5/6/2025 - Agentic AI Tools: A Deep Dive and Comparison

Transcript

1. 0:00Come on, you little fat doggy. Hey, you little fat doggy. What's going on, you little fat doggy? Huh? [Music]
2. 0:20Hey, what song is this? [Music]
3. 0:36[Laughter] [Music]
4. 0:52I met my baby at the F club underneath the town cafe on a corner life and four street late one Wednesday she offered me a red strap I
5. 1:05said baby not a chance I could go for tall cool ginger ale how about we just dance we did the boom shaboom the do I
6. 1:14did it mocha chokata and the nitty-gritty we did the coochie cool the lazy ballerina baby make me feel like I'm a California dreamer [Music]
7. 1:32Well, I asked her for her number. She said, "Here's my phone at work." I put two and two together. Damn, she must be
8. 1:39with some jerk. Two summers later on the vineyard came another opportunity. She came all that way just to see me play,
9. 1:47but I was with Sweet Lisa Marie. After so many months of waiting, well, finally come that spring, she was free and clear
10. 1:56and good to go. And hot damn, we did our thing. We did the boom shaboom the rumble in the jungle. The who knows who
11. 2:03in the hast stack shuffle did the I don't know the words and the something something girl from me.
12. 2:18Yeah. Forgot that song. That's a good old Martin 6 to one called Boom Shaboom. Good evening, good people. Hope
13. 2:31you're doing fine on this evening. It's late. It's late for a very important date. [Music]
14. 3:08Oh, you and I here all along. and all Sunday morning at
15. 3:23home. Sky's blue and the coffee is strong. It's true. Then I open my eyes to a dream realized in front of
16. 3:40me and I haven't got a clue what in the world is happening to me. Think I think I'm happy like first day summer vacation.
17. 3:55Happy got to get a little rest and relaxation. Happy like a choir on Sunday morning singing true. [Music]
18. 4:37[Music] Hello. Hey. Hey. What are you doing? Why are you rapping? What are you rapping for? [Music]
19. 4:51We're sing We're singing We're singing just folk folk music. Folk tunes. We're not getting all aggressive. We're getting all
20. 5:02aggressive. Sometimes these dogs, they just don't know how to stay in a genre, you [Music]
21. 5:18know? Feel like I'm watching three different views of Kyle, but they're all different points of [Music]
22. 5:30lag. Oh man. Side hustle meet me in the house. What's happening? What's happening? What's going down? Yo, fantastic cams. Yeah, we do that.
23. 5:43We'll do that. That does look a little slower tonight, doesn't it? Something's not right. Streamyard. Oh, Tik Tok. [Music]
24. 6:07[Music] Cam Katkin, I feel so lost with agents.
25. 6:22Say more. Like you're trying to you're trying to do them and they're just not behaving or you're feeling like they're
26. 6:29just off doing [ __ ] and you don't have control. Like what is it about agents that you're not grooving on? I mean, if
27. 6:36you're frustrated with them, I would say we don't really have agents right now. We've got semi semi-agentic things.
28. 6:43We've got manis, which is agenty. Or are you trying to build them? [Music]
29. 7:12Um, I don't think we've gotten we haven't had the chat GPT moment yet for agents. I don't know who it's going to
30. 7:22come from. I mean, Manis is a little bit. Um, I would say I would say uh Gen Spark is probably slicker slicker still than Manis.
31. 7:38But Manis feels a bit more robust. Anyone tried the Lortix Suna open-source genurpose agent? I have [Music]
32. 7:50not. I have no idea how to use Manis. Is that something you could talk about or Gen Spark? Yeah. Um, yeah. Let's go. We'll we'll go do
33. 8:02something. I mean, [Music] to a great [Music] degree. You can you can almost think of
34. 8:17it like this. You can almost think of it like this. Okay, we're going to the boxes. We're going to the boxes. We're going to the
35. 8:31boxes, people. Just hang on. Hang on. Just calm down. Everybody calm down. We're going to get
36. 8:46there. But we've got the toolbox, too. Ah, okay. I Ah, okay. I got it. Okay. I got it.
37. 9:04Okay. Okay. It's good. It's good. Good. Good question from the crowd. Good. Good question. Cam Ken from Cleveland. Okay.
38. 9:15So, here's what we're going to do. Here's your large language model, right? That there's your GPT. And
39. 9:29then here's your chat GPT. All right. And that's going to be this is going to be your basic 40 model. That might look
40. 9:38like magenta to you, but that's just 40. So here's all your knowledge. We took all of the knowledge of humanity. We put
41. 9:44it in this box. Okay? And then we wrote a little chatbot that you chat into. You got the chat hole, right? And the basic model 40. This is
42. 9:5540. All right. Perfect. Then along comes the reasoning models 03 04 mini 01 all those
43. 10:10right where the the toilet paper comes in. So it fits perfect. It's a perfect metaphor. Okay. So, okay. So, before the toilet paper,
44. 10:25you ask chat JPT for something and it goes into the knowledge base into the into all of humanity, humanity's knowledge and it gets you an
45. 10:36answer and it gives it to you. And what happens with the reasoning models is you ask it a question and it goes into all
46. 10:43the knowledge of humanity and then it just keeps asking itself questions like like it's just like searching the
47. 10:51internet going back into the humanity just just pulling out lots and lots of questions and and then it gives you an answer. Okay.
48. 11:01So, this is 03. It's solid. I'm liking this. This is working. I'm so dead. No, but it gets better. Okay, this is a box of tools.
49. 11:21Okay. So, so the new the new 03, the the the the big MacDaddy 03 that they just launched, it can actually write its own
50. 11:33tools. So, let's say let's say it goes into all all of the knowledge of humanity and it pulls out some data and
51. 11:40it it creates a little plan for itself, but then it says, "Hey, I actually need to analyze some of this data and come up
52. 11:47with a trend." But, but I can't do that by just, you know, talking to it. I actually need a tool that's going to do
53. 11:53trend analysis. It can go into its little box of tricks and write Python, write a tool that will do that analysis.
