Live From Coinbase CODE:NYC Hackathon

[00:00:01] Speaker A: Hey, Paul, how you doing? [00:00:02] Speaker B: Hey, Greg. Good, I can hear you. Perfect. Fine. Where? [00:00:06] Speaker A: Awesome. [00:00:06] Speaker B: Tell the people where you are. [00:00:08] Speaker A: Yeah, I'm here in, in Brooklyn, in New York at the Coinbase Code NYC hackathon. I'm representing one of my customers. Nice. Really smart guy named Chayton who has a. He has a company called Flora that does. It's a marketplace for MCP servers. So if you're doing an MCP server and you want to have it monetized, he has a protocol called Monetize MCP that you include in your MCP server. So for an autonomous AI agent, you need a marketplace in order to discover. Right. So once you build one, it's out there doing stuff, you want to charge for it. You can use this gateway. You also need a way to discover them and connect them. And so he has a marketplace for it. So yeah, we're here at this hackathon, there's a lot of other people doing MCP server, AI agent stuff. Coinbase has a new protocol, X402 that MCP or monetized MCP is compatible with. And so the big value prop is that if you use mcp, monetize msp, you can work with like multiple gateways and stuff. So. [00:01:24] Speaker B: Yeah. [00:01:25] Speaker A: There you go. Nerdy weekend. [00:01:28] Speaker B: That's awesome. [00:01:29] Speaker A: Yeah, no, it's, it's fun. It looks like they just got started. So there's some people here. Here, I'll show you. Nice. [00:01:39] Speaker B: That's. [00:01:40] Speaker A: So I believe the. The event is sold out. There's 200 people that registered. So for the hackathon. Yeah. And they will, they will code all through the off of the night and then Shade and I will judge the apps and Flora has a. As a category. [00:02:04] Speaker B: Okay. [00:02:06] Speaker A: Most, like most creative, that was something I suggested. I wanted like, hey, let's do something like super creative. I think hackathons are an awesome opportunity to like really push the boundaries of what you're building. And then the other category is most viable. Right? Like, hey, this one's gonna make, make the most money. [00:02:26] Speaker B: That's it. So you guys are judging this? There will be a winner. [00:02:32] Speaker A: Yeah, I think all the sponsors have bounties and categories, so it's really cool. As a developer here you can use multiple tools from the sponsors and put it into your app and you can win or enter and conceivably win in multiple categories. So as a entry, like you could, you could run the tables and win a bunch of money. If you over successful in a number of different categories, you can enter as many as you want? [00:03:01] Speaker B: That's awesome. [00:03:02] Speaker A: Yeah. [00:03:03] Speaker B: What. Have you already talked to some people? What are they building? [00:03:06] Speaker A: Yeah, it's a good question. I talked to one guy upstairs who had a. Like, a service that was a tutor. [00:03:13] Speaker B: Okay. [00:03:14] Speaker A: For kids. It was really interesting. So in New York City, I think they have. In other areas, too, if you have a child in public school, you can pull them out and put them into different charter schools that are, you know, hopefully a better educational environment than some of the public schools. And he had a tool that was to provide tutoring for kids that leave public school and go into the charter system. A lot of time, they're perhaps behind their peers in the charter school. And so he had a whole thing with, like, agents that would do tutoring, and the thing had, like, a payment system where if you're one of the kids and you do good on your homework, you can get crypto coins. So there's a monetary incentive. Yeah. [00:04:07] Speaker B: Yeah. That's awesome. It's a gamified study. [00:04:13] Speaker A: That was a cool, cool idea. Yeah. I think, like, there's been a lot of people who've talked about this, like, why not give economic incentives for kids who go to school and stuff? I'm a big advocate of it. I think, you know, I. I like capitalism. I think incentives are everything. I think it's actually a good idea. My suggestion was, like, you could also be penalized. So he didn't have that, but I was like, look, you should have the coins go in the kid's wallet, and if they don't do it, they miss a day. They have to pay a fine, actually, which I actually am an advocate for. Like, I think it should go in. In both directions. Like, you should be rewarded, but you should also be penalized, because that is how I think the real world works. [00:04:58] Speaker B: You get. You get slapped on the hand. You lose some coins or you lose some rewards. [00:05:04] Speaker A: Yeah, I think there's, like, people who've tried this with fitness apps. You get penalized. Like, you miss your workout and stuff. I'm. I'm a big. I'm a big believer in this. I think it's a. It's an interesting, interesting model. [00:05:17] Speaker B: I love it. Cool, man. What do you want? Do you want to talk about, like, GPT5? Any of the stuff from News this week? Lots of happens. [00:05:26] Speaker A: Yeah. Let's walk around a bit more here. Like, we're getting set up. [00:05:31] Speaker B: Yeah. [00:05:31] Speaker A: Dude. I have to say, like, it's really great to be back in New York. I love it here. And my wife was really excited. She's like. She's like, it's perfect for you. You're you. I'm getting a little old for this stuff, but. But she's like, you love New York. And I was big in hackathons from, like, 2010 to 2014 in the mobile era. Did a lot of them in Nice in the Bay Area and stuff. Yeah. So it's fun. It's fun to be back doing this. Doing this stuff. Yeah, we should definitely talk about GPT5. I'll jump onto the. Onto my PC and see if. See if we can do it from there. [00:06:13] Speaker B: Yeah, yeah, walk around. Let's see what people. People are doing. All right, how's this? Can't. Can't hear you. [00:06:43] Speaker A: It's. [00:07:08] Speaker B: It. Nothing yet. Eh. Oh, there you go. [00:07:57] Speaker A: Yeah. Okay, it's back on. All right. There you go. I knew he'd have a little bit of a challenge there, but that's okay. It's live fun. [00:08:04] Speaker B: This is perfect here, you know. [00:08:09] Speaker