had didn't have good boundary control. That's chat GPT projects. Rationing of GPU clock cycles. I something like that. you know, they're
80. 15:35either rationing the the GPUs or just I I think what's probably more likely is they've got sort of levels of thinking that they can
81. 15:48move up and down. So, they're probably not doing it on a GPU byGPU basis, but they're probably doing it on, you know,
82. 15:55kind of features, what the thing can actually do. So, my guess is is that chat GPT is a lot dumber right now than
83. 16:02it'll feel when they have enough GPUs. Um, but I don't know. I don't know. That's just chat TMZ. Chat TMZ.
84. 16:12All right. What else you got going there, people? Oh, yeah. It's on Twitter if anyone wants to sympathize. Oh, you you posted
85. 16:22it on Twitter. That's good. That's good. Um the uh the guy last night I showed this guy Fernando who's a musician in LA and
86. 16:38he's um he's been playing with Sununo, right? He puts his songs in it. Well, he's put out a bunch more videos and his um his LA
87. 16:49Moon I think was the name of the song. the the the original video that he did, that video has gone viral. Um he's
88. 16:57gotten a whole bunch of hate from AI haters that are like, "You're evil. You're in bed with the devil doing this
89. 17:04AI stuff." And he's just like, "Hey, man. It's the future." Like, this is really cool technology. And he keeps
90. 17:09being impressed by it. Um so we can go look at a couple more of those if you want to tonight.
91. 17:28Brian Whitney, someone in the building. Someone is building a huge Dana Center here in northeastern Pennsylvania.
92. 17:35That's interesting. Techie Becky, I love getting my Friday night entertainment right here. That's right. You don't Let
93. 17:41me tell you something, people. This is a public service announcement right here. You don't You don't need Netflix. You
94. 17:47don't need Paramount Plus. You don't need Hulu. Hulu, Hulu. You don't need any of that. You come to the AI learning lab. You get you some
95. 17:57learning. You get you some entertainment. You got it all right here. Good lord, people. You think you got to
96. 18:06pay those Hollywood executives their high bonus salaries to create crap, import movies from Sweden. Uh-uh. You just come right here. I got a
97. 18:20microphone now. You don't think this is professional? You think this is just a crappy home office? You don't even know the sound
98. 18:27stage we got here. It's astounding. This is all just a curved LED screen. You You think those are actual shells? All right.
99. 18:40I don't think that was funny, but you know. Yeah. Thank you for the uh whatever that was on my head. API is application protocol interface.
100. 18:55It is an AP API basically allows one one piece of software to talk to another piece of software. So here a good way to describe an API
101. 19:06is let's say I'm creating a mobile app and in that mobile app I want to have a map function right so I can you know
102. 19:16people can find my business. Well, I could program an entire map application and launch my own satellites
103. 19:27and, you know, spend the next 3 years, 5 years creating a solid map application. Or I could say, "Huh, Google's already
104. 19:36done this. Why don't I just use their API?" And then basically what happens is you get a special key from Google and
105. 19:44you put it in your app and then you just make a little box and you say, "Stick Google's map here." and then it does it.
106. 19:51So that's that's kind of the that's kind of the thing. It it allows software developers to share what they've created
107. 19:58and make money on it, right? Every time you use the little Google Map application, they get, you know, a
108. 20:03hundredth of a penny or whatever it is and over enough of those, they make a fortune. [Music] Uh yeah, it's in the software stack. Um,
109. 20:16I finally got so fast frustrated with chat's memory I deleted it all and it still doesn't forget. Huh. Well, wait. But Tom, so in order to
110. 20:29delete it all. Okay, so there's a couple of things. Let's go let's go look at memory. So if you wanted to if you wanted to
111. 20:40really get rid of your chat memory, there's a couple of things you'd have to do. So, let's go into settings.
112. 20:49Turtle image generation didn't work. Oh, that's interesting. Um, let's go into data controls. So, improve the model for everyone. So, this
113. 21:05means it's sharing my data. Interesting. So I just turned Oh, I'm not sharing, am I? Yeah, I am. Wait.
114. 21:20Yeah, I am. Okay. Um Okay. I just turned off model improvement. So model improvement. This sends data to OpenAI. I just turned that
115. 21:32off. I had that on and I'd turned that off before. So that's that's one that they probably when they updated to five,
116. 21:39they're like, "Let's just reset everyone's preferences." Remove browser data. Let the agent mode remote browser reuse cookies between
117. 21:48sessions. All right, that's fine. I'm fine with that. But if you want higher security, okay, share links, archive chats.
118. 22:03So the date the debt management guidance I guess you can archive okay you can archive all chats you can export data
119. 22:18security multiffactor authentication secure signin with chat and then let's go to personalization so in personalization
120. 22:30you've got reference chat history which I had turned off. You've got reference record history. What's that? Oh, recording. Record
121. 22:39history. Let chat GBT reference all previous recording transcripts. I guess that's advanced voice. And I had I had reference chat history
122. 22:53turned off. I'm going to turn that back on. So, you could just turn off both of these without deleting stuff. The other
123. 23:00thing you can do is manage your memories. So these are the explicit memory. So there's two different kinds of memory in chat GPT.
124. 23:08There's the explicit memory where you say, "I want you to remember this." And then there's it remembering all of your
125. 23:13chat histories. And you can turn those on and off on personalization. So So Tom, you might want to just try turning
126. 23:21off both of those and see if that helps. Tik Tok gifts from Lord Digital Gods. Thank you, sir. Thank you, kind sir. always appreciated. Um,
127. 23:34[Music] so for you YouTubers, does this audio sound better? Does it sound too hot? Is the mic too hot? Is my voice annoying?
128. 23:50You're like, we were better when we couldn't hear you. I sound a bit tiny. I bet that's because that's too loud. Test, test, test. How's
129. 24:02that? Is that better? Testing, testing, test, test, test, test, test, test. Audio sounds good. Yes. Okay, so I was
130. 24:11clipping. Hang on. Let me actually go over to this software. Hello. Test test test. Okay, now I'm not clipping.
131. 24:20All right, there we go. All right. I'm a professional. You liked it before. Okay. All right. There's somewhere between where I was
132. 24:30before and where I was now. What's up, champion? Hang on. Hold, please.
133. 24:46All right. Champy. Champy lacked a little confidence.
134. 25:00He was hopping up on the bed. He was a little little scaredy cat. He was a scaredy dog. Um, all right. So, let me show you a uh
135. 25:13a prompt that I did today that was really quite good marketing email for pharma. Okay,
136. 25:30beauty. That's good. All right, let's see. [Music]
137. 25:46I saw Champ's relative yesterday all posting irregulars. That's cool. Um, okay. So, so today I was having a talk with my
138. 26:00co-founder at StoryVine and we were talking about marketing and one of the things going on in the pharma industry,
139. 26:05which is, you know, it's going on everywhere, but there's a lot going on in pharma right now, is big pharma
140. 26:11companies are doing like big cuts. They're doing like $3 billion restructurings and they're, you know, defunding,
141. 26:19you know, budgets for certain brands and, you know, prioritizing their core brands and dep prioritizing other
142. 26:26brands. There's all sorts of stuff going on. And, you know, historically, what we would try to do is navigate to
143. 26:36the companies and the brands that had bigger budget, right? and we would try to avoid the the chaos. And and what hit me today was, well,
144. 26:48wait a minute. What if we leaned into the fact that um that these pharma companies are going through tough times
145. 26:55and we market story because Storyine's a really efficient video creation tool. And so what if we leaned into it and I
146. 27:03thought well maybe chat GPT could help us identify like either who are the companies or what are the trends or
147. 27:09things like that. So I put it in GPT5 thinking mini which is my new favorite model and then I put it in did I do deep
148. 27:20research? I think I did do deep research. Does it say? Yeah. Uh, I don't remember if it was deep research or just
149. 27:28turned on the web web search, but it was one of them, but but I wrote, I want you to write me up a marketing email that
150. 27:34targets pharma. This was a bit of a a chain of consciousness, stream of consciousness kind of prompt. This this
151. 27:42isn't a very This isn't very well written, not very well structured, but chat GBT doesn't really seem to care
152. 27:49about that. You give it a bunch of context and it figures the [ __ ] out. I want you to write up write me up a far
153. 27:55pharma a marketing email that targets pharmas that may have had recent cuts or layoffs where cost cutting is more
154. 28:03obvious and I want to position storyvine as an alternative think something like just because budgets have been slashed
155. 28:10doesn't mean you need to give up video or if your salesforce has been reduced what if you could triple their
156. 28:16effectiveness with personal promotion and then I put these subjects are not literal suggestions but the kind of
157. 28:22positioning I'm looking for, right? So, just tell it, you know, don't copy me exactly because if you don't say that,
158. 28:29it'll just copy you exactly. It won't think. And then I said, then I had this f this further thought. Why don't you do
159. 28:36some research and identify the kinds of pharma scenarios where this kind of positioning could make sense, categorize
160. 28:43them, and write a marketing position for each category with an example email per category. And then I said, "Go." And it thought
161. 28:53for 24 seconds. And this was my the first response. This was my first response. I I find this just astounding.