54. 12:00And so, it sort of combines all these things together. tools, these things. Okay, but 03 is really just using
55. 12:09Python. It's kind of a single tool. Okay, now so that's that's that's gotten us to where we are today with chat
56. 12:17GPT. Manis and GenSpark and and things like it, they're this, but they have like 80 different tools, right? I think Gen
57. 12:28Spark has 30 or 40. Manis has like 80. So, Manis can do things like instead of going and searching the web, it can
58. 12:36actually spin up 10 different web browsers and have them all doing things independently. So, so Manis and GenSpark
59. 12:44have a really good coordinator role that's coordinating all these different tools that it's using and all the
60. 12:50different stuff out there. So, you give it a goal, it comes up with a plan, and then it figures out, okay, of all the
61. 12:56tools I have access to, what am I about to go turn loose? Um, so hope that helps. Um um if you've heard of MCP, which is
62. 13:09model context protocol from from Anthropic, what that is is that allows developers to take applications like say
63. 13:18Blender or uh Excel or [ __ ] whatever you know some your Macintosh operating system and they can create sort of a a
64. 13:30series of instructions that allow a large language model to talk to that application. So there's been some some examples on [Music]
65. 13:43um on Twitter where you know they just tell Blender to like make me a a 3D model of a little town with lots of
66. 13:51buildings in it and you just watch Blender like making [ __ ] right? Because it's got model context protocol. There's
67. 13:59going to be thousands of those things. And and what Anthropic just announced last week is that you can there's now
68. 14:07like effectively a search engine for all the MCP servers. So what we're going to have like in in very short order. Again, I
69. 14:16don't think agents are there right now, but we'll we'll go play with Manis and GenSpark and I'll show you how they
70. 14:21work. But where we're headed is you're going to be able to put a request into the chat hole and then the thing will start
71. 14:31reasoning with itself and then it'll basically go out and look and say what are all the different possible
72. 14:37applications I could use to help me in pursuit of this goal. And so it might just go out and do a bunch of [ __ ] right? Like maybe you
73. 14:47want to start an import export business and it's finding, you know, all the laws in the different countries to get the
74. 14:54thing from where you want it to where you are and it's it's figuring out all the tariffs and it's building
75. 15:00calculators and it's using websites that specialize that. Like that's that's the kind of thing where it's headed. So, so just just
76. 15:10think your simple request is getting increasingly more complicated on for what the systems do with it. Now, what
77. 15:18do you do with that power? I don't know. Um, last night I shared an article that I wrote. If you go to my ex channel um
78. 15:28Kyle Shannon and click on articles, there's a there's an article in there about how to use 03 as a non- STEM
79. 15:38person if if if you don't have physics and math problems to solve how do how do you use a reasoning model? So I wrote an
80. 15:46article on that. So that that might be something worth looking at. Um we'll come back to this thing. Let's Let's just go
81. 16:01to Let's go to Let's go to Gen Spark because I think Gen Gen Spark is like a friendlier, simpler Manis. All right. So,
82. 16:15GenSpark.ai. And I if you're if you're trying to get to Manis, it's mus dot Wait for it. I am manis.mim and genenspark.ai.
83. 16:32Okay, so gen Gen Spark has this thing called the super agent. Um, it can also generate video. It can generate images. Um, it can do
84. 16:43slides. Um, why don't we do slides? Because slides are kind of interesting. Why don't we do um find all of the tools
85. 17:02that are like manisim and genspark and then I'll I'll let it know it's it's it's that's it.
86. 17:18Jen Spark, that's you. Oops. Find all the tools that are like
87. 17:32Manis and Gen Spark that are agentic and powerful and can use tools
88. 17:49and create a presentation that describes what they are, the pros and cons of the
89. 18:08main tools. and at least three slides of use
90. 18:23cases beyond research and math. All right. So you can give these things relatively complicated
91. 18:38uh instructions and then they just go and douche [ __ ] So So we're in slides um and I'll just hit return. So we've got our
92. 18:50little area over here where this is where our chat is and then over here on the right we'll get the So so it says using tool search
93. 19:02manisim Agentic AI tools using tools search using tool read so it's got a reader tool and and you can view any one of
94. 19:13these right now it's searching again claude agentic AI tools And then here's a bunch of [ __ ] it's finding, right? Pretty slick. Search
95. 19:30thinking. Okay, now I have enough information to create a comprehensive presentation using tool. I don't know what tool it is. Oh,
96. 19:39presentation. New eight slides. So, it's got a presentation tool, right? And then if you notice over here thinking I'm creating a title
97. 19:53slide I'll use tail tailwind CSS for styling now it's writing the code right so off off go do
98. 20:09research read some [ __ ] comprehend some [ __ ] figure out of all the [ __ ] you just learned, what eight slides you need to
99. 20:19write. Go think about slide one now. Code slide one. And then you'll see here in a second, we'll see slide one. And
100. 20:27then it's going to do this for all all eight slides. So, and like, you know, that's pretty slick. Leading agentic tools, Manis,
101. 20:36Gen, Spark, Claude, and Perplexity. All right. I, you know, I I wish it would have given me one I hadn't heard of, but
102. 20:43whatever. I've heard of all those and I I would say, you know, perplexity. Yeah, Perplexity is in the neighborhood of one of these. Clawed
103. 20:53with MCP certainly is, but like it gave us a nice little design there, right? It knew which word to aentic AI to
104. 21:06highlight. Perplexity has agentic tools. Well, kind of. If you've used um in perplexity, have have you generated a
105. 21:18page in perplexity? While this is doing this, well, let me jump over to perplexity and show you something that's in the neighborhood of
106. 21:25agentic. And again, everyone right now, everyone, [ __ ] everyone is calling everything an agent, right? Because Sam Alman a year
107. 21:40ago said, you know, agents are coming when whatever the [ __ ] it was. So now everyone's calling every automation that they create an
108. 21:49agent. You know, think think of an agent is an agent is doing something on your behalf without you needing to guide
109. 21:57it. Um so what perplexity does that's in the neighborhood, it's not great, but it's in the neighborhood is
110. 22:08Where is it? Uh, oh [ __ ] they've moved things. Um, okay. Let me just go I'll go I'll go research something. Uh, find me the top agentic
111. 22:24tools as of today. [Music] So, it's searching. So, like one of the
112. 22:39things that Perplexity does is it searches a bunch of things. It gives you the sources. Um, so this is kind of doing
113. 22:48what that does, but you know, it's it's not really. So, what did I come up with? Relevance AI co-pilot studio Watson X
114. 22:58orchestrator. Um, let's see. These all suck. So, let's go. No things that are more like um Claude MCP
115. 23:16um Manis.Mim and Genspark dot what was it? AI
116. 23:35Add to space. They've gotten rid of something here. Discover home plus spaces. Did they get rid of pages on
117. 23:54perplexity? Export answer. Share. Rewrite Pro Search Copy. Am I losing my mind? This is a
118. 24:09master class with a laugh. Thanks, Avi. Oh, my boxes. That was a master class. Um, I thought Mr. Smith was an
119. 24:23agent. Um, which model did you use to help script this new analogy? Hey, that was just that was just off. I'm like
120. 24:31Harry Mack with the uh with the analogies. Um, where the [ __ ] are pages? Oh,
121. 24:46library threads page. Jesus, they buried the [ __ ] out of it.