A: All right. [00:08:11] Speaker B: Oh, yeah, let's. Let's get. Let's let the people behind you settle in, and then maybe we can walk around. What do you think? [00:08:22] Speaker A: You want me to walk around more? I'm back on the PC, buddy. [00:08:25] Speaker B: Okay, let's just talk about GPT5. [00:08:30] Speaker A: Yeah, they're all hacking now. [00:08:32] Speaker B: They're all hacking away. Let them hack and then we'll walk around. [00:08:37] Speaker A: All right, so what are you. What are your thoughts on. [00:08:43] Speaker B: Yeah, so I. I think. I think it's cool. I think it's the first time that GPT is pushing towards the tool calling. If you read some of the press releases or if you read some of the. The blog content from people who got their hands on this model first, they are heavily pushing and explaining that this GPT has access to tools. So it's kind of like directly competing with mcp and you have to use it in a way that makes sense. [00:09:21] Speaker A: Yeah, it's obvious that, like, so. So they're gonna try to create their own ecosystem and everything. Like, if I was like, my own autonomous agent protocol came for all of that stuff. Because the tradition or history of Silicon Valley and technology is that the market leader always tries to, like, own as much of an ecosystem as possible. Right. Like, Facebook works that way, Apple works that way. Right. The ecosystems are fairly closed, and usually the nearest competitor tries to make an open systems like Android and Google. [00:09:57] Speaker B: Yeah, I. I'm like. So a couple things there is, yes, they're competing on mcp. It seems like they're competing not directly on coding. And also it's not. It's not AGI. It's nowhere near AGI. So. [00:10:16] Speaker A: No. Yeah. I mean, it's a lot of heat, like, because, I mean, I'm not surprised by it. But, you know, Almond's been out there perhaps overselling it or hyping it up, which I think is his job. [00:10:30] Speaker B: Of course. [00:10:30] Speaker A: Of course. But yeah, I mean, what I think the average person thinks AGI is. Is like, not going to be achieved in the next couple of years. But for me, that seems pretty obvious. And I don't know, like, I'm not just. I think. I think the tools are incredible. I think the new rods. Great. It was funny. Like, I went on Reddit and like the number one thread that I found when I did search was like, chat GPT5 is garbage. [00:10:54] Speaker B: Yeah. [00:10:54] Speaker A: Just all these angry comments. But, like, there wasn't a lot of things that were specific other than people were mad that, like, there's some limits now and you are forced to upgrade to the $200 a month. [00:11:07] Speaker B: Yeah. [00:11:08] Speaker A: Package. Which is like, expected. I mean, they're all obvious things if you've been around and that, like, there are a lot of people who said. What I have already believed about Chad was that the quality of the writing is not very good and I haven't done enough with it. I use it for some other things, but I've been disappointed with the writing capabilities of ChatGPT for a while now. You haven't really clear that. It's just. It's not a good writer. And I would like. I like to joke, like, totally tasteless. Like, I would even use it to like, do gas station advertisements. So I'm not surprised by. But I mean, my answer is that I think they've uncovered. Everyone's uncovered the AI tools. The big market. The big market is like Vibe coding software engineering, autonomous agents that do a lot of those types of things. And that's where I think they've forced. They made a conscious decision to go down the path of like, building a tool that does that. And I think they're trying to replace Google Search. I think there's like two use cases that are. [00:12:12] Speaker B: Yeah. Really correct. [00:12:14] Speaker A: Yeah. Place Google Search and become a coding assistant. And they're like now perhaps neck and neck with. With Claude, depending on your perspective on it. I've been using it to do Vibe coding. Right. Python scripts. It's like, I mean, for me, amazing to be able to use natural language and do coding and do things that like, I could not do before or very hard or I needed assistance and usually what was happening to me was like I'd set up something and then I'd run into some roadblocks and then run to another roadblock. And they'd be simple things, but they were to add up and be like, was in the right directory or I didn't have this like library downloaded and everything else. And I just get like. And I just go like this is not worth my time. I just go and find someone else to do it for me. And so now I like will do that. It's, it's mind blowing. So with AI, it's like I have that person now that I used to be when I was like, I don't understand this. And then Chad V or Quad goes oh, Gregory just blah blah blah and like, boom. Done. Like I'm like, whoa. Like, like I understand it all. I just, I don't want to be a professional software engineer. So the tools for me are like giving me superpowers. [00:13:18] Speaker B: I think that's, that's the distinction though. So Claude is becoming tools for the developer that actually knows how to develop GPT is for everybody else. So now I'm starting to see that GPT5 might actually allow people like yourself code as well or even better than software developers using Claude. That's exciting. I think that's huge. [00:13:41] Speaker A: Mind blowing. I mean like I was talking to team of developers the other day who like they sold some like library or framework and they were just like trying to pivot and like consumer like trying to do a totally different business because they don't think that there's going to be any market for like frameworks, libraries. There were like few businesses around these things. Like it's all just going to be built into these systems or like the system just create the library. Like it doesn't, you don't need to get a library anymore. I mean it's just like it's evolving so quickly. I mean that's what I saw with like in five years. I think I could build pretty sophisticated application with these types of tools. Yeah. Which is like important for me. It's a really important point. Like everyone's panicking about