162. 29:03Um that's probably a little too big, but that's a little better for TikTok. Um so it says, "You want a crisp,
163. 29:11deployable email copy and strategic positioning that leads into the reality pharma marketers are living in right
164. 29:18now." And so it did some research uh and then so so quick evidence snapshot and and these all have citations.
165. 29:26Layoffs and workforce reductions across biioarma continued into 2025 with thousands of roles cut. Industrywide
166. 29:33cost pressure is being driven by drug pricing and reimbursement debates. Pharma is actively shifting to omni
167. 29:40channel and digital first HCP engagement which that leans right into us. personalized video and short form video
168. 29:47s perform significantly better than generic content which leans right into us. So so two of the trends that it
169. 29:53found are right in the pipeline of StoryVine and then it and then here's the categories it found. Category A
170. 30:00salesforce reductions commercial team slowdown. Here's the scenario. Here's the positioning. Turn every
171. 30:07remaining rep into a high lever content engine. personalized compliant video outreach to multip that multiplies touch
172. 30:14points and preserves the human connection like this is right out of our positioning which it might have pulled out of our positioning.
173. 30:22um key benefits to the client, subject line options, right? So, here's subject lines and then here's an email example. Subject preheader,
174. 30:32there's your email, here's your little bullet points, there's your call to action, right? PS, we can spin up a 60 60-second
175. 30:41personalized sample video. Category B, marketing budget squeezes. And then same thing for that. Category C, mergers,
176. 30:49acquisitions, or major restructuring. There's a whole category for that. Category D, launches or life cycle
177. 30:56management. There's a whole category for that. Category E, regulatory or pricing pressure, right? And so because it's got memory,
178. 31:06it knows what Storyvine is. It knows what our value prop is. And then there's wrap up and next steps. So this was in
179. 31:14the three minutes I had in between calls, you know, I had a a conversation with my my partner and she was like, "Oh, maybe
180. 31:23see if you can write up a like an email." And so I wrote this prompt and like bang, 24 seconds later, or maybe it was
181. 31:3430 because it took six seconds to write this. Um, so that is an example to me where ChatGpt 5 significantly outperformed,
182. 31:49you know, four. It just it just went deeper. The writing's really good. The thoughts really good. And that's just
183. 31:55the mini model. Again, my my this thinking mini model for me is has been pretty pretty damn good. So anyway, thoughts, questions.
184. 32:09[Music] Oh, I got to do Black Bar. That's horrible. All right, side hustle Mimi. Sorry. Did I ever tell you guys about
185. 32:24the time I checked into a hotel in New Mexico? One of two. What does the agent in chat GPT do? I
186. 32:39haven't tried it yet. Okay, so this is really cool. Um, well, let me see. Chat GPT agent isn't really cool. It is
187. 32:55It's going to be really cool when it doesn't suck. And the only reason it sucks right now, the thing I mentioned earlier that that
188. 33:06chat GPT is completely overloaded right now. They don't have enough GPUs. And so, um, the agent capability is
189. 33:15really slow. Sometimes it just craps out. Sometimes it doesn't finish it thing its thing. But I I'll do a little demo here
190. 33:23and and we can go play with it. But um in a nutshell what agent does there's think of think of chat GBT kind of has
191. 33:33three basic modes. The first mode is just regular chat GBT mode. So when you flip into Oh wait two of two. Two of
192. 33:42two. They gave me a complimentary plate of nachos upon checkin. It was weird. Like I had I had luggage. That's really funny.
193. 33:58Oh my god. Yeah. Some some middle manager there at that hotel was like, "I got an idea." You know how what what what is the one
194. 34:08that gives you a cookie? You know how they got warm chocolate chip cookies? We're going to one up them. We're going
195. 34:14nachos. I'm talking jalapeno peppers, the cheese sauce. It's going to be awesome. How do people get it up to
196. 34:21their room? They just balance it on their luggage. It'll be awesome. Okay. All right. Here we go. Is this better?
197. 34:34That That's a little easier to read, right? Yeah, that's good. Except it puts me in the middle of everything. All right. Whatever.
198. 34:42Okay. Um chat GPT agent. So before I get in in demo agent, um there's there's kind of three different modes with large language models.
199. 34:59There's the basic chat single chat response. You put in a prompt, you get out an answer. So when you choose the instant model in chat GBT um
200. 35:16let me make a new chat here and just say um make me a poem about hotel nachos.
201. 35:33You put that in and it, you know, here's a little ode to hotel nachos and then out out it comes, right? You put in a
202. 35:39thing, you you you get out an answer. The second primary mode of chat GPT is this thinking mode and that is you put
203. 35:49in a prompt and it thinks about it for a little bit and it what it's doing is it's actually prompting itself
204. 35:55internally going well the user asked me for a poem and what should I do and so let me let me create a new a new chat
205. 36:03here we'll do it in thinking midi uh thinking mini write me a poem about nachos. Now, this may or may not kick
206. 36:11off much thinking, but let's see. Oh, I just about nacho. So, there it says thinking. It thought for a couple of
207. 36:18seconds and then it wrote a poem. So, that's not a really good example, but that's kind of mode two. And then
208. 36:24mode three is this agent mode. And what happens in agent mode is it takes that thinking model and it basically gives it
209. 36:31its own virtual computer. And so it literally creates like a screen where where it can do a bunch of things with
210. 36:40tools. So it can think, it can come up with a plan. So you tell it to go do some research. It comes up with a plan
211. 36:47and then it starts to execute that plan and it has access to tools. And so one of the tools it has access to is it can
212. 36:53surf the web. It can actually use its vision to like look at websites and click on them. The other thing it can do
213. 37:00is it can use deep research to go look at a bunch of websites and get all the data from them. The other thing it can
214. 37:06do is it can write and execute Python code and do math and create, you know, its own tools to accomplish whatever
215. 37:14goals you gave it. So, the way you you kick this thing off, I'll put it into full thinking mode. We'll go new chat.
216. 37:23And then you click plus here. And and you've got agent mode. And notice I've got 36 left that will reset on August
217. 37:3023rd. So tomorrow my things reset. So I should probably play with them tonight. Um so we're going to put it into agent mode. And then
218. 37:42you can turn on different sources if you want to, but I'm just going to have web search on is fine. And then I'm going to say um go research
219. 38:01Cost cutting activity in major pharma companies.
220. 38:18We'll just say in in in the pharma industry, analyze
221. 38:33what you find. Um, present
222. 38:49this as key trends that a company like Storyvine could
223. 39:04exploit and make me a presentation I can share with my co-founder
224. 39:19for new go to market strategy. All right. And so now I kick that off. And so it says setting up my desktop.
225. 39:32I'm going to make this a little smaller so you can see. Absolutely, Kyle. I'll research the most recent cost cutting activities. Blah
226. 39:40blah blah. Basically restates what I did. But now, okay, so notice this blue thing. I'm starting a text string based
227. 39:45search. And so it's now off it's now off using deep research to um to to find activity. And the this little blue this blue bar
228. 40:02around it, this is the virtual computer that it's spun up. It's spun up what's called a container where inside that
229. 40:10blue box, it's like you've got a whole separate computer, but chat GPT is the one out surfing the web. You don't have
230. 40:17to do it. It's just off doing it. All right. Um, find activity means what? I don't know. I mean, oh, because I
231. 40:26asked it to go find activity. That was part of my prompt. Analyze. Let's see. Go research cost cutting activity. So, so, so what this
232. 40:38thing does, I I should be able to There we go. So, along the bottom of this window is
233. 40:52is a little um slider. And what this is is this is actually tracking the history in this window. So, you can actually go
234. 41:02in. So the first thing it does, I'm going to research cost cutting activities and then you can kind of
235. 41:08slide this thing forward and you you see what it does. So it's it's basically recording all of its activity. Now some of these things
236. 41:17like this may take 10 or 15 minutes and I don't have to tell it what to do. like it's out finding this stuff and and one
237. 41:25of the things that it will often do is it'll create like a plan of here's the things I'm going to do and then as it
238. 41:31finds things sometimes it'll update that plan and realize oh I should change this to go in a different order and then I
239. 41:38should do this and based on what I found I'm going to do this but what this is doing is I can now go just spin up a separate chat GPT window
240. 41:47and this thing will just keep going. Kyle, can you test LinkedIn? Give me the best marketing contacts based on a job title.