122. 25:00Okay, up at the top in Perplexity. Let me make this a little bigger. In Perplexity, up at the top, there's this thing called
123. 25:14library. And if you click on library, it'll take you to threads, which I guess are searches I've done. And then there's a plus
124. 25:25button which if you click it, you can create a thread or a page. I don't know why they buried this so much. Maybe it's costing them
125. 25:37money. So you choose the audience. Do you want beginners or experts? Let's go with beginners. And we'll say create a page
126. 25:46about Why can I not type in here? Oh, what's your page about? Um, Agentic Tools. I I'm so pissed at these
127. 26:02companies. They [ __ ] change the interface every [ __ ] week. Like, how are we supposed to use your dumb [ __ ]
128. 26:10tool? Hey, Gentech tools like Manis.im I am and Jen Spark um Jen Spark um find me others and
129. 26:28explain what the hell I am supposed to do with them. Okay. Boom. So, where this thing gets kind of
130. 26:44agentic is it's off doing searches right now. And what it's got is it's got kind of a layout tool built into it. So
131. 26:53rather than just doing a search with sources, what it's actually doing now is finding images, it's doing sort of
132. 27:01magazine layout. So what it's creating is like a little mini magazine article about AI agents uses and alternatives
133. 27:12right and if you scroll down so similar AI's similar AI's dynamic learning capabilities Rasa for contextual conversation design think of
134. 27:27Rasa wait where did Rasa go as a Lego set for building smart chat bots unlike simple chat bots that follows. Okay, so
135. 27:36it's finding us all this [ __ ] Cool. Are we done here? There's genark my experience. So I think with any of
136. 27:52these copy it, link it. They've completely changed this. Continue editing.
137. 28:09Okay. Ah. Oh, you Jesus Christ. Now you have to say continue editing. Used to just be able to edit this stuff. Okay.
138. 28:18So So see it picked this image for Gen Spark and it's the the media here is not Gen Spark. But let's so we can change out the
139. 28:29image. Super agent showdown. Best free alternatives. Now there's that's a cool thing. Gen Spark. And then here I can
140. 28:40add media for similar. So if I say add media, that'll go find media. And so it threw in that chart. I don't know what that chart is, but
141. 28:48whatever. We'll throw in that chart. And then let's see. This has got um bullet points, but let's say we want
142. 29:00that to be a table. So I think if I go edit here. Yeah, I can just turn that into a table and then hit that. And now it's
143. 29:10going to rewrite whatever I just did as a table. Yeah, here we go. Boom. Now you got a table there. So, how this is not agentic is like I'm
144. 29:25having to do a lot of this, right? Um, but the fact that it went out and it found three or four things and it wrote
145. 29:33articles and it knew how to do it and it did did all that without me, that makes it it's got elements of of Agentic,
146. 29:40right? Think of Agentic as it's just off doing its thing. Haven't made a page in a while, but I still get them dropped to
147. 29:46me. Yeah, never had a subscription, so I never used pages. I don't think you need a subscription for it. I don't think I
148. 29:53have a subscription anymore. I had the pro account for a while. All right, let's go look at what GenSpark did. Oh,
149. 29:59it's still [ __ ] working on our presentation. Let me see. View and export. Play slides.
150. 30:16Nice. All right. What are agentic AI tools? Can you see this? Oh, no, you can't because I went full screen and chat GPT decided
151. 30:28to or no, Chrome. Oh, no. OSX decided to reorient the windows in some random [ __ ] way. Um, traditional AI versus agentic AI.
152. 30:41Traditional AI assistants respond directly to prompts, are limited to pre-trained knowledge, require explicit
153. 30:48instructions, have no ability to use external tools. Agentic AI tools, autonomous decisionmaking, that's
154. 30:55probably the biggest one. The the toilet paper role, that's it. Sort of planning and thinking about and and changing its
155. 31:03plan based on what it learns. External tool usage, multi-step planning, self-improvement over time. Yep. Good. I like
156. 31:17it. Leading tools. Manis. Okay. So, here we go. Autonomous execution, multimodal tool integration, adaptive learning.
157. 31:26Manis, state-of-the-art performance in Gaia benchmark, greater than 60%. Genspark, agentic agent with dedicated
158. 31:34AI agents for search, deep research, and various automated tasks. Multi- aent teams nice deep research content generation 80 plus tool
159. 31:44integrations in GenSpark Claude AI assistant from Anthropic with agentic capabilities for planning. I I I if they're not including
160. 31:56MCP stuff here I that's that's a stretch and perplexity is a stretch frankly both of their their other two they're stretches for
161. 32:05me. Here's pros and cons for both use cases. Business process automation content creation and media production personal digital
162. 32:22assistant. So there you have it. That's cool. All right. I I How do I get out of here? Did it Okay.