jobs. So my takeaway from this is that you still need to know computer science. And like you and I, like we've worked with computers a long time. I know a lot about how they work. I've seen them from their inception perhaps. And so for me a lot of those things are very obvious. But knowing all the very intricate details, like how these work is a lot of knowledge. And you still need all, all of that knowledge where like I know what Python is, I know what a library is, I know what a directory is, I know I like and like, it might sound silly, but it's only silly if you talk to somebody who hasn't had a lot of experience with this stuff. Like the, the like, you forget the like layers of sophistication on top of that. And it's only when you talk to someone a long time and try to like, like explain how you do some coding, you can see the knowledge is just tremendous. So I think that like people should still study computer science because I think, I don't think had a lot of coffee today. I know I did. Yeah. I think that like this, this point for me is really important. So the coding piece gets abstracted out, which means like the computer science piece becomes I think more important because I think the computer science piece becomes more complicated. So if the coding piece now is solved by agents, the way you string together hardware, components, libraries, languages is going to be really important. So understanding the strength of Python and the strength of MongoDB and the strength of like JavaScript and how this works, like you still need to know all of those things. You just don't have to be an expert in coding it. [00:16:14] Speaker B: I, I actually kind of disagree on this. I think it's going to be more important to have multi disciplinary. So yeah, software, software developers, you could study cs, but it's going to be more important for you to know something else. So like you have to be engineer, you also have to be, let's say a marketer so that you can directly apply, apply all of the, the stuff that you do to a complete adjacent field. So like what I'm trying to say is it's much more powerful to be a marketer who knows how to code than a developer who knows how to market. [00:16:56] Speaker A: I mean this is a really interesting point. So, so I mean it depends what you, it depends what we're talking about. So like will there will there be a need for like computer science degrees in the future? Yes. And actually into demand will increase. [00:17:12] Speaker B: Right? [00:17:12] Speaker A: I think, I think the skill set of like I'm an expert in this, this language is going to go away and there's a lot of people who like they just, there was teams of people that just did like Python scripts to like move data around and stuff and we don't need any of those people. And they had managers and they had product managers, like so I mean it's a huge change, right? Like not everybody was a computer science person, they were they went to like a coding bootcamp and whatever. Right. And so, and so that's where I'm saying I think is important. You're making what I would say is a different point but important point that like if you want to be really formidable. [00:17:51] Speaker B: Yeah. [00:17:51] Speaker A: Knowing I mean business and economics will always be valuable and that's why like it's more popular. There's a big joke that like all the Americans all learn economics and business degrees but most let's like foreign like people. I'm trying to think of the right term where within other markets where it's popular to do like it services up. Everyone's very focused on the production in terms of like getting an engineering agree. I've talked to, I know a lot of engineers who were like we're joking about this like Americans and like Silicon Valley. What you're saying very focused on like the business and the business model and less on the like implementation. But the implementation is really important I think understanding like how it works. But you don't need to necessarily be the like. [00:18:51] Speaker B: A carpenter to do the actual work. Yeah. Gotcha. So like to do. Yeah. [00:18:56] Speaker A: Doing the carpentry you still need to be an architect and you still need to understand project management, business, all these things but like just being an expert at like and that's where like a lot of these other. There's a lot of people that are just focused on those roles. Right. And they're less, they're less popular in the United States or like less people who like I'm gonna be like just so awesome at databases or whatever. It. [00:19:19] Speaker B: That's such a good point. So yeah I, I'm. I think like we'll have to still test out GPT5. I. I feel like I'm. I need to figure out the way to prompt it. Give it a lot more tool calling abilities. So to give it kind of inside of the prompt I want to define if it should be searching websites it should be using these tools instead of the way that I've been prompting it which is giving it lots of detail of the tasks that I want it to perform. So that could be a massive difference in its capabilities. And I will say currently a lot of people are complaining about how they're losing access to their old GPTs big for all where they saved up a lot of memory and saved up a lot of conversation so that. What do you think? What do you think on that point? [00:20:19] Speaker A: Yeah, I mean so I, I've heard something about this. Like it. It's it like the new systems not Necessarily compatible with like the old. [00:20:29] Speaker B: Correct. Huh. [00:20:32] Speaker A: It had an. There's an upgrade issue they moved to. [00:20:36] Speaker B: So essentially you lost all of the access to the old GPTs. It's all kind of rolled into 1, into 5 GPT. 5. I. I wonder why they did this. So like the API cost of using 5, it's dramatically lower now. Like much lower. So it could be a cost saving. Cost saving thing that they did. [00:21:07] Speaker A: Yeah. Who knows? I mean, I'm tempted to say it's a vision thing. I think Almond really is aiming to do AGI whether achieve it or not. I think he 100 is like sincere, but that is his mission in spite of all the hype. And so if you're. That's your mission. One GPT to, you know, Rain would be the