241. 42:03Um, I can let me let me try that as a new as a new window. Um, I I I'll have to log into my own chat GPT, which means I'm gonna have to find
242. 42:17my own passwords. My own passwords. So, I might need to Brandon, just wait. Can you make you a dinner reservation? I'm so
243. 42:30sick of the dinner reservation thing. Let me let me see if I can get um uh LinkedIn up. I'm going to have to log
244. 42:44into it. But go ahead and share my screen, Brandon, and then the minute it it it kicks me over to log in then then just turn it off.
245. 43:01There you go. Okay. Um, so I've put it in agent mode and then what did you say? Test LinkedIn. Give me the best
246. 43:11marketing context based on job title. Okay. um help me identify um pharma marketers
247. 43:34with major launch brands that uh Wait, help me identify pharma marketers
248. 43:49uh with titles of decision making authority
249. 44:04on LinkedIn. Let's see on my LinkedIn. So that should kick us over to log in on my LinkedIn
250. 44:20who work for major launch brands that
251. 44:34recently launched or will launch. in 20 26. Put the names, links,
252. 44:50and contact info into a spreadsheet. Okay. So, I'm going to fire that off
253. 45:04and then we can flip back over to this other one. And this is still going. So like I don't know what it's doing. It's still doing
254. 45:14research which one of the things that is a absolutely mind-blowing if if you haven't played with there there's other
255. 45:21there's other tools that you can play with that have this similar functionality. One is called Manis manis.im m- anus.
256. 45:30There's another one called genspark. Gen s p a r k. Both of those are agentic kind of research tools. GenSpark in
257. 45:39particular is really good. It's fast. It's welldesigned and it has a lot of features. Um the chat GPT one is janky
258. 45:49and slow right now. They just launched it and it's kind of crap. Um so if you want to see a good one, go to GenSpark.
259. 45:55U Manis is also really quite good. It's just not as pretty. Um, but these these things are going to get very very capable in 2026.
260. 46:07These agents are going to be able to go do sort of remarkable amounts of stuff. They'll do it very very quickly because
261. 46:13even chat GPT can spin up 10 separate agents on its own. It may decide, hey, this is going too slow. I want to I want
262. 46:22to go do things in parallel. So that's going to these things are going to get quite amazing. All right, let's see. Understood. I'll
263. 46:32conduct research. Okay, let's see. I'm waiting while LinkedIn Oh, okay. So, this is interesting because I'm I'm logged into LinkedIn here.
264. 46:45Um, it looks like like I'm logged in to this LinkedIn right here. So, it's got access to my LinkedIn, which is cool.
265. 46:56That's the other thing. There's a there's a Perplexity now has a browser and Perplexity um it's called Perplexity Comet. And
266. 47:07what's cool about Perplexity Comet is it can do this kind of agentic work as well. And any tab that you have you've
267. 47:16logged into, it can access any tab. It can basically go outside of its its tab which which this is as well because I've
268. 47:23given it permissions to be able to do that. Yeah. God. Yeah. Yeah. They make the data slippery. This is the other thing
269. 47:32to know about these agentic things is privacy and security are assume that they're garbage because they're garbage. But assume that like
270. 47:44don't don't put anything in these things that that is not stuff that if it went out to everyone you'd be okay with. If
271. 47:51if it's anything less than that, you should be careful. Oh, this is this is really cool what it just said there.
272. 48:02I don't know quite how to go back, but it basically found this woman Jessica and it said she's a second degree
273. 48:08connection for you. So, it's it's identifying people that either, you know, I'm not connected with, it's it's
274. 48:15at least aware of who I'm connected with. It's pretty amazing. All right. Um, okay. Tik Tok is about to die soon,
275. 48:26isn't it? Uh, maybe. I don't know. Um, the White House just started a Tik Tok account like last week and they left their comments on
276. 48:40and and and democracy is is taking advantage of that particular little fact. Um, I don't know. Trump said he liked
277. 48:50Tik Tok, so maybe maybe they maybe it sticks around for a while. I don't know. Um it it does look like they're changing
278. 48:57their um community guidelines sometime this month and any of the political Tik Tok channels look like they're just
279. 49:04going to get decimated. So whatever. I think we'll be fine. But I don't know. Like the reason I dual cast on YouTube
280. 49:12and on Tik Tok is for precisely this. If it goes away, it goes away. Some people like Tik Tok, some people like YouTube.
281. 49:18So, but I've got more followers on Tik Tok, so we'll just keep doing it like this for a while. Um, something happened in this Tik Tok
282. 49:27streaming media since last time I watched. Some people's moved. I don't know what that means. block.
283. 49:48All right. Um, let's go look at the other the other agent. Is it still doing its thing? Still doing its thing.
284. 49:57How video training is revolutionizing pharma 4.0.
285. 50:17All right. Uh, let's see. We can go Oh, you know what might be fun? Let's go. Let's go watch some videos of what's his name? Um,
286. 50:29Tik Tok. Fernando.
287. 50:47Um, all right. Let's see. He does a rock and roll song. I turned my guitar into an orchestra. Testing Sununo with my most
288. 50:59complicated song. Tested Sunso with an anti-AI protest song. Let's do the orchestra one. Hopefully he talks a little louder here.
289. 51:32>> Oh, wait. >> That said he was completely quadriplegic. Hang on. >> All right, another test for Sunno.
290. 51:40Waltz for Doug. I was born and raised in Miami, Florida. I went to Miami Beach Senior High School. We had a teacher
291. 51:46named Doug Burus who taught classical guitar ensemble recording studio class and a thing called rock ensemble. The
292. 51:54thing that's crazy about Doug is that he was completely quadriplegic because of multiple scerosis. He could only move his head. And
293. 52:02>> let me give you context of this if you weren't here last night. So, so this guy's a working songwriter and music
294. 52:08producer in LA and two nights ago, two days ago, he posted a a video about, you know, my job, my job is in jeopardy. He
295. 52:18basically discovered Sununo's ability to take one of his songs and completely produce it. Um, and so rather than
296. 52:25hiding from that, he's been kind of leaning into it and learning what it makes possible and testing it and doing
297. 52:31it publicly. One of the things that I talk about all the time is play with these tools and learn out loud. Share
298. 52:38what you're learning. He's gotten a lot of hate for this, but he's also like he's establishing himself as a
299. 52:45thoughtful musician who may not like this, but he's also like these tools are remarkable and they're not going away,
300. 52:51so why would I not learn them? Which, you know, that's this is what he's doing is very much in the spirit of this channel.
301. 52:59you can hate this [ __ ] just don't ignore it. And that's kind of where he is. He's like he he's he says, "I'm a
302. 53:06I'm a combination of offended and impressed." Right? And that's that's where a lot of this stuff is >> us by telling us what to do. Um he
303. 53:17taught thousands of students for 40 years and he was amazing. And when my dad passed away, he became my second
304. 53:22father. I wrote this song for him. It's called Walts for Doug. It's a classical guitar composition. I would like Suno to
305. 53:29turn it into an orchestral thing. So, here we go. >> Can you hear this on YouTube? Oh, yeah. You YouTube is coming through this thing.
306. 53:37[Music]
307. 54:02Champy's singing. [Music]
308. 54:36All right, let's see what it does. Hit instrumental string quartet. Who's he playing the guitar champion? Did you like that >> classical?
309. 54:52No guitar. One, Mississippi. Two, Mississippi. Three, Mississippi. Four Mississippi. Five Mississippi. Six Mississippi. Seven
310. 55:08Mississippi. 8 Mississippi. 9 Mississippi. 10 Mississippi. 11 Mississippi. 11 seconds. [Music]
311. 55:22[ __ ] Today is [ __ ] All right, another test.