163. 32:36Is it done? Yeah, it's done. So anyway, oh god damn that thing. Um, so yeah, here's all the slides it made and it did all that, you know,
164. 32:51these are not it's not using what's amazing about this. It's not using a slide program, right? Like there's slide
165. 33:00programs out there like I forget what they're called. Whatever doesn't matter. This is actually writing the code and
166. 33:08the CSS to do the design of these slides. Can you say move? Can you say more about personal digital assistant?
167. 33:21Sure. Yes, I can read the slide. Hang on. Um, view and export. Your Tik Tok cam is off left. How's
168. 33:36that? Better. Hang on. Let me let me go. Hang on, Cam. Yes, I will read this slide for you. I know. I didn't really read it.
169. 33:47Okay. Personal digital assistant. How agentic tools can enhance daily productivity by autonomously? Well, you
170. 33:53know what I'll do? I'll I'll also export Oh my god. I'll also export this um and and put it in the uh in the salon, but
171. 34:03let me let me just read through it here and see if it's worth exporting. Um, Agentic tools enhance daily productivity
172. 34:10by autonomously managing personal tasks, calendar, and scheduling. Autonomous meeting scheduling across time zones.
173. 34:18It's kind of cool. Dynamic priority based rescheduling. Intelligent buffer time management. I don't know what that is.
174. 34:27Oh, this is cool. It's also saying which tools would do this. Well, Manis and Gen Spark would do that one. Research and
175. 34:33learning, Perplexity and Claude. Travel planning, Manis and Gen Spark. Financial management, Manis and
176. 34:43Perplexity. Time saved weekly, 15 plus hours. Decision quality plus 42%. Stress reduction minus 38%. Not if you got to go learn all these
177. 34:55[ __ ] tools. Gamma was the other one. That's right. Um, but let me let me um let me export
178. 35:10this. We'll export it as a [Music] PDF. Uhoh. What did I do?
179. 35:33And a gentic tools comparison 2025 desktop. All right. So I am going to the AI salon. If you don't know what the AI
180. 35:42salon is, it is swell. We had a lovely meeting tonight. We had uh Telicia White from lovable.dev dev uh which is one of
181. 35:51them their vibe coding tools. Uh she came in and demoed Lovable for us and was just awesome. It was a really good
182. 35:58meeting tonight. I can't get one to work on clearing out my Gmail inbox. Yeah. Yeah. Like I said, listen all everybody
183. 36:10don't think you're missing the agent bus. You haven't missed the bus. The bus hasn't been assembled yet.
184. 36:19Right now, we've got like an agentic unicycle and and it's only got one pedal, so it's really [ __ ] hard to
185. 36:33ride. Um, where do you want me to put this? You want me to put this in the regular channel? Can I upload a PDF to this
186. 36:44thing? I bet I can't. That's gonna piss me off. [Music] Uh the pro there was a problem. There's
187. 36:58always a [ __ ] problem. Shut up. Just all of you leave me alone. Leave me alone. Export. Export. Quartz filter. Reduce
188. 37:15file size. Save. Replace. Let's make things smaller. This shouldn't be that big because it's all [ __ ] code. But let's let's go back
189. 37:30here. Let's try this again. Nope. Anybody? Ber, anybody know how I
190. 37:45can upload a PDF to uh Mighty Networks? No, I'm exhausted. I'm exhausted, people. Tired
191. 38:01Tuesday. I just want a real assistant to handle all of my 111,239 unreads. Yeah, exactly. Save them as JPEGs. Yeah, I know. I just
192. 38:13don't want to upload [ __ ] eight different JPEGs cuz I'm lazy. Um, nice artwork, dude. What the [ __ ]
193. 38:28Are you okay? That's Amelio's wife's handiwork right there. Oh my god. Save it in a Google Drive and
194. 38:44paste the link. Okay. Thank you, Vicki. Once again, serve serving as my as my brain. Okay. So, here's what I'll do.
195. 38:55I'll go to I'll go to here. I'll go to here. We'll take a screenshot of that bad boy. [Music]
196. 39:11Uh, we'll save that to the desktop. Boom. Then we'll go to drive and I'll put it in just my drive loose because I
197. 39:25know how to live. I know how to do file management. You know how you do file management? You save everything in the root
198. 39:36folder and make sure that it's not dated or has any kind of name that would make sense. Then you got to find it with your shitty naming
199. 39:50convention. By the way, I do consulting. You can hire me for file management strategies. I'm reasonable. Um,
200. 40:02share share with anyone with a link. All of you can have this. Copy link. And then we're going to go back to uh, what's it called? AI
201. 40:15salon. We're going to go get our picture that we took a picture of. Come on. Where is it? Where's my screenshot? Uh, I have a second
202. 40:24desktop. Damn it. Screenshot. I listen all of you people that know how to use computers. I don't want to hear it. I don't
203. 40:33care. Is that it? That's it. Uhuh. Hey, Marge. Hey, Marge. Marge. Yeah. Hey, listen. I'm Oh, I know you're watching the wheel,
204. 40:47hun. Yeah, I just figured out myself. Marge is not as patient as she used to be. I think it's her
205. 41:07bunions. All right. So, this is uh what is this? This is the here is the genspark.ai AI um aentic tool compare is son
206. 41:26presentation I made. People are going to read this and be like what the hell is he talking about? But there you go. We're not going
207. 41:38to notify all of the world that we put up a PDF. All right. Got it. You're a complicated man, Kyle. By the way, lovable oneshot my app
208. 41:54using 03 instructions based on my half-ass prompt. So, archetype, that's that's where we're headed. Like, this is
209. 42:02the thing. Okay, so I'm I'm going to we're going to we're going to jump now. I So, Cam, I hope that I hope the masterful um explanation of
210. 42:15large language models and tools and toilet paper. I hope that cleared things up. Um, but I think what what archetype just
211. 42:28said right there is is absolutely right. And uh we so we we read this on the salon tonight. So I know some of you
212. 42:38have probably heard this, but I think this is really important and I'm I'm probably going to talk about this for a while.
213. 42:45Um so this is from Mika Kaufman. He's the CEO of Fiverr. And if you don't know what Fiverr is, it's like a freelance marketplace, right?