product vision. It's conceivably cheaper, but it's also conceivably more expensive and more complicated to manage. So I don't, I don't know. I don't know enough about how it's built. [00:21:46] Speaker B: Yeah, let's see. Let's see where all of this goes. It seems like the new new agents coming out every six months and I wonder like what the next one would be. You got any predictions? [00:21:58] Speaker A: I didn't understand that the income, the backwards compatibility was a huge issue. I had read something about that, but I didn't, I didn't understand that it's. It's that dramatic. Like you literally can't use the, the old conversations. [00:22:15] Speaker B: You can't. So it seems like people are testing the new GPT and it's just completely different. Like the way that it responds, it's not as affirmed. It's a lot more like business, like, let's say. So people are not happy with it. [00:22:33] Speaker A: I heard people tell me that like, like, like the, the more let's call it like mental health and wellness side. [00:22:41] Speaker B: Correct. [00:22:42] Speaker A: Just like that type of stuff that people were like, oh my God, the personality is totally different. Like I lost my friend. I actually heard a bunch of people like, yep. You know, like some joking about it, some actually kind of serious about it. Like that it. It made such a huge difference in the personality, which is fascinating. [00:22:59] Speaker B: There's so many reasons and people are. It's kind of scary, don't you think? [00:23:03] Speaker A: It's like people, yeah. [00:23:07] Speaker B: And like I heard this like ridiculous story where the Prime Minister of Sweden or somewhere who tweeted or talked about during an interview that he uses GPT to make a lot of decisions. He converses with GPT. So you think, you think these Kind of shifts are. Are literally like, nation moving and kind of like, scares me a little bit how powerful these AI companies are starting to become. [00:23:40] Speaker A: There's some moments where, like, you. You see. You see how AGI, I'm gonna say, like, will emerge someday. Like, I think it's an inevitability. Like, when that happens, I don't know. I'm not even confident it's in my lifetime. But a hundred Years from now, 500 years now, like, it seems inevitable that, like, there will be, like. Like I even believe that. We've talked a little bit. Like, the whole machine civilization theory, like, seems plausible. [00:24:15] Speaker B: Like, it's very plausible. [00:24:18] Speaker A: Plausible. Like, like that, like, AI becomes so sophisticated, starts creating other AIs. They all have different personalities. They're attached to machines. They have machines that make machines. Like, there's a whole machine civilization. Like, it's. It's like, it seems like it. Like, I don't believe it'll happen on my lifetime, but, like, the idea that this could. Could exist or will exist, I think is. It's very real. Like, I don't think it's science fiction. [00:24:41] Speaker B: I don't think so either. I think it's. [00:24:43] Speaker A: No, I think we're. [00:24:44] Speaker B: We're very, very close there. So that. That's like. I'm more scared of that. Right. Where humans start relying on these GPT AI things that's not fully AGI, so it has no conscious. It has, like, no actual intelligence. [00:25:03] Speaker A: So they're announcing the food trucks. I'm gonna go on mute. [00:25:07] Speaker B: All right, let's do it. This is cool. What is. If you're typing in the chat, maybe I could tell the stream what is happening. What's. [00:25:18] Speaker A: They're. It's like, it's lunch time, so they're announcing the. The food trucks are outside. [00:25:24] Speaker B: Nice. Okay, gotcha. Are you gonna go grab some lunch or. [00:25:28] Speaker A: No, no, dude, I'm. I'm 100 vegan again. [00:25:32] Speaker B: Oh, nice. Gotcha. That's awesome. [00:25:38] Speaker A: I. I'm. I'm the most annoying person again. Like, ride a bike everywhere. And vegan. Like, everyone just. I, like, it's embarrassing almost. I had to do it for. For health reasons. I. I had some, like, issues with my cholesterol, and it's, like, genetic, so I have high cholesterol, which is like, I would need a. [00:25:59] Speaker B: Really. Which is surprising. [00:26:01] Speaker A: Yeah. And so, yeah, so I had to change my diet and even went on some statins to, like, reduce cholesterol because it's. It's not healthy for you. And so. Yeah, so, like, Now I have. And I was a vegan when I. For like many years. Now I joined my doctor. I'm like, well, at least now I have like a real medical excuse, like, to be a vegan. Like, yeah, just not just annoying. Like, I have a real condition, which is cool. I'm like, I actually enjoy vegan food, so I don't know what they'll be for me to eat there, but I'm not too worried about it. I have to like, bring my own. [00:26:40] Speaker B: Food everywhere now you're European, Brooklyn, I'm sure you'll find a. Yeah, exactly. Heft vegan spot. [00:26:50] Speaker A: All right, so it looks like they're done with the announcements. Yeah, chat GBT5. So, so there, there was some stuff that I read. I'm playing around with it enough. So, like, they're trying. There were people trying to find the article that you sent me. We should show that they put the one up. Yeah, yeah. Article. I mean, if you believe it, like, the guy made some pretty interesting statements about the capabilities and its ability to code and solve problems that it couldn't solve before. Did you believe this article? What was your take on it? [00:27:27] Speaker B: I think this is correct. I agree. GPT5 is moving towards tool calling. This was probably the most informative graph, this entire article, and it's kind of changing the way that I'm thinking about GPT5 and all AI models that ChatGPT will or OpenAI will release in the future. Right. So like this maybe. Let me summarize. We're slowly moving away from a data science problem of improving LLMs. We're now moving towards engineering problem. And the engineering problem is essentially, essentially chaining these tool calls and creating a context around these tool calls for the LLMs to have access to. So another thing that they're talking about is this is like the stone age of. The