312. 55:38>> All right, for my next test of Suno, let's see. >> Hey, why does it keep changing? >> Hi, I'm Fernando. All right. Another
313. 55:50Mississippi five Mississippi 6 Mississippi 7 Mississippi 8 Mississippi 9 Mississippi 10 Mississippi 11 Mississippi 11 seconds [Music]
314. 56:08[ __ ] [Laughter] I bow down to you soon. [Music] >> That's such a that's such an honest dual
315. 56:25reaction. So I talk about in the stages of AI adoption, stage three is this stage of wonder and fear combined.
316. 56:34So his initial response is [ __ ] because he doesn't think Sununo can do the classical music thing, right? He's
317. 56:41tested it on rock and so he's like but I'm sure it can't do classical and then it does [ __ ] and then he's like I bow down to you.
318. 56:49Like he doesn't know what they're doing. He doesn't know how they're doing it but he's like holy [ __ ] Here's this new
319. 56:54thing right? I'm both offended and impressed. And that's if you're not feeling that weird pull, you're probably not deep enough. You're
320. 57:07probably haven't, you know, sort of peeled back the onion enough about your own personal work and the things you
321. 57:13care about. Like he clearly, this is his life, right? And so this is fairly existential stuff. You know, someone
322. 57:21made the comment here, I don't know if it was Side Hustle, Mimi, someone made the comment here about, you know, why
323. 57:26are people resisting this? so much. There's already so much technology out there that does a lot of this stuff in music.
324. 57:33It there there's something different about I mean this is for him to produce that song traditionally, he would have
325. 57:41to go find musicians or he would have to, you know, orchestrate it himself and use MIDI tools and things like that.
326. 57:49This is 11 seconds after him playing some crappy notes on the guitar. Um, you know, it it turned into this. So, it's
327. 57:56just the fact that it's generating music. It's not just processing our inputs. It's actually generating new
328. 58:03stuff is is what's freaking people out because he knows how long it would take to produce that track traditionally. [Music]
329. 58:14>> I wonder what Doug would have thought of this. [Music] That's pretty good.
330. 58:30[Music] My god. I wrote this when I was 17. [Music] I wrote this when I was 17. So he wrote
331. 58:46that little thing when he was 17. This is one of the things when when you watch Timberland who's a a music producer,
332. 58:53he's got a video of his first experience with this and he has a similar sort of reaction. He's like, "Wait a minute."
333. 58:59Like after he sort of after he sort of processes what's happening, he goes, "Wait a minute. That means I could take
334. 59:06any little SOG fragment I have and and explore it, right? And I could explore it quickly." Um, Tik Tok pin. Um, I
335. 59:16don't understand with how all the technology exists that people are still resisting or still scared of AI. Um, yeah, just
336. 59:23because just just because it's um, you know, it here here's a here's a thing that I I could be very off base with this. Um,
337. 59:37one of the things I hear a lot in the AI salon when when people first start playing with AI is that they feel guilty
338. 59:50that the thing that they can now do in 2 minutes or 5 minutes or 10 minutes that they know should have taken them three
339. 59:58hours or three days now it's just done that they feel guilty. Um, I I personally feel like
340. 1:00:14there's an overhang from the industrial revolution that time equals money. That in order for something to have value,
341. 1:00:23you have to put hours into it. It has to have toil in it. I think in the creative world, that that's a trope that is
342. 1:00:31particularly true. I've got to suffer for my art, right? We talk about suffering for your art. I think that's a very industrial revolution
343. 1:00:41mentality. And so I think the initial instinct of I feel like I'm cheating. I feel guilty. This is this feels like cheating is
344. 1:00:52because I shouldn't be able to create something of this perceived value in this little amount of time. And so I
345. 1:01:00think the AI resistance and fear part of it is that gap that it just shouldn't be that easy. And if it's that easy, what is my value? Right?
346. 1:01:12So it's not just that the product may or may not have value. It's that well wait a minute. If if my value is in the time
347. 1:01:21I put in to produce this stuff, what am I worth now that this stuff can just be done instantly? Right? That's that's the
348. 1:01:28kind of stuff that he's confronting right now. Tik Tok question. Do you know all of the content ownership issues with
349. 1:01:35Sunno? Can you explain? I am I am not an IP attorney at all. Um I have opinions on this. Okay. Couple of things. First of all, you need
350. 1:01:47to go go read the Puno terms of service. Um because I think I think my my co-founder just went and read them I
351. 1:01:58don't know three weeks ago because we were wondering about the rights for for using these for video projects. What they basically say is
352. 1:02:08you can do anything you want with this music but if someone sues you that's on you. Okay. So so they they are they are
353. 1:02:18severing themselves for any liability. Now, the thing about the thing about these tools, Sunno won't let you say,
354. 1:02:27"Give me a song that sounds like the Eagles." It'll block that that um that that generation. Um but you
355. 1:02:35could go into Chat GBT, for example, and say, "Describe in detail this specific song from the Eagles as if I were
356. 1:02:44explaining it to a music producer." And then you could take that over there and you could generate a song that sounds
357. 1:02:49like Hotel California, right? So if you're using these tools to produce content that is directly trying to sound
358. 1:02:59like copywritten works, you may get you may get banged for that. Now, that's that's likely not the the direct
359. 1:03:08question you're asking about, but I but it's an important thing that that if you push these tools into copyright
360. 1:03:14infringing territory, then you're infringing copyright and you know, you you may get sued. The the bigger
361. 1:03:21question is this. The first one is how did they train these models? Did they train these models on copywritten work?
362. 1:03:29almost 100% guaranteed that they trained these model on cop models on copywritten work. Almost 100% guaranteed.
363. 1:03:38Now the next question is how does it produce what it produces? Is it copying and pasting from those original works?
364. 1:03:46And the answer there is no. And that's one of the things that people most misunderstand about generative AI
365. 1:03:53is just because it can produce something that sounds like something it was trained on does not mean it's copying
366. 1:04:00and pasting from that original work. The way these things work is when the models are trained, the original data gets
367. 1:04:08turned into all these bizarre mathematical fragments that live in thousanddimensional mathematical space.
368. 1:04:15The original works actually cease to exist. There's there's current law out that basically says that the act of
369. 1:04:23training a large language model is what they call a transformative act. You're transforming in a significant way, in a
370. 1:04:30material way the original input document into something completely different. And therefore, if the training document was
371. 1:04:39obtained legally, it is considered fair use. If you stole it, then it's not fair use. But if you bought a copy of this
372. 1:04:48book and trained it into your model, because it's a transform, materially transformative, it is um it's considered fair use.
373. 1:04:58So when Sunno makes one of these songs, it's actually generating an original song that's never existed before.
374. 1:05:07And so based on that, you can use these things for anything. Now, can you copyright them? That's a whole other
375. 1:05:14thing. It depends. In his case, he wrote that song when he was 17. He can document that he wrote that song when he
376. 1:05:22was 17. He's got a Tik Tok video where he plays it into the tool and it takes that that song that he wrote and it
377. 1:05:29basically just reproduces it. It it produces it in a new and unique way, but it's his song, right? So, he's got a
378. 1:05:37chain of craft that he could probably take to the copyright office and say, "Yeah, this is my song. I put it in
379. 1:05:44there." You could potentially do the same with lyrics. You could potentially do the same if you do 400 iterations of
380. 1:05:50a song before the one that you publish and you track all that. Um, so this is incredibly complicated, incredibly
381. 1:05:58nuanced territory that I don't think is going to be settled for 20 or 30 years. So, so my, you know, my thing is this.
382. 1:06:08If you're working for, say, a big pharma company like I do at StoryVine, we need to be much more careful about what we
383. 1:06:15use. I think for most people, for most use cases, don't worry about it. Just go make music and play. Let the lawyers
384. 1:06:21figure this [ __ ] out. If you're trying to infringe on someone else's um copywritten material, then you're going
385. 1:06:27to get in trouble. But if you just want to use it as a regular creator, maybe not. Okay. Tik Tok pin. The the pin's
386. 1:06:32not there because I was talking too much. So, pop it back up there. Uh but really good question. >> Fascinating times we live in.