214. 42:57Um, I'm pretty sure it was Fiverr. Yeah, I'm pretty sure it was Fiverr that just added um AI tools for freelancers to be able to do [ __ ]
215. 43:10faster, right? So, they could actually charge five bucks and, you know, make money. No one charges five bucks on
216. 43:16Fiverr anymore, but whatever. It's like Dollar General stores. Nothing's a dollar in a Dollar General store anymore. So,
217. 43:23um, Manis did it, too, but it's uglier. Yeah. No, Lovable is really good design-wise. So, anyway, so the CEO of
218. 43:31Fiverr sent this email to his staff and then he posted it to X saying, "Hey, this is probably going to come out
219. 43:37anyway, so you might as well hear it from me. I've always believed in radical cander and despise those who sugarcoat
220. 43:44reality to avoid stating the unpleasant truth. The very basis for radical cander is care. You care enough about your
221. 43:52friends and colleagues to tell them the truth because you want them to be able to understand it, grow, and
222. 43:58succeed. So, here's the unpleasant truth. AI is coming for your jobs. Heck, it's coming for my job, too. This is a
223. 44:08wakeup call. It does not matter if you're a programmer, designer, project product manager, data scientist, lawyer,
224. 44:17customer support rep, salesperson, or finance person. AI is coming for you. You must understand that was once
225. 44:24considered easy tasks will no longer exist. What was once considered hard tasks will be the new easy, and what was
226. 44:33once considered impossible tasks will be the new hard. If you do not become an exceptional talent at what you do, a
227. 44:44master, you will need to face Wait, you you will face the need for a career change in a matter of months, right? This is not a five-year
228. 44:55thing. This is like before the end of the year, you got to get your [ __ ] together and [ __ ] level up, right? He's saying this to his
229. 45:07staff. Um, if you do not become an ex exceptional talent at what you do, a master, you will face the need for for
230. 45:16need for a career change within a matter of months. I'm not trying to scare you. I'm not talking about your job at
231. 45:22Fiverr. I'm talking about your ability to stay in your profession in the industry. He's like, I don't give a [ __ ]
232. 45:29if you work here or there. If you don't do this, you're not going to make it anywhere. Um, are we all doomed? Not all of us,
233. 45:40but those who will not wake up and understand the new reality fast are unfortunately doomed. What can we do?
234. 45:49First of all, take a moment and let it sink in. Drink a glass of water. Scream hard in front of the mirror if it helps you. Now,
235. 45:57relax. This is it's a well-written letter. Panic hasn't solved problems for anyone. Let's talk about what would help
236. 46:04you become an exceptional talent in your field. The fact that you're here tonight, it is [ __ ] 9:40 in Denver.
237. 46:11So, if you're on the East Coast, it's [ __ ] 11:40, right? Even on the West Coast, it's almost 9:00 at night. There
238. 46:19are 52 people in the Tik Tok, 48 people on YouTube and X and LinkedIn. I'm hopeful for you. you're here. I'm here. Like, why am I
239. 46:34here? Cuz this has been coming. But, you know, we now got someone articulating it quite well. Um, okay. So, he says,
240. 46:44"Let's talk about what would help you become an exceptional talent in your field. One, study, research, and master
241. 46:52the latest AI solutions in your field." Right? Listen to people that use boxes and toilet paper to explain how AI
242. 47:03works. Okay, this one is definitely this channel. Try multiple solutions and figure out what gives you
243. 47:10superpowers, right? What am I doing in here night after night? Just [ __ ] around with stuff, trying different
244. 47:15things. You know, we did it tonight. We played with uh Gen Spark and then we went to Perplexity and Perplexity's been
245. 47:22kind of downgraded and is weird now. Jen Spark kicked ass. Okay, so we know that. Um, archetypal architect said that
246. 47:30Lovable did better than Manis did because it was prettier. Okay, good. So, he's understanding, you know, where the
247. 47:37superpowers are. By superpowers, I mean the ability to generate more outcomes per unit of time with better quality per delivery. Listen to
248. 47:50this. Superpowers. the ability to generate more outcomes per unit of time with better quality. So if you're used to charging a
249. 48:04week's worth of your time to deliver research, you now have to do that in a day or two days and it has to be better
250. 48:18than what you did when you took a week. Right? That's super power there. There used to be this old adage in
251. 48:23advertising, good, fast, and cheap. You can have any two of the three, right? You can have it good and fast, but it ain't going to be cheap.
252. 48:31You can have it fast and cheap, but it ain't going to be good, right? You not anymore. Not with AI. Good, fast, and cheap is now the
253. 48:43expectation or will be. Um, programmers code customer support tickets. You know, he he's given examples here. Then he says, and and
254. 48:57again, that's what this this room is for is for number two. Find the most knowledgeable people on our team who can
255. 49:04help you become more familiar with the latest and greatest in AI. This isn't just about me teaching what I know. This
256. 49:10is about all of you in here as you're commenting to one another, as I'm talking about stuff and you're going,
257. 49:16"Oh, I did that over here. I did that over here." The power of that knowledge sharing is [ __ ] insane. Basically, what he's saying
258. 49:26is learn everything you can on your own and then get your ass in community. Connect with other people that are
259. 49:32curious about this [ __ ] and hungry about this [ __ ] Time is the most valuable asset we have. If you're working like
260. 49:39it's 2024, you're doing it wrong. you are expected and needed to do more faster and more efficiently now. Um,
261. 49:50become a prompt engineer. Google is dead. LLM and Gen AI are the new basics, right? It's not a big deal that
262. 49:57you can use chat GPT. That's now expected. Um, and if you're not using them as experts, your value will
263. 50:05decrease before you know what hit you. Five, get involved in making the organization more efficient using AI
264. 50:11tools and technologies. It does not make sense to hire more people before we learn how to do more with what we have.