stone age of LLMs, which kind of makes sense. Right. We're now teaching LLMs to use tools. It's raw intelligence kind of peaked out and then now it needs to become engineers and use tools and that's how we will move the progress forward. [00:28:58] Speaker A: Yeah, it's. It's quite like. I mean, it's all very logical. Right. And like, there's a lot of other people on the, in the, like, you know, AI world who have said that, yeah, the, the intelligence piece is getting tapped out, which is, it's logical. Right. You've got this, you've got a acceleration curve, I guess is the right term. [00:29:20] Speaker B: Yeah. [00:29:20] Speaker A: And that there's a point where, like, it's just diminishing returns. Right. This stuff's clearly smart. I don't think it needs to be so smarter. I think it needs more capabilities and it needs to be able to do more things and do them effectively, do them repeatedly, do them correctly, like names a whole host of things they need to fix what they're trying to work on. And they made, made some really interesting progress. [00:29:45] Speaker B: I think so. So I'm, I'm hopeful of all of these things. I think this is the right move. What this means is we're going to see essentially a marketplace blow up of all of these different tools that support AI agents. And what this hopefully means is in two years or so we'll get to replace all of the browsers. The way that humans interact with the computer, if we can do that and that. That makes a ton of sense. [00:30:23] Speaker A: Yeah, I mean that's. So we were talking a lot about that today because with Flora and Cheat and all the MCP server stuff, like yeah, it's, it's clear to me that, that that framework is, is like you and I, we talk about a long time. Like it's, it seems like that's a replacement or an alternative to HTTP. Like some of the core elements of the Internet and like a lot of services are going to migrate to these autonomous agents. [00:30:50] Speaker B: Correct. [00:30:50] Speaker A: And they're going to move away from the browser. You can even see like, you could see it now. Right. Like I guess like if I was a free at really high level AI mode on Google, the reduction, and the reduction impact of like traffic search, the whole methodology and like way that we're familiar with the entra page metaphor. Right. Like apps, apps kind of change the game in terms of like showing people what's possible. [00:31:18] Speaker B: Yeah. [00:31:18] Speaker A: Like an app on your phone. Right. Which I always thought was like a much better experience at least than like a web page. [00:31:26] Speaker B: Correct. [00:31:26] Speaker A: Like autonomous agents. Like. Yeah, I think a lot of the like web metaphor goes away and like you'll have an app or something you interact with and then the agent like just does what it needs to. Needs to do. I think for the average person. Yeah, they're just like, they go on their phone and they'll just turn their agents on and probably built right into like iOS or something and they'll have an agent that does a lot of things for them. [00:31:51] Speaker B: Correct. [00:31:52] Speaker A: And so it's, it's really going to just change how the Internet like works. [00:31:56] Speaker B: I think. So like Internet will actually disappear into the background. Right. Like that's the point. Which right now everybody is plugged into the Internet literally every single day. Like, this call is happening over the Internet. But in the future, like, I. I really do believe if everything. All of this goes well, I stop using computers. Like, full stop. [00:32:19] Speaker A: It's never gonna happen. [00:32:20] Speaker B: Never gonna happen. [00:32:21] Speaker A: You'll do everything from your phone or you'll be glad. Like, what do you mean? [00:32:25] Speaker B: Like, you know, anything. Right. Like, I sent the omi AI thing where you just have a device and it's constantly tapping to your. Your everyday. Right, Right. So, like, some kind of wearable where you talk to it. It'll give you some kind of feedback, and things get stunning in the background. [00:32:51] Speaker A: I don't. I don't know. Dude, this. This. This demo is corny. [00:32:54] Speaker B: It's corny. It's terrible, right? This is, like, straight out of Black Mirror with a chip on. On their head that's, like, fully Black Mirror. [00:33:06] Speaker A: And then he's, like, trying to use it to pick up the girl. It was the nerdiest marketing. Like. Like this. So, dude, this is why nerds need good marketers to not make videos like this. [00:33:23] Speaker B: Nowhere. Yeah, absolutely. No one would be. [00:33:27] Speaker A: Well, like. Like, just look at this. Really? Basically, like, it's like, if this is your pitch, it's not for girls. No, Right. Like, the first thing I thought about, I really went, like. And then, like. And like, there are some guys who are good at meeting people. They don't need it. So you're talking. It's just like, oh, my God. This marketing is just. It's just terrible. [00:33:51] Speaker B: Absolutely. [00:33:53] Speaker A: And, like, we've seen this before, too, with, like, all the glasses. Remember all the glasses demos from, like, I don't know, seven years ago. There was one that was, like, exactly the same as this. It was a different alternative, and the guy was doing the same thing. Like. [00:34:06] Speaker B: Yeah, yeah. [00:34:07] Speaker A: I don't know. Like, I mean. I mean, we're looking at these demos for a long time, and so, like, will these things be a reality? I think it's a long list. It's got to go a long way to, like, replace the computer or the phone, perhaps. [00:34:24] Speaker B: Or the phone. Right. So I don't think it will replace the phone because the form factor is just too good. [00:34:32] Speaker A: It's too good. [00:34:33] Speaker B: It's just too good. But replacing a laptop, I could be convinced. I could be convinced. I mean, lucky to stop using a laptop or, like, an operating system in general. Right? Like, I. I don't want to be logging onto the Internet using a browser, clicking around, using a mouse, a keyboard. [00:34:54] Speaker A: I mean, maybe. I mean, I Don't know. I wear glasses, so I'm more. I'm more apt to believe that like, at least I would