387. 1:06:52>> I still the three hours because my experience shaped the prompt. Yep. That so that that's that's interesting side hustle Mimi. Um,
388. 1:07:04I think that the the core business model of the advertising world is time and materials, right? How how ad agencies for the most
389. 1:07:14part make their money is by having 50 copywriters sit in a bullpen and write copy all day and charge $165, $250 an
390. 1:07:24hour for it, right? That's their business model. We now have a tool that can just generate copy in seconds. Um,
391. 1:07:34you can do what's called value pricing, right? You can say, "Hey, hey, you know, client, you need, you know, 50 Tik Tok
392. 1:07:41ads. Um, in a time and materials model, that would have cost you whatever $50,000. I'll tell you what I'm going to do
393. 1:07:51because I've got professional copywriters and editors who are going to make sure that the content's still good.
394. 1:07:57But now, but because we're using these AI tools, we can be more efficient. Rather than 50 grand, we'll charge you
395. 1:08:0435 grand. And then maybe instead of it costing, you know, the agency 15 grand to get that 35 grand, maybe it only cost them
396. 1:08:14five grand, so they can actually charge less and make more money. Um most agencies right now are resisting
397. 1:08:26that shift like the plague which which says to me there's a major opportunity to to have a a startup agency that can
398. 1:08:36undercut traditional agencies um and overd deliver under undercut them on price o overd deliver on output.
399. 1:08:45She slug of doom. I need to get my new glasses. I thought the Chiron had changed to Kyle Shannon, alligator. Oh, AI learning lab.
400. 1:09:01Okay, wait. We've got we've got results. Oh, wait. The one Hang on. The one result is still going. The original
401. 1:09:07prompt we did is still going. This was the one. What did I do? Oh, this was the farmer research one, but the the LinkedIn one is done. Okay,
402. 1:09:19understood. I'll conduct a list on LinkedIn to identify pharmaceutical marketers. Worked for 20 minutes,
403. 1:09:28right? So, here's here's what it did for 20 minutes, right? Did all this stuff. Was on LinkedIn. It was cruising around
404. 1:09:38LinkedIn. It wrote some code. Did it switch their tabs? >> Oh, did it? It did not. That's rude. Anyway, okay. So, it thought for 20
405. 1:09:49minutes. Um, I compiled a list of pharma marketing leaders with companies major launch brands. For example, Fierce Pharma reports that
406. 1:09:59Vertex's blah blah blah triple therapy is one of 2025's most anticipated launch. All right, here's here's our spreadsheet.
407. 1:10:09So, let's go look at it and see if it gave us Here's my prediction. So, it thought for 20 minutes. Oops. Oh,
408. 1:10:21this is better than I thought it would be. So, let me get pe rid of people's real names here. Um, so if found
409. 1:10:37one, two, three, four, four people from Vertex, two people from InMed, four from Astroenica, six from or five from Dichi Sano,
410. 1:10:51some from GSK, and some from Santa Fe. Um, this is better than I thought it would be. I my my guess is is that this
411. 1:10:59would have found um like two or three and been done. The fact that it found What did it find here? 20.
412. 1:11:07It's not bad. It's not bad at all. All right. Well, there you go. So, that's how agent works. You just turn the thing
413. 1:11:16loose. It goes does [ __ ] and then, you know, makes you a spreadsheet. Boom. Bing bang. Boom. Oh, and then this thing. Oh, no. No.
414. 1:11:28This thing's still going. Okay. That one's doing the research. All right. It's still It's still Oh, you know what it how training
415. 1:11:38how video training is revolutionizing pharma. I have a feeling this hung, which is one of the things it does. But
416. 1:11:53maybe not. We'll just let it keep running. Desktop view. All right.
417. 1:12:08The Chroma Awards in the Salon. Okay. Hang on. So, let me pop up the uh the salon address there,
418. 1:12:22Brandon, if you'd be so kind. All right. So, see that address right there? If you're not part of the AI salon.
419. 1:12:30So, so this live is brought to you by the AI Salon. The AI Salon is a community of about 3,000, more than
420. 1:12:383,000 now, 3,100 um AI optimists, people trying to figure this AI stuff out. Um if you're not a part of a community and
421. 1:12:48you're trying to get your head around AI, it's going to be very lonely because everything's changing really fast. It's
422. 1:12:53very hard to keep up with things. This is a really remarkable community. So, um sign up for this if you go to community.thesalon.ai.
423. 1:13:04AI. Um, that'll bring you here. Um, you can learn about our cycle of AI readiness, play first, mindfully create,
424. 1:13:11generously lead, what that all means, the five stages of AI adoption, our values. The second thing to do here is
425. 1:13:18introduce yourself. So, if you if you have not, go into the section called introduce yourself and introduce
426. 1:13:25yourself. And by the way, here's another thing that I would love to do. We did we did this the other night.
427. 1:13:31Um, if you're in a regular here, do me a favor. Go in to introduce yourself in the salon right now and just um, you
428. 1:13:40know, welcome people in and and uh, like give them give them a little a little hearty heart or an applause or respond
429. 1:13:49to them. Let's let because I know there's a bunch of new people in here. So, let's let Oh, that's really cool.
430. 1:13:55Let's let uh, some folks know that they are welcome to be here. thumbs up, hearty hearts, things like that. So, if
431. 1:14:03you if you all if you irregulars would go into the introduce yourself section and make people feel welcome, that would
432. 1:14:09be swell. Okay, now Vicki has put something. Where did Vicki put her her stuff? So, agent will do it, but the
433. 1:14:18plus account won't. No, agent, you have access to agent within the plus account. Hey, Kyle. Decided to pay for chat GPT, but disappointed.
434. 1:14:28still must wait if images docs are involved. No, I don't know what that means, Darren. You should be able to
435. 1:14:37generate images. You should be able to generate docs. You should be able to ingest docs. Ingest images if you're
436. 1:14:44paying 20 bucks a month for chat GPT. Now, the only thing that's possible, Darren, is if if you're doing it on
437. 1:14:51mobile and you didn't download the official chat GPT app, but you downloaded an app that's one of these um
438. 1:15:00uh parasite apps that's not really from ChatGpt. It's possible you paid for something that's not an actual uh
439. 1:15:08subscription to ChatGpt. Challenges and competitions. Okay, cool. So, so in in the AI salon, you've got start your
440. 1:15:16adventure. So, there's a little onboarding section here, um, in the community section, there's announcements
441. 1:15:23and a showand tell area and learn out louds where you can, um, teach things if you want to teach things. And then under
442. 1:15:30latest news, there's challenges and competitions. And so, the Chroma Awards are coming. Free trials for a bunch of tools. Very
443. 1:15:39cool. a film, music, and games competition toi unite creators, communities, and companies around the world. If we go to the Chroma Awards,
444. 1:15:51film, music, video, or games. Very cool. A6Z, LTX Studio, Bolt, Cling, Open Art, Fairground, and Image Art. So, the way
445. 1:16:03these contests tend to work, I am sharing this. Oh, I'm not. The way these contests tend to work is they want you to create content that
446. 1:16:16they're going to, you know, give awards for and things like that, but they cut deals with these these companies. 11
447. 1:16:24Labs, Freepick, F, um, Dreaming AI, Cap Cut, A16Z, which is they're an investor, LTX Studio, Bolt, Clling, Open Art,
448. 1:16:34Fairground, and Image Art. So, you will likely get credits for all of these different companies if you apply for
449. 1:16:43this contest. And if you're like, "Well, I don't know how to use any of those tools." Let's go back to to to the the AI salon for a second.
450. 1:16:54Let's go to the very first page you're dropped into. And the very first graphic that you see after our welcome video is this.
451. 1:17:06play first. Play first. So, if if you're thinking that you don't have the skills, you don't have the talents to join this
452. 1:17:14contest, [ __ ] Go do it anyway. Um, Becky Rue, who's in here, reminded me. I didn't realize that I did this to her,
453. 1:17:25but but when Becky Rue first joined joined the lives here, she she made the mistake of commenting, "I'm too old to
454. 1:17:32do this AI stuff." And I was like, "Bullshit. No, you're not." And then she she went and played. So, if you feel
455. 1:17:39like you don't have the the uh the talent or you're not creative enough to go make a film or a music video or a
456. 1:17:45game, guess again. You do. You just got to go play. So, go join this and play. All right. Ann Murphy in the house. Ann
457. 1:17:54Murphy, how does my voice sound? Huh? The new microphone. I'm all professional. Like, sound good. Um, so let me take you to another uh
458. 1:18:08another site. If you go to um you can pop this up on screen too. Uh Brandon, are you ready for ai.com? Are you ready? Number four, aai.com.