265. 50:18It does not make sense to hire more people until we understand how to do more with what we have. He didn't quite say there's a
266. 50:27hiring freeze, but it doesn't sound like he's going to be hiring. Understand the company strategy well and contribute to helping it
267. 50:34achieve goals. Don't wait to be invited to a meeting where we ask you to participate uh each participant for
268. 50:41ideas. There will be no such meeting. Instead, pitch your ideas proactively. Stop waiting for the world or the place of work to hand you
269. 50:51opportunities and to learn and grow. Create those opportunities yourself. I vow to help anyone who wants to help
270. 50:58themselves. Right? So, it's a bit it's a bit like, you know, go suck it up, buttercup. But suck it up, buttercup,
271. 51:05right? That's where we are. If you don't like what I wrote, if you think I'm full of [ __ ] or just an [ __ ] who's trying
272. 51:11to scare you, be my guest and disregard this message, I love all of you and wish you nothing but good things, but I
273. 51:18honestly don't think that a promising professional future awaits you if you disregard reality. So, um, and then he wraps it
274. 51:28up. Brandon made a good point there. There were two two other CEO letters that preceded this one. I have a feeling
275. 51:36we're about to see a rash of these. This was the one that was the most directly worded where it was just like [ __ ]
276. 51:44you're going to have to [ __ ] deal. So, anyway, Google is not dead while they are rising from the dead. Well, I
277. 51:51would agree with that. I I I think I think Google's AI is good, but I saw a thing today. It was some SEO
278. 52:01dude and he said he said SEO is dying, right? He said they they were having traffic go from like 10,000 somethings a
279. 52:11day to 3,000 like like you know twothird drop off. Um so becau because people are not using Google as much. They're using chat GPT. So that
280. 52:23search traffic is is diminishing now. Might there be LLM optimization? Yeah. So, so yeah, there there How do
281. 52:36you get How do you show up in a chat GPT search when someone's asking about Ecoin toothbrush or toothpaste? I don't
282. 52:45know. Someone's gonna have to figure that [ __ ] out, too. So any anyone that says Google is dead is being a bit hyperbolic, but SEO's
283. 52:55toast. I could see Google's agent space making some waves. They showed they're very serious with 2.5. Yeah, it looks
284. 53:02like the the a new version of 2.5 just came out that's supposed to be really good at coding. Google's problem is not their
285. 53:10technology. Google's problem is they fired their [ __ ] marketing layer. They fired all the managers. They fired
286. 53:18all the all the non-engineers. They're like, "This is just about technology." And so now they've got a a
287. 53:28horrific product strategy. Well, they don't have one. They They're literally just launching features into one of about a dozen
288. 53:37URLs. No one can find them. Like, the only ones worse than Google right now are Microsoft. And the only ones worse
289. 53:43than Microsoft are Apple because they haven't done [ __ ] anything. Oh my god, it's crazy. It's crazy.
290. 53:59Um, absolutely bonkers. Agent space is going to be huge and agent to agent is going to be huge. Yeah, don't don't worry that you're not
291. 54:10caught up on agents. Probably I'm going to guess sometime after GPT5. So, I'm going to guess GPT5 comes out in June or maybe
292. 54:24July. And then I'm thinking probably August, September, October is when we start seeing real autonomous agent systems show
293. 54:38up, right? Because MCP came out what, three months ago, two months ago. So developers are developing all their
294. 54:48servers and the infrastructure is now catching up and then you know OpenAI is going to have their version of it.
295. 54:53They're probably going to support MCP. They're going to have their own agentic framework. Like all of this stuff is
296. 55:01coming. Here's what to pay attention to because this doesn't really exist right now. start paying attention for companies
297. 55:12like OpenAI or Anthropic or or any of these talking about the um the coordination layer, right? Where they're going to
298. 55:22have specialized LLMs that their only job is to take your they're going to they're going to act like project
299. 55:29managers, right? There's going to be a project manager AI that you talk to and it's going to go manage all these different activities.
300. 55:39Um, I think we'll see some of the frontier companies starting to talk about that layer and GPT5 could could
301. 55:46have that conversation in it, but that that's what to pay attention to. When that starts showing up, then then agents
302. 55:52are probably on their way. News brief. Google miner finds an app nobody knew existed. Exactly. Story at 10. Yeah, exactly.
303. 56:06Why is Apple so unambitious with AI? I don't [ __ ] know. Well, I I do know. I do know. Steve Jobs, you know, on his way out, he had
304. 56:20he had two options and he chose Tim Cook. He chose a CFO. The reason you choose a CFO is because
305. 56:35you don't want your company to go out of business. The reason you would choose another innovator is you want the company to keep
306. 56:44innovating. He chose the C CFO, right? He chose Tim Cook. So naturally, as the innovation pipeline that was in Steve Jobs head dries
307. 57:00up, if the priority of the company is sell more [ __ ] and keep the margins, you know, there's not a commitment to innovation. I don't
308. 57:10[ __ ] know. I don't get it. I don't get it. It's It's like one of the saddest things in in technology history.
309. 57:18out. It's possible that Apple, like Apple's never been a leader. They they always sort of sit back, watch everyone
310. 57:25sort of [ __ ] around, and then they're like, "Oh, you want to see how to really do it? Here you go." They've had that
311. 57:33opportunity now for a year and a half, and it's just their their AI [ __ ] sucks. It's it's like unusable. It's it's
312. 57:42depressing. It's depressingly bad. So, I don't know. I don't know what's going on over there. Does anyone else fall for the April
313. 57:53Fool's saying Apple acquired sesame? No, I didn't see that. Apple's historically been the perfector. Yeah, but but like
314. 58:05normally normally you've got some sense that they're perfecting something, right? That they've got some capability.
315. 58:14They're like, "Yeah, we're not going to talk about that." But you know, something's cooking now. I just get the
316. 58:20sense they're just in I don't even think they're in [ __ ] panic mode. It's bizarre. It's bizarre. I think they
317. 58:27fired someone from something AI and they they they collapsed the the goo the goggle team, the the [ __ ] VR goggle
318. 58:37team, and put them on I I don't know what's going on. I don't know. I don't know. I don't know. can't talk about it.
319. 58:44Makes me sad. Like I'm a I'm a diehard Apple fanboy and I just, you know, about four or five years ago, their product
320. 58:53quality just [ __ ] went off a cliff. Maybe even longer than that. Seven, eight years ago, and now they're like, they're
321. 59:01missing the AI boat. They've got more [ __ ] cash. They could be buying up all these companies that are doing good work.