wear. They lose the normal glasses. You wear glasses too. So like, for me, like, AI glasses I think is like, I see the benefit and like, I don't think. I think. I think there's less and less stigma all the time too. Like wearing glasses. Yeah, like. Like a little chip on your head or like. I don't. I don't know. I don't know. I mean, maybe. I mean, I do think that, like, that's what I think. It's what the device Johnny Ives and Sam Alton will make. I just think we'll have better marketing than this. Perhaps. [00:35:33] Speaker B: If they make a wearable that's anywhere close to this. I feel like we've already seen enough that nobody will be open to it. [00:35:41] Speaker A: I think it'll be a little like mirror from this. A little pod thing that you walk around with and like listens everything that you say. [00:35:47] Speaker B: Yeah, yeah. [00:35:49] Speaker A: But like, the problem with all these device ideas is that it's listening to everything I say. You can just build that feature into the phone. You can build that feature into your smartwatch. Like, you don't necessarily need a third device. You can build in the glasses. Like, if it just listens and then does stuff for you. Like, you don't really need a new device. [00:36:14] Speaker B: Right. So the form factor is still a question. [00:36:20] Speaker A: I think it's really hard call on these things. I mean, I guess the only. The only data I have is that the keyboard and mouse still exists and they're popular. So I think it's actually a really. Yeah, a really important point that like, the guy. I mean, there's videos the guy on bend the mouse saying like two things. He says, this is a dumb name and I'm sure someone comes in better. And then he says, I think someone will invent something much better in like five years or something. Like. And it's still around and still called him out. Nice. So he invented something like, quite good. I think it's like a testament to his design ability. [00:36:59] Speaker B: Yeah, I totally. I see your point there. I agree. [00:37:04] Speaker A: But I don't know. I think it's cool. I want to try them. I like the smartwatch. I have a smart watch. I think they work really well. [00:37:14] Speaker B: That would be a cool hackathon idea, is to come up with the form factor that people actually would use. [00:37:21] Speaker A: What other options are there? You got a phone, you got glasses, you got a watch, like, and you could walk around like a Little necklacey friend thing. Like a little. Your little Pokemon, you know, I don't know. Maybe they're like a cute little friend. Maybe they'd be more. They kind of for kids, but. But it was like a little like a robot dog that actually I would love that. I want a little eyeball robot dog. [00:37:48] Speaker B: Thing that just follows around. [00:37:50] Speaker A: Yeah. It's my little AI companion. [00:37:57] Speaker B: Interesting. Interesting, interesting. Yeah, yeah, yeah. Or it just gets built into like everything, you know, I mean that seems. [00:38:06] Speaker A: The obvious angle too. Is that like you. Why would my, like my personal. Let's explore personal agents. Why would my personal agent not just live in the cloud and just wherever I am, it exists. Privacy. [00:38:17] Speaker B: That's. That's the. [00:38:20] Speaker A: Yeah. Besides that, of course. [00:38:23] Speaker B: Do you want some like intelligent thing that knows everything about you to living in the cloud? [00:38:29] Speaker A: Well, I think. Dude, that's probably. Well no, I think that. I think that one direction that Apple could conceivably go in is that they do a local LLM and like you're like your personal agent is local on your device for security reasons. [00:38:47] Speaker B: Right. Oh, and then everywhere you go see that future I'm open to. To. Right. [00:38:55] Speaker A: I mean. Yeah, it's quite logical. Right. Like it would have to be. So it's like you're like this thing has. My personal LM is on there with all my stuff. Like it has to be a better model that is say like I would love for like it to be like a vault kind of how crypto works. It's a volunteer. [00:39:12] Speaker B: A wallet. Yeah, yeah. [00:39:13] Speaker A: Everything is everything. Like it's like, it's like a personal data wallet. [00:39:18] Speaker B: Right. [00:39:18] Speaker A: All of my life is there and it's this hardware thing. [00:39:22] Speaker B: Right. [00:39:22] Speaker A: There's good bad to it because like people would want to make it really valuable. Want to steal it. I know you got a back. There's. I guess there's ways to solve for it, but I would be a fan of that. [00:39:36] Speaker B: Yeah. And then you walk around with it in your pocket. I see, I see. We're getting into like. Yeah, you walk around it and then maybe if you go to a store, it syncs up with the store. Store. And the store knows exactly who you are. [00:39:53] Speaker A: I mean it could have all kinds of like location based information and it knows where you are and you could like return. I mean there's. There's a lot you could. You could do with it. I think it's cool. [00:40:06] Speaker B: I think. [00:40:06] Speaker A: Anything else you want to talk about? GPT5. We've covered all. [00:40:10] Speaker B: I think we covered everything. GPT5 was this week. Yeah, I think that's, that's pretty much it. [00:40:22] Speaker A: That was big. Any other. [00:40:25] Speaker B: It's big but like, you know, I don't think it was that it wasn't mind blowing. Which I. As a. From a marketing. From a marketing perspective, do you think the GPT5 marketing was a success overall? Is that a hard question to answer? Yeah. [00:40:45] Speaker A: Yeah. So, okay, this is really good. I think it's a good point. What is marketing's job? My job is to get people focused on the product. I think that, I think, I think it was successful from that perspective. Like he did capture the attention of world and like, you know, did he over hype it? Let me do that. I think that's a sign that the marketing works. Like what products that like all the fanatics are excited about and then they see it. Are they underwhelmed? It seems to always happen. I can't imagine. I can't come up with a concept where like someone came out and everyone's like, this