459. 1:18:23It's going to take you to this beautifully designed website built by Vicky Baptiste um who also did the instructional design
460. 1:18:33for the AI readiness training program. So, a couple of things to talk about here. Last between Christmas and New Year's in
461. 1:18:422024, we did this thing called AI Festivus, AI for the rest of us, where we brought in 34 speakers from four
462. 1:18:49different communities. We had thousands of participants um and and it was 24 hours of programming o across two days. So 12
463. 1:18:59hours at night um on the on Friday, 12 hours on Saturday in between Christmas and New Year's. We're going to do that
464. 1:19:07again this year. But what happened was all of the people that came got so excited about the presentations, they're
465. 1:19:14like, "This needs to be turned into a training." Because what happened last year that had that had never really
466. 1:19:20happened before was every single presentation even though they used AI tools were not training people how to
467. 1:19:28use specific AI tools. Every single presentation was about the mindset of how they think about using tools or how
468. 1:19:36they think about privacy or how they think about doing creative work when they're using AI. And so an and Vicki
469. 1:19:44and I got together and at the request of the community put together the AI readiness training program which is not
470. 1:19:51about tools but it's about the mindset. So there are five different training modules within the AI readiness training
471. 1:19:58project. It's a really unique program um and in incredibly powerful. So if you want to elevate your AI skills, check
472. 1:20:08this training out. It is. It's robust. There's a lot here. All of the videos from AI Festivus are in here. They're
473. 1:20:17sort of the source material that that you learn from. And then we extracted out of them all of these universal ideas
474. 1:20:25across five different sections from business to creative to privacy and ethical use of AI to putting it all
475. 1:20:32together in business. There's there's a bunch of different things. Um really powerful. So go check that out if you haven't. Okay. Um,
476. 1:20:44[Music] all right. Fantastic. Fantastic. Yeah. All right. What's happening, Ann Murphy? Anything going on? Anything else we
477. 1:20:57should talk about? Oh, the other thing we should talk about is AI Festivus is happening again in 2025.
478. 1:21:04So, December 26th is a Friday. December 27th is a Saturday. So, we're going to do the same thing again. It's going to
479. 1:21:13be split over two days, 24 hours of programming. The one thing that we're going to do this year that we did
480. 1:21:18different from last year is everyone was so excited last year to hang out after the event. Um, we're going to have an
481. 1:21:25official hangout area after the 24 hours. Anna and I are going to go to bed because we'll be hosting the damn thing.
482. 1:21:33So, we're going to go to bed, but you all are welcome to hang out. So, um, we don't have we don't have the, uh, the,
483. 1:21:40uh, anything up and running just yet, uh, for you to sign up for it, but just save the dates of December 26th,
484. 1:21:49December 27th, the Friday and Saturday between, um, you know, Christmas, Hanukkah, those kind of things, and New
485. 1:21:56Year's uh, is is when it's going to happen. All right. Oh, yes. Yes, there is another thing. So, one of the things Ann and I put
486. 1:22:07together as a as a parallel initiative to the training program is we created the AI readiness project, which is a
487. 1:22:16podcast where we interview people and we talk about how ready for you, how ready for AI are you and what does that actually mean? And um so
488. 1:22:27so if you go to aire readiness project.com um or you go search for the AI readiness project on um Apple podcasts, Spotify
489. 1:22:40podcasts, I think it's on Amazon podcasts. Um please go subscribe to your podcast delivery vehicle of choice and
490. 1:22:49do whatever you do there. I think you subscribe to our podcast, but leave a review, listen to them. Um, we just we
491. 1:22:57just celebrated our 25th episode, which if you've ever done a podcast, most podcasts die before they hit double
492. 1:23:05digits. They die at nine, almost all of them. So, the fact that we hit 25, I don't know. I'm a little proud. I don't
493. 1:23:13know about you, AM, but I'm a little proud. Um, how do you feel about MIT saying 90 91% of initiatives are not meeting
494. 1:23:22expectations? Um, I'm surprised I'm surprised that Well, I think it was 95% not 91%. Um, I'm surprised that that
495. 1:23:32many are successful here. Here's here. Okay. Almost all AI initiatives that big
496. 1:23:46corporations are rolling out right now are being rolled out as IT initiatives where they like okay you got access to chat GPT go
497. 1:23:58and people like uh okay what do I do with this the people aren't being trained um these initiatives. Look,
498. 1:24:14the reason it makes sense for me to go live five nights a week here is because nobody knows exactly what to do with
499. 1:24:20these tools because we've never had capabilities like this before. One of the mistakes that big corporations are
500. 1:24:27making right now is they're treating generative AI like they've treated historical AI, right? So, okay, we use
501. 1:24:33Salesforce, we're going to add AI tools into it, and we're going to roll that out to all the people. That's a very
502. 1:24:38specific vertical use case where if you're doing a sort of vertical use case like that where you're where AI is
503. 1:24:45amplifying an existing process, those are probably more successful because you know exactly how to do the process
504. 1:24:52before. Now AI is making that a little bit easier. But if you just take generative AI as a general capability
505. 1:24:59and you throw it at the organization, nobody knows what to do with it. Nobody knows how to use it. Nobody use knows
506. 1:25:08why to use it. What does success look like? So it is not surprising to me at all that these initiatives are failing.
507. 1:25:19What the impact of an article like that are going to have this is [ __ ] insane. We're we're we're just we're going to just watch a
508. 1:25:29series a rolling series of Kodak moments where where big companies are are going to start getting their asses handed to
509. 1:25:36them by little startups. And and what's going to happen is this. People are going to take that article and they're going to go, "Oh, AI was
510. 1:25:45just a bubble. I told you it was just hype. It was just like NFTTS and blockchain. It was just hype. We can ignore AI now."
511. 1:25:54and and so so focus on a AI in a lot of corporations is probably going to get pulled back. Now what's going to happen
512. 1:26:02is smart corporations that put together little pods of people to actually explore generative AI and figure out on
513. 1:26:09a department by department basis um what generative AI actually makes possible and do things from the from the workers
514. 1:26:19up rather than from it down. those companies are going to start to innovate and start to start to really
515. 1:26:28leverage the [ __ ] out of AI. So, I think what you're going to see is you're going to see a bifurcation starting this year,
516. 1:26:34a bifurcation of companies that pull back on AI investment and those that invest more heavily into it and actually
517. 1:26:41understand how to train their employees, how to roll these things out in a smart and ethical and uh intelligent way. and
518. 1:26:50and those two kind of companies are going to separate really quickly because the ones that lean into it and really
519. 1:26:56figure out how to use this stuff are going to start to amplify what they do. So that's that's that's where I think
520. 1:27:02this goes. Um so I'm not surprised at all that that article came out. You you know media companies are going to do
521. 1:27:11sensational headlines like that. um most people are not going to read the articles and they're going to completely mischaracterize why
522. 1:27:22why the success rate is so low and and it's it's it's not going to be good for them. So anyway, there you go. There you
523. 1:27:29go. There you go. There you go. There you go. There you go. Um we won't consult with any org if they
524. 1:27:37have us work with it. That's Ann Murphy. So Ann Murphy is is doing this on a regular basis. She's working with
525. 1:27:44organizations to teach them how to use AI to think about it. And yeah, if if you think about okay, think about AI
526. 1:27:56like you think about the internet and the worldwide web. When when when I started building websites in the mid 90s
527. 1:28:04for big Fortune 500 companies, one of the very first things that we did as an agency was we would go in and we would wrestle
528. 1:28:15the control of web development away from it. Now, I'll tell you a really specific example of this. CocaCola put out a in
529. 1:28:25this was I think 1996 they put out an RFP for an agency of record pitch for people that for agencies that knew how to build websites.