322. 59:10They're not [ __ ] doing a thing. They're not doing a thing. I feel like Apple's given up. I I feel like that's
323. 59:17that's the thing that makes it depressing. It's like where's the [ __ ] panic? Where's the urgency? They're like, "Oh, you can make
324. 59:25emojis of yourself now in in at the Eiffel Tower. Really? That's what we're going with? [ __ ] fancy mimojis. [Laughter]
325. 59:49Maybe Ethel knows that the aliens are going to land soon, so there's no point. Yeah, I like that would be that. You
326. 59:56know what, Shane? At least that would be a [ __ ] explanation. I I I honest to God, I don't get it. I don't get
327. 1:00:04it. They should bin Vision Pro and compete with Neurolink. They should do [ __ ] something. do their car, compete with Tesla, do
328. 1:00:15something. I I I just it just makes me sad. And it's Listen, I I have experienced like a teeny tiny micro version of this. When I when
329. 1:00:28I started agency.com in the 90s, you know, we started as a small scrappy digital agency and then we grew really
330. 1:00:36fast and we grew through acquisition. And I ended up leaving the company because it had become the company we used to make fun
331. 1:00:46of, right? The company I founded ended up being one of those companies that couldn't make decisions. It was always
332. 1:00:54political. People were [ __ ] doing shenanigans. Who does who says that? M. Baabab says, "Hi, Dad." Hello. I don't know what that
333. 1:01:11means. Um, Vision Pro had such promise and it just disappeared. I listen, I I don't think they're going to kill Vision
334. 1:01:19Pro. I I I think there's got to be a version two and then probably a version three. Version three will be in a form
335. 1:01:26factor that most people can use and can afford. I mean, this is how they've always done it. So, I I don't think Vision Pro goes away. Um,
336. 1:01:37but it wasn't the runaway hit they thought it would be. Like, it's got some seriously impressive technology in it.
337. 1:01:44But what the [ __ ] with the AI? Anyway, all right, people. Um, my voice is getting crispy. It's been a long one.
338. 1:01:52I'm curious to see what they say about AI at Worldwide Developer Conference. I'm assuming they realize how far behind they
339. 1:01:59are. Why are you lying? L I G H I N G. I don't know what that means. What am I lying about? Lighting. What am I
340. 1:02:14lighting? Um, Apple deals with teachers. Teachers hate AI. Maybe that's why they don't push it. Teachers hating
341. 1:02:22AI. Okay, let's let's run this logic train, shall we?
342. 1:02:37teachers. Their job is to impart knowledge and teach kids how to learn. And along comes the most
343. 1:02:51profound knowledge access tool in the history of humanity. And schools are are saying kids are going to use it to cheat. Let's not let
344. 1:03:06them use it. [ __ ] really really really that that like the only thing less excusable than Apple not innovating on AI is the
345. 1:03:22education sector [ __ ] missing this boat. It is shocking to me. It's [ __ ] shocking. I mean, education's been
346. 1:03:32broken for [ __ ] decades because we're still teaching based on the the the [ __ ] curriculum from the industrial
347. 1:03:45revolution, right? We haven't even updated to like the internet age, much less. We now have knowledge on demand at at our fingertips.
348. 1:03:57students are going to enter a world that does not resemble the world we have right now. It's It's not even going to resemble
349. 1:04:04it. And we're ignoring AI. It's it's it's it's [ __ ] shocking. It's it's it's shockingly destructive. Apparently Claude had a
350. 1:04:14mass massive context window in preferences. Oh, has a massive context window in preferences. Let's go look at
351. 1:04:21that. I don't know what massive is. They've always had like 200,000 tokens. Is it bigger than that? Claude.ai. [Music]
352. 1:04:43Oops. Preferences settings. Let's see. Artifacts on feature preview on appearance wouldn't be there.
353. 1:05:00Account cla settings. Uhuh. I don't see it.
354. 1:05:18They're going to find out real quick that the kids can dance circles around them at home. Yeah. Well, but the thing
355. 1:05:25that's [ __ ] up is you've got like what what's here here's the result here. Here's what I'm going to what I'm seeing
356. 1:05:34is you've got Gen Xers kicking ass at AI. You've got millennials like, "Yeah, whatever, Dad." Millennials are just
357. 1:05:43kind of like pissed off that they're not working right then you've got Gen Z who are have been told that AI is bad and so
358. 1:05:57they're afraid of it and then I think whatever whatever Gen next or whatever the [ __ ] next one is below Z I think
359. 1:06:04they're going to be fine because I think I think kids in grade school right now like the education system's going to
360. 1:06:11[ __ ] work something out before they get to high school I think they're going to be fine but I think this Gen Z where
361. 1:06:16they're like afraid of it, you know, combined with like, you know, who their who they look up to, the millennials and
362. 1:06:23who their parents are. I don't know, man. I It's just It just doesn't seem good. It just seems like there's a big
363. 1:06:29[ __ ] black hole around those those two categories of people. Alpha then beta. Yeah, beta is going to be absolutely fine, right? Beta
364. 1:06:40beta are going to um they're going to grow up native with AI. Like like Brandon's kid, he's four and he says, "Daddy, I want to listen to
365. 1:06:50the song I wrote that he wrote in Sunno." And Adnan's daughter who's seven that co-wrote a book with her
366. 1:06:59dad, you know, based on the the adult version of the book that her dad wrote with Brian Moahan. Those kids are going
367. 1:07:05to be fine. The cusp kids that are probably junior high right now, yeah, they'll be okay, but touchy. The
368. 1:07:15high school kids, people going into college right now, I think they're [ __ ] I think they're [ __ ] because I
369. 1:07:21I also like if you're graduating college in the next four years, where's the entry level? Like, where are the entry level
370. 1:07:31jobs? What are entry-level jobs? [ __ ] low-level tactical execution work, right? Oh, he's new. Just give him uh
371. 1:07:43tell him to do those TPS reports, right? Tell him to write the copy and paste this [ __ ] into the into the deck for the CEO. Those jobs are