is even better than I thought. Like, what is that product? Like I know the Apple Watch launch and it's cool. It's exactly expected. Right. Or the Vision Pro launch and it was kind of a disappointment. Or each iPhone comes out and like it's cool. But like I try to think of the part where there's a lot of hype. They came out, everyone's like, this is even better. [00:41:39] Speaker B: Oh, it, it had a lot of hype. Correct. Correct. Right. It's usually like not a lot of hype. People try it at, their minds are. [00:41:47] Speaker A: Blown and it exceeds their expectations. [00:41:49] Speaker B: Right. [00:41:49] Speaker A: Which is, which is rare. That happens more with like startups or like people like throw out some kind of product that takes off. But it probably has like a lot of, a lot of hype, a lot of marketing behind it. Generally there's people who are just like too excited about it and then they get it and then it doesn't live up to their expectations. I think it's very human. Like, like in general, like the new album comes out or whatever it is. Like they get the pair of shoes and you're, you're like, oh, that's it. Right? [00:42:18] Speaker B: Exactly. Yeah, it's like 3.5. Right. The first time you tried 3.5, your minds were blown. You're like, I've never heard of GPT. [00:42:26] Speaker A: Awesome. Right? No one, no one ever heard of it. And so it was like mind blowing. [00:42:31] Speaker B: So. [00:42:31] Speaker A: So from that perspective, I think like they did a good job. They got everyone focused on it. And he continues to like push the hype, which I think is his job. I think he should, like, you want people to want to capture attention. That's the job of marketing. And people are disappointed. You know, that's bound to happen when you have something that's that, that highly anticipated. You know, I think for like rollout. The chart stuff. Oh, we didn't talk about that. Chart stuff was a bit of an embarrassment. So that's interesting. Did you see this? It's well documented all over like X. [00:43:09] Speaker B: So you're talking about doing their presentation, right? [00:43:12] Speaker A: During their presentation there was some chart. [00:43:14] Speaker B: Yeah, yeah, it's all over. Right. [00:43:16] Speaker A: And like, dude, like, like a joke, I'm not a mathematician, but I couldn't figure out like why those charts were set up that way at all. Like, they were just wrong. They were like super odd and everyone saw it and it just like it really, like it was bad in the sense that like, I think it like really reduced confidence. Like, because where I think AI is struggling right now is in the enterprise. And so enterprises have a lot of like real concerns. [00:43:43] Speaker B: Yeah. [00:43:43] Speaker A: Security hallucinations, repeatability, quality. And so their own rollout has like big issues with like charts being calculated wrong. Like doing things that are like quantitative. So it should be able to do them. It's not a qualitative thing. Like it should be able to just do those correctly. And for whatever reason, like it was wrong. They did fix the charts, like if you go and check if they're all correct now. But I thought that was a pretty big gap and like, look, they're trying to go faster. Startup, like I, I know the other argument or the other side of it and I would still say that like getting things out there and like having some issues is part of the process. [00:44:25] Speaker B: Right? [00:44:25] Speaker A: I think so. I think that's what like high tech startups are all about. It's not a medical device. Right. You don't want to release your pacemaker and say, oops, we didn't catch that. Right. Like that. So different categories have different requirements in terms of like expectations on rollout of products. So I think that like that they, that they took the bold step of like releasing it. Some flaws is okay. [00:44:48] Speaker B: Yeah. [00:44:49] Speaker A: But it's not going to do them any favors in enterprise. [00:44:55] Speaker B: Got it, got it. So we'll see. I mean, I don't know. Where do you think that they are going? Well, they must be right. Going after enterprises. [00:45:07] Speaker A: They like, in order to meet. Yeah, in order to meet like their targets with like their evaluations and stuff. [00:45:15] Speaker B: Right. [00:45:15] Speaker A: They've Got to have everybody on the planet be a subscriber to this thing. And they would need, like big companies need to adopt this in a big way. And also they need to deliver on the efficiencies. [00:45:28] Speaker B: Right. [00:45:28] Speaker A: So this is another interesting debate. I think it's starting to emerge. Like, is the investment in AI worth it? Will the efficiencies emerge? And so actually, what do you think? I haven't really, actually I have a clear like perspective on this. But I'm curious what you think. [00:45:48] Speaker B: I think we need something. I think we're, we're like super stale for the next tech, tech innovation. [00:45:56] Speaker A: So like, do you think all this investment like spending, like we're spending 2% of GDP on AI infrastructure in America? [00:46:07] Speaker B: I think we have to. I think we have to. Right? Because we have a clear enemy or we have a clear enemy that's, that's doing this regardless if we do it or not. [00:46:19] Speaker A: I saw a cool chart that said, okay, the railroad build out in the 19th century United States had peaked at 6% of GDP. [00:46:29] Speaker B: Right? Right. [00:46:31] Speaker A: Telco build out I want to say was like four or five. So 2000s. Like remember we had dark fiber. Do you remember? Dark fiber was like all that fiber I got laid. They get to use eventually got used, right. And all the railroads got used too. So we're at 1 to 2% right now. My take is that if you look back at all these booms, the capacity does eventually get used. So I don't think the build out is kind of like where you were coming from. We got to win the battle in AI and we need to like spend an enormous amount. I wish I had the number. Like the US space program to put a man on the moon started to eat up a big percentage of GDP. I want to say like 3 or 4%. It was an enormous amount that was spent on like Apollo emissions. A