530. 1:28:38And so we got invited to pitch it. We took all of our best people. We're like this is Coca-Cola. We've got to it's got
531. 1:28:44to be polar bears and refreshing graphics and we've got to figure this [ __ ] out, right? So, we put together this really
532. 1:28:55beautiful pitch. I probably have it here somewhere. We put this together this really beautiful pitch. We went down to
533. 1:29:01Atlanta. We're sitting in a room with about I don't know, seven or eight or nine Coca-Cola executives and seven or eight or nine of us
534. 1:29:10and you know, we're talking about what the web is and why it is and what we've done and things like that. And then we
535. 1:29:17get to this slide. And what the slide was was two screenshots. One was a screenshot of Coca-Cola.com and one was a screenshot of coke.com.
536. 1:29:28Both of those pages had been built by the IT department. One was light yellow with blue hyperlinks and the hyperlinks
537. 1:29:37went to all the annual reports in the history of the Coca-Cola Corporation. And the other one was a white page with
538. 1:29:44black hyperlinks that went to I don't know uh other documents some documents that it thought were important
539. 1:29:54and and our headline for that slide was where are the polar bears. The Coca-Cola executives stopped the meeting. They said they said stop.
540. 1:30:10We'll be back. They all left. They all walked out of the room. We're like, "What the [ __ ] did we do? What's
541. 1:30:21going on? Are we getting fired? Like, are they going to ask us to leave? Are they getting security? What's happening?" Um,
542. 1:30:30they had never seen their website. They had never gone to it. They've never looked at it. when they saw how their
543. 1:30:40brand was being represented to the world, they absolutely flipped out. They walked back into the room after about 15
544. 1:30:48minutes and said, "Can you start Monday?" No one had ever shown them this before. So, it's not that IT groups are bad or
545. 1:30:57evil. It's that the worldwide web and generative AI affect every single part of the organization from marketing to
546. 1:31:07operations to sales to product all of it. And you and it's it's not just like an enabling you know CRM that it has to
547. 1:31:19install and maintain. It's a completely new capability and it's got to be thought about. I just got restricted. This is primarily
548. 1:31:33directs people off the platform. This is in line. Whatever. I just appealed it.
549. 1:31:50whatever restriction I want to hear more and talk about this forever. Yeah. So, so the the thing is is
550. 1:32:01if I if it is treat is is treating chat GPT like it like like it's Salesforce. It's not. It's just it's a fundamentally
551. 1:32:12different thing. Like nobody knows how to use generative AI to run their business right now. Sure, you can do
552. 1:32:20automations. You can use AI and something like Zapier or NATO to create an automation to make an existing
553. 1:32:28workflow more efficient. You can do that. A lot of people are doing that. A lot of consultants are making money doing that.
554. 1:32:35But that's using AI to do what you already know how to do. Generative AI lets you do things in completely new and
555. 1:32:43fundamental ways that you can only discover through innovation and experimentation and play. And that's not the purview of it. It is
556. 1:32:53there once we have something locked down. We're going to make it solid and more efficient. That's not the stage
557. 1:32:59we're at with with AI right now. We've got another decade to figure out how we use this stuff, how it impacts all the
558. 1:33:06different areas of a business. So anyway, that's my two cents on that. Um, [Music] the people side is so important. The
559. 1:33:20people side's everything. Listen, these are large language models. These are not large programming models. They're not
560. 1:33:28large engineering models. They're large language models. So how we interact with them is very different than how we've
561. 1:33:35interacted with technology in the past. And so this is all about people skills and connecting and speaking and communicating because
562. 1:33:45one set of skills is how you interact with people. The other set of skills is how you take that language and you
563. 1:33:51interact with large language models and get the best results out of them. They're very related. Right? So there
564. 1:34:00you go. There you go people. You understand what I'm saying? All right. All right. Yeah. Um, let's go see if our other agent is still stuck.
565. 1:34:26Still stuck. Yeah, this hung. This hung. Wait, maybe it didn't. What is
566. 1:34:41share this tab? Um, here I am. This thing looks like it's still stuck here. Um, I'm just going to say where is my
567. 1:35:01presentation? So, remember before when I said this thing is janky and it might not work all the time.
568. 1:35:09This is janky and it didn't work all the time. This one sort of worked. The one the one where someone asked me to go
569. 1:35:14look at my LinkedIn and find, you know, target target uh prospects. That worked. This one did not. Okay. I'm still in the process of
570. 1:35:25compiling research and crafting my presentation based on the latest cost cutting trends. I'll focus on how these trends align with
571. 1:35:33StoryVine's offerings once I've completed. So continue. [Music]
572. 1:35:47All right. I guess we we got it back going back live again. All right. Whatever. Um, here here's another thought I have on on
573. 1:36:02these agent things. Even if you go play with Gen Spark or Manis, which are which are more capable than chat GPTs. The reason chat GPTs is
574. 1:36:14significant is they've got 700 million weekly users. Now of those probably 37 are using chat GPT deep enough to
575. 1:36:23understand what agent is right it's like not a lot of people are using this right now but you know relatively speaking but
576. 1:36:31as this tool becomes more transparent and more effective you know people are going to have instant access to
577. 1:36:37incredibly powerful tools oh it looks like it's building us our our deck now the report has been created all right good Um, don't
578. 1:36:51one of the biggest things I would caution against right now is don't dismiss any of these technologies
579. 1:37:00because they suck right now. You know, one of the things we always say in here is, you know, these tools are as bad as
580. 1:37:06they're ever going to be. The fact that this thing can go like look at my LinkedIn and find people and do that
581. 1:37:12without me having to be there to do it. The fact that this can go do research and create a presentation for me even if
582. 1:37:19it's janky and even if it's a crap report means that imagine what that's going to be like a year or two from now. So by
583. 1:37:29playing and by exploring what these tools can do today, you'll be ready when these things get good. And when these
584. 1:37:36things get good, it'll be you will occur to the rest of the world like a wizard. Because very few people know what's
585. 1:37:46coming. Very few people know what's here. Even people that say, "Oh yeah, I use chat GPT all the time." Do you know how
586. 1:37:54many people like on a weekly basis tell me, "Oh yeah, I use AI all the time." And then I show them some simple thing
587. 1:38:02within chat GPT, they're like, "Oh, I didn't know it could do that." It's like regular all the time. So, so
588. 1:38:11just keep coming back to these lives. Join the AI salon. Hang out with people. Build relationships with people. Okay.
589. 1:38:20Spin B3. Dip time. Good night. Peace out. Spin B3. Try three months, not two years. Yeah, I know. I know. This stuff is moving too fast.
590. 1:38:33All right. This is now This is now successfully building us slides, which is cool. So we should we should have
591. 1:38:39slides to uh to look at here shortly. Worked for three minutes. Yeah. After it hung for an hour. All right. So there's our report.
592. 1:38:54It looks like it's in a canvas. Report comprehensive thing. Um is this a canvas? Maybe not. Cost cutting activity
593. 1:39:09industry contest shrinking margin and patent cliffs force restructuring. Merc's cancer drug Kituda. Did you know
594. 1:39:18Kituda for Merc is 80% of their revenue? 80% one drug. It's like it's got like 46 indications, 46 different ways it can be
595. 1:39:29used to battle cancer, but it's coming off patent. pipeline prioritization, outsourcing, digital transformation and AI adoption,
596. 1:39:43generative AI and analytics for efficiency, decentralized clinical trials, implications for StoryVine, reducing marketing and comm's cost,
597. 1:39:54supporting decentralized clinical trials, scaling training and compliance programs, enhancing supply chain visibility, internal comms,
598. 1:40:06Conclusion. There's that. Where's my presentation? Oh, here's my presentation. Look at this. Preview. Play slideshow.
599. 1:40:16All right. So, whoever asked about this before, what's an agent do? It went and did some research. Even
600. 1:40:23though it hung and I had to kick it, we now have a deck. Now, is this deck any good? I don't know. Let's go look at it.
601. 1:40:30Um, one thing one thing that I do know about uh chat GPT's presentations is they're ugly as sin. They're just really
602. 1:40:37bad uh visually. Cost cutting drivers key trends and operations opportunities for storyvine go to market strategy. Why
603. 1:40:46cost cutting? Almost 60% of pharma executives plan to optimize operations and reduce costs employing restructuring, offshoring and layoffs.
604. 1:40:5660% Um 6,000 jobs cut from Merc. Yep, that just happened. Mona, so these are real.
605. 1:41:1110,000 out of Mona. Biotech. 5,000. Generative AI and automation. Content cost reduction.