372. 1:07:55gone. Where do those kids go? Especially if they've been told, "Don't [ __ ] use AI. It's evil. It's the world's greatest plagiarism m
373. 1:08:05machine. You're cheaters, you [ __ ] losers. Learn our book stuff. I don't know. They got to get self motivated. They They should read
374. 1:08:18the Fiverr CEO's letter. Anyway, beta are going to be cyborgs. I agree with that. Kim Katkin, I think there will be so many new fields
375. 1:08:27of research. Zool linguistics learning interspecies. Yeah. Listen again. Clean the windows. Exactly. Gen X will be
376. 1:08:37totally rad. Exactly. [ __ ] them, man. Like, listen. We have We have all of our lives, we just want to [ __ ] [ __ ] up,
377. 1:08:44right? Like AI is the ultimate [ __ ] [ __ ] up tool. It is. All right. Me got book. Yeah, exactly. Did you see that the
378. 1:08:55number of translator roles has increased since chat GPT is released? That's fascinating. There are going to be jobs we simply
379. 1:09:04can't imagine right now. And you know, and I I can promise you who gets those jobs are going to be the people that
380. 1:09:13radically get radical about learning AI. Get like radically curious, radically adventurous. Just [ __ ] go in, start a [ __ ] Tik Tok
381. 1:09:24channel. It's all about asking questions. It's absolutely about asking questions. Someone asked me the other day, their
382. 1:09:36kids in college I think might be studying computer science and they're like, you know, what else should they study? I'm
383. 1:09:41like, instead of studying, they should be learning AI. And I said, quite frankly, they should probably get a minor in
384. 1:09:49philosophy. And I actually I firmly believe that philosophy, psychology, drama, something about language, right? Language, sociology,
385. 1:10:04understanding the human condition because those people that understand people are going to be the ones that kick ass at
386. 1:10:13AI. Throw in sprinkle in a little technical skills with that. They're going to be amazing. All right. I'm amazed how heavily
387. 1:10:25represented Gen X is around this. I It's Well, Gen X So, a couple of things. Gen X, we got to experience the worldwide
388. 1:10:33web. Like, we were there before the worldwide web and we got to sort of live through it changing the world. So, we're we've got a visceral
389. 1:10:45understanding of what AI is about to do. like we we have no sense of what it is because it's just so much bigger than
390. 1:10:54the worldwide web was and so much faster. But I think that's why you're seeing Gen X do well in this because
391. 1:11:01it's like, "Oh [ __ ] this is a party. Let me jump on this bandwagon. Let me tell my boss to go screw himself. Go reinvent me a career."
392. 1:11:17Oh man, my degrees in technical and creative writing are now serving me very well. Yes, Joker. I'll get my for folding chair and popcorn for
393. 1:11:27that. Oh man. Round two fight. We are the Mad Max generation. All right, my voice is gone. I'm out of here, people. Um, happy Tuesday night.
394. 1:11:40Hope you had fun. Um, I'm I'm I'm really digging my new metaphor, the box, the toilet paper, and the uh the tools. It's
395. 1:11:50working for me. Did I mention the Oh, no. Um, so starting June 1st, so we've got the AI salon. So, if you go to the the AI
396. 1:12:02salon website, which is the salon.ai, and click join our community. Um, we've been around for two and a half years.
397. 1:12:10We're coming up on two and a half years. Um, tonight we announced we're launching what we're calling AI Salon Mastermind,
398. 1:12:18which is a subscriptionbased area in the salon for people who really want to go dig deeper with AI, build tighter
399. 1:12:26relationships, things like that. Um, and we're going to have like a ridiculously cheap price for all of 2025, starting
400. 1:12:33June 1st, which is 1995 a month. And then in 2026, it'll probably go up to 50 or 60 or 70 bucks. I don't know what
401. 1:12:40it'll be. Um, and you you come in as a a founding mastermind and and you get that that price for life if you come in in
402. 1:12:492025. So, if you go to the AI Salon website um under the community corner section, there's an FAQ about the AI salon
403. 1:13:00mastermind. You can go learn about what we're up to there. So, do that. And then, um, tomorrow, there's two things to pay attention to.
404. 1:13:08One is at 4 pm mountain time. Um the AI readiness project podcast with myself and Ann Murphy happens. Um we have an
405. 1:13:17amazing artist on tomorrow who you're going to [ __ ] love. So come to that. Um if you go to airespro.com you can see past episodes
406. 1:13:27and there's information about where you can watch it there. And then tomorrow at 5:30 Mountain is at 5:30 Mountain. Is that correct,
407. 1:13:37Brandon? Hold, please. Yes. Okay. 530 Mountain is the AI salon life hacks um club. And if you haven't been to one of
408. 1:13:49those, they're a blast. It's just like how to use AI to do cool [ __ ] in your [Laughter] life. Hale Lima is going to be fabulous.
409. 1:13:58Did you see the artwork that she made, Ann? She made this really cool piece of artwork she put on LinkedIn that
410. 1:14:03featured you and I and her with a cool groovy background. It's really cool. So, um, Ann Murphy's in the house. Ann Murphy,
411. 1:14:14rockstar. Hey, Ann, I have a very serious question for you. How many pairs of sunglasses do you own? It's got to be over a hundred. Am I
412. 1:14:28wrong? every every video you're in. I'm like, "Well, there's a banger set of glasses and then the next one. And then the next one." She's
413. 1:14:40laughing. She's got hats going. She got the hair going and just always some hip glasses. I lose them all. Oh, it's an add
414. 1:14:54thing. That's the best. Okay, cool. Um, all right everybody. Yeah, I'm going to get out of here. So, tomorrow night's Wednesday. I
415. 1:15:05don't think I have anything. Let me just look at my schedule quick. Do I have it up here? No. Should be 8 o'clock tomorrow.
416. 1:15:13Wednesday. Um, yeah, I'm pretty sure that's right. And, uh, think that's it. I think that's it. Hope you had
417. 1:15:24fun. Wear your party hats. Oh, yeah. I have a uh I achieve another decade [Laughter] tomorrow. I don't like it.
418. 1:15:41later.