lot of people actually are not aware of this at all. Like how much money the US government put into the space program to. And I bring this up because there's always this like, I wish we could get back to how cool it was the Apollo missions and space program. It's like, yeah, dude, we spent like 3 or 4% of GDP on it. Of course it was cool. Like we're not spending that much money on NASA. But like, the point is that these types of like, I want to call it like inflection points in technology require enormous amounts of investment. And I don't think they're unwarranted. I would even make an argument that they're not because of the hype. They're because they are real technologies that are like battles that need to be won. And if you don't win, second place is a terrible place to be in these things. So you want to be the leader for the United States in AI? We need to spend more, we need to win, we need to build out. [00:48:17] Speaker B: We have to correct all the infrastructure. [00:48:19] Speaker A: I have another theory. I never shared with you, another theory about, like, why there's this massive AI infrastructure build out. And so my theory is that this is gonna get everything. This is. This. This is gonna border that conspiracy theory land. I think. I think the Pentagon, okay, the defense industry, I think they actually. I think that actually estimates how much AI capacity they need to win an autonomous war. And I think if you imagine that, like, the future is autonomous battles and autonomous war, you could calculate how much capacity is required to power the number of drones and systems required to win a war against China. Yeah, my guess is that this is where I like to joke like, not mathematician. But my guess is that it is enormous and that I ultimately think that, like, power consumption, power generation, let's say power generation and AI processing capacity at peak is the number one thing required to win an AI or autonomous war capacity than your enemy. You need to be able to stand by more drones, more autonomous tanks, more like, aggressive hacking initiatives. Like, you need to be able to overwhelm them with more. So the only way to do that would be to literally have more processing capacity and the power to, like, in the general power generation to power it to win the war. And I think there is a whole plan in place that we don't know about, or it's not. At least, let's say public, public, that the Pentagon and the military, all of those dia, all those people are fully aware of and are fully implementing 100 consciously. And that, I think, is the main driver behind the AI infrastructure build out in the United States. [00:50:25] Speaker B: All of this. So basically, it's computational power plus energy. [00:50:29] Speaker A: That's basically the reason I've heard, like, anecdotally, right? There's been generals who make statements like, if, if this is what came. If, if. If, like, China was to invade Taiwan, they're like, we are gonna rain down like, a thousand drones. Like, they made these statements, right? About, like, yeah, which we've never seen, like, what type of, like, capabilities we would deploy. Which is where I came to the conclusion, like, God, like, if you're gonna fly, like, thousands and thousands of drones, like, you need massive processing power in order to do that. Right? Which means that, like, the war is actually won by having that. Like, you can't be like, oh, we don't have any more chips left, sorry. We're just gonna, like, not be able to launch any more drones today. No web, no WI FI access from you. Right? Like, it can't work like that. And so that's where I became convinced that. That that was really what's powering it. And I think, like, Microsoft, Google, these companies are like, very close to the US Government. And Amazon, right, they all have contracts. Palantir, they have contracts with the federal government, and that's what they're working on. [00:51:42] Speaker B: 100. I completely. I completely agree here. [00:51:47] Speaker A: I hope no one. Like, I haven't heard anyone. I've never heard this theory floated anywhere. It's my own crazy theory, but I think I believe it. [00:51:59] Speaker B: I. I believe it too. And just the fact. I mean, like, just the fact that China is also doing this. Right? Like, I mean. [00:52:06] Speaker A: Right. I mean, China's probably doing the same thing. And so it'd be easy to, like, to calculate this. And so I don't know if you've ever heard this, but, like, there's still a rumor, I believe, that the US did the same thing with the Soviet Union and that we knew the Cold War was a way to bankrupt Soviet Union. I've heard other people say, like, this puts too much. It gives too much credit to CIA, But I think there are really bright people who work at all these agencies. Don't ask me how I know this, but they're not foolish. Are the kind of people who. They are the kind of, like, thinkers that would engage in this kind of thing. Like, hey, over 20 years, if we build more tanks and we're very public about it, right? Like, the Cold War was super public, that it'll bankrupt the SO Union. And we have economic estimates and ways to calculate gdp and like, we understand what the nature of their system is. And. And that's essentially what happened. So, like, I think the same type of strategy at a very high level is being employed in, like, Battle of China. [00:53:19] Speaker B: Everything. Basically. That's. That's kind of like the. The strategy is we have to outspan. [00:53:26] Speaker A: And I know that now. I'm on a list that we made it public in this video. But. But that's okay, buddy. Yeah. I mean, that's. That's how I see it. [00:53:40] Speaker B: I love it. [00:53:42] Speaker A: Awesome. All right, dude, I'm gonna go check out the food trucks. [00:53:45] Speaker B: Okay. Good luck on finding good options. [00:53:50] Speaker A: Yeah, yeah, I'll do some more tweets and stuff. I'm, like, gonna post all the things that we're doing here today, and tomorrow we'll announce the winners, and so I'll write some stuff on social media for that as well. So. [00:54:04] Speaker B: Awesome. Enjoy your day and have fun. [00:54:06] Speaker A: Awesome. Take care, buddy. [00:54:08] Speaker B: Bye.