606. 1:41:26Generative AI adoption yields up to 12% savings. AI could generate 60 to 110 billion annually. Content creation costs
607. 1:41:36drop 30 to 50%. Which is that's that's right where our company plays. Digital marketing of engagement video
608. 1:41:49content retains 95% of a message. Text only 10. This these are good numbers. 84% purchase conversions, 56% increase in ROI for video
609. 1:42:03impressions, digital training, decentralized trials. There's opportunities for StoryVine. Go to market strategy. Target the leaders
610. 1:42:12digital training. All right. Conclusions and next steps. Pilot StoryVine with marketing and training teams. Integrate into
611. 1:42:21decentralized trial platforms. There you go. All right. Not great, but fine. Fine. Fine. Fine. There was some There was
612. 1:42:31some new and interesting stuff in there. All right, everybody. I'm getting a little tired. We You having fun tonight? Hey. Yeah.
613. 1:42:59All right, let's see. Let's see what we've got here. [Music] Any comments? Any thoughts, questions? All right, everybody. Uh, I'm going to
614. 1:43:13take my new fancy microphone to bed with me. Hello, little microphone. um homework for this weekend.
615. 1:43:28Please, we don't need to hear that. Me touching the microphone. Yeah. The one thing I noticed about this microphone is it's got really bad
616. 1:43:38um if you touch it like like good microphones if you touch them, you don't hear anything. This one's got really bad
617. 1:43:44versions of that. No, you mean Oh, mic in bed. Yeah, you know, great demo, Kyle. Thank you. Thank you,
618. 1:43:56Champion Irregulars. All right, let's go look at Champion Irregulars. Hey, did you all go um connect with
619. 1:44:07people, give them thumbs up, and and respond to their introductions in the salon? If you didn't, please do. If
620. 1:44:13you're an irregular, please go welcome some people in. Um, okay, the Irregular's channel. Let's go look at Ireulas. The Irregulars.
621. 1:44:26Yeah, Champy's cousin. That does look like Champy. Although this is like the skinny version of Champy. Champy's face is squattier,
622. 1:44:38but very cute. Oh, wait. I'm not sharing. Share this tab instead. Look how cute. Yeah. Nice, nice, nice. What brand is it? It is a
623. 1:44:54Mayono. The Mayo Mayono 300 300 DX or 300X. Yeah, Mayono 300X.
624. 1:45:08And it came with a little swing arm. And I was going to get the 400X, but This is newer, so I got it. I got it.
625. 1:45:20Yeah. Here you go. All right, people. Tik Tok question. Do you have a site we can check to learn about setting up agents? Um,
626. 1:45:37I do not. What I can tell you is, well, okay, setting up agents. I don't know what setting up agents means. One of the
627. 1:45:46things that's happening right now is everyone is calling everything an agent. An agent might be like a chatbot that
628. 1:45:55you interact with like an agent like that. That's that people are calling those agents. People are calling
629. 1:46:01automations agents and then people are calling agentic tools agents. The only thing that are really agents
630. 1:46:09are the agentic tools. Um so it depends what you mean by the question. None of the agentic tools in my opinion are
631. 1:46:18really ready for prime time right now. They're they're all kind of in experimental mode. So what I would say
632. 1:46:24is the the four things I would recommend playing with are Chat GBT agent mode, Genpark, Manis, Mus, and then the Perplexity Comet browser.
633. 1:46:38if you can get access. I think it's in it's not open beta right now, but it's in like invited beta. People people who
634. 1:46:46have access to it might be able to get you an invite code. Um, those are the four that I'd play with right now. There are probably a lot of
635. 1:46:54other ones, but those are kind of the major ones that you can at least start to see the state-of-the-art for agentic
636. 1:47:02workflows for automate fully automated things that are going out and using tools on your behalf. Um, and then I would just say be patient
637. 1:47:13about it. Um, if you want to do things like build automations, there are lots and lots of courses and workshops,
638. 1:47:22people offering that. Um, and then if you're talking about agents where you're interacting either with voice or text or
639. 1:47:30like a a video call, those kind of agents, that's a whole different category. Naden has tutorial on their on their
640. 1:47:39site. Yeah, a lot of them probably have tutorials. There's probably a lot of YouTube videos. Quite frankly, you can
641. 1:47:44just use chat GPT. Put chat GPT in voice mode, turn on the camera, and point it at your screen and have it give you
642. 1:47:52instructions on how to set those things up. That might be a way to do it as well. Thank you for the suggestions. Love what
643. 1:47:59you're doing. Very inspiring. Thank you very much. I appreciate that. Very, very cool. Um, cool. All right. So, it is the
644. 1:48:05weekend. I hope you guys have a fantastic weekend. I would continue if I were you. I would continue to explore,
645. 1:48:12start pushing all the buttons in chat GPT. Let me let me actually share my screen and show you what I would play
646. 1:48:19with over the weekend. Other than my microphone. Okay. Um sharing this tab.
647. 1:48:35All right. So, here we are in chat GBT. First thing I would I would experiment with if I were you, if you haven't done
648. 1:48:42it yet, if you go to your settings and go to customize chat GBT, this is what's called your um custom instructions. You can play around with
649. 1:48:54all of these things, but there's one one little hidden right hidden in the middle of this is this little line right here.
650. 1:49:02What personality should chat GPT have? And if you click on this pulld down menu, there's five choices. Default,
651. 1:49:09cynic, robot, listener, and nerd. Take the same prompt. So, take a prompt that you've used before or take a prompt
652. 1:49:18where you know you're wanting to get a particular kind of answer. Use the same prompt flipping into these different modes and
653. 1:49:28just look how it changes how chat GBT responds. What you might find is that one of these personalities changes chat
654. 1:49:38GPT into something you absolutely love where otherwise you might not have liked it or vice versa. There might be one of
655. 1:49:45these that you absolutely just hate it. Um someone today told me they were like um they were like why is chat GPT being
656. 1:49:53so mean to me? Oh Dr. J Claire was like all of a sudden chat GPT started being really tur and and sort of snotty to
657. 1:50:01her. She's like, "What's going on? Does it not like me anymore?" She had it set to robot efficient and blunt, right? And
658. 1:50:09she forgot to turn it off. So, go play with those. That's a thing I'd experiment with. Then the other thing is um
659. 1:50:18in your in your model chooser, the default mode is auto, right? And so what happens in auto mode when you interact
660. 1:50:26with chat GPT, it is deciding which of the different models to use. What I would do if I were you is manually switch into these different
661. 1:50:37modes and again take the same prompt and apply it in different modes and see what answers you get. See the different kinds
662. 1:50:46of answers you get. You can also under legacy models, you can go back to GPT40, which is how chat GPT used to used to act,
663. 1:50:55but then use GPT5 instant, GPT5 thinking mini if you have it, GPT thinking if you have it. They're very, very different
664. 1:51:05models. My favorite right now is thinking mini. I don't know why. It's just I like the answers it gives me and
665. 1:51:14its creative writing ability is way better than uh GPT5 instant which sucks. I think this model sucks. Thinking Mini is the one I like. But
666. 1:51:26that's what I would do is just keep playing and experimenting with GPT5 because we're all trying to figure out
667. 1:51:33how this works. What model are you using for your custom GPTs? I have not even touched my custom GPT. Chef Clly, I will
668. 1:51:40probably flip them into I'll probably leave most of them in GPT40 mode if I can. And then I think there's a point at which they
669. 1:51:51automatically flip over. I don't know which one I'll use. I might use Thinking Mini. I just I like that one better
670. 1:51:56right now. I feel like their GPT5 Instant if it it feels like it must not be a My guess is they have a better version
671. 1:52:07of that model. that's too big for them right now because they don't have enough GPUs. So, I have a feeling that's a
672. 1:52:13crappy that's a crappy model. So, that thing may get better over the next couple of months, but right now it
673. 1:52:19sucks. So, I don't know. I haven't really thought about my custom GPTs. I can only pick GP T5 thinking. No Mini
674. 1:52:27for me. Yeah, I don't I honest to God don't know why. Um on on one of my calls today, I had a couple of people on the
675. 1:52:34call had access to Mini and some did not. I don't know why. Again, this is probably OpenAI doing stuff to manage their GPU usage is
676. 1:52:45my guess. Um, we'll be back Monday. Tuesday, there's no AI salon, so it's just a normal week next week. So, have a great weekend,
677. 1:52:56everyone. And, uh, I will see you I will see you Monday. All right. Have fun this weekend. Peace out.