I Made a New Friend (and It's Not Michael)

[0:20] Matt: He sent me a reminder of what we said we wanted to talk about this week. Yeah. And I said, perfect, thank you. And he wrote back, have a great recording. Looking forward to hearing the episode later tonight.

[0:30] Michael: That's great,

[0:34] Matt: man.

[0:34] Michael: So I think that just

[0:35] Matt: tells us we need to go to the shores because

[0:38] Michael: we're sitting there so much right now. So Matt made a new friend.

[0:45] Matt: That's right.

[0:46] Michael: I think he's going to replace me.

[0:49] Matt: So tonight we're going to talk about artificial intelligence. Yes. And OpenClaw. So we kind of mentioned this at the end of the episode last week and we wanted to get into it, but I had just, you know, dipped my toe in, so I wanted to wait. I've now had like a week and a half with this thing. I

[1:09] Michael: would love. I want to know your synopsis of, like, I've got a viewpoint as far as, like, you getting into it, but then I want to hear where you are right now with it.

[1:19] Matt: Well, first let's explain what it is. Okay. For anyone who doesn't know. And it's kind of hard to explain what it is, actually. So OpenClaw is this open source code project started by a guy named Peter Steinberger. Is that right? Right, Steinberger.

[1:35] Michael: Something like that.

[1:35] Matt: Something like that. And it. So what a lot of people are using it to do is they're purchasing a Mac Mini and installing this software on the Mac Mini. And then the software hooks up to the API of what you currently think AI is. So it hooks up to the API of ChatGPT or Grok or.

[1:58] Michael: Anthropic.

[1:58] Matt: Anthropic. And then it just runs 247 on your computer and it uses the API to these chatbot AIs to then start orchestrating and executing on tasks. So you can give it access to things like email, you can give it access to things like code repositories. It has different channels for communicating with you. You and I are both using the Telegram channel. So we go onto Telegram to talk to it, ask it to do. And conceptually, I don't know what world we're living in. You got into it. And I was very nervous to do it because there's some pretty obvious security concerns around having an autonomous agent running 247 with access into all kinds of systems. What's it gonna do?

[2:54] Michael: Yeah, totally.

[2:55] Matt: And so I was hesitant to get into it, but it was starting to be talked about enough in the circles that I travel in on X, that I just had. Well, you tried it.

[3:04] Michael: I wasn't convincing enough to. I'm sorry.

[3:06] Matt: So I finally tried it and I'm struggling with some things. Like, I. I don't know how to conceptualize the world we're in. I don't know how to conceptualize what, like, what my purpose in that world is. So here's one way to put it. My position or theory on AI from within my own industry, which is tech. You know, I've been an engineer and a developer and a designer for a long time. I don't know, three or four months ago. My position was for people who know what they're doing, AI is going to be a really powerful tool for people who don't know what they're doing. It's going to lower the barrier to entry. But you still need to understand the core concepts of things like coding web applications in order to really, you know, oversee what the AI is doing. I thought that it would be an inconvenient, convenient tool, but I didn't think that it was going to disrupt the industry all that much. I'm going to tell some stories about why, how I came to this. I

[4:14] Michael: would love to hear them.

[4:15] Matt: But my position now is there is no reason for a human to write code ever again. And the implications of that are pretty. I don't know if it's scary.

[4:27] Michael: Well, first of all, I don't know if you guys know, like, what Matt does. Like, he's worked on projects, huge projects like Lockheed Martin, City of California, I mean, state

[4:38] Matt: of

[4:38] Michael: California, you know,

[4:39] Matt: United States Postal Service.

[4:42] Michael: So this is not like a little, oh, he knows how to code. Yeah, kind of guy.

[4:49] Matt: And

[4:49] Michael: run the projects.

[4:50] Matt: I've spent a lot of time running projects for really large companies and developing and committing code for software that they use. So, like, for example, for Lockheed Martin, the team that I was running built a piece of data analytics software with the intent of helping Lockheed Martin reduce the budget overruns of the F35 aircraft project, which was like, famously over budget. Each. Each plane was costing something like 120 million. They were only supposed to cost something like 90. And they didn't really know, you know, why. And so we built them this software. So we wrote code, you know, in a highly secure environment for a Department of Defense contractor. So, yeah, I've seen, you know, what it takes to develop software that works and deliver it reliably. Um, I am now convinced that not only do. Does no one have to write code anymore, you would be stupid to do it. So another kind of paradigm Shift. A lot of people were using AI as we previously thought about it, hooked up into their code writing software and AI was helping them with the code that they were writing. And that was why I was like, oh, that's a helpful tool. I have from my sort of previous entrepreneurial days. I have an application that I built that was intended to support my parents therapy practice. They were running some software that my dad put together on FileMaker Pro in the early 90s and years ago I started building them a web application version of a solution for this. Manages their calendar, their billing, their session notes, things like this. And I planned to launch it as a software, as a service for all kinds of reasons that never really got around to it. But my parents have been using it. So I had this code base sitting around, I haven't touched it in five years. And so I took my OpenClaw and I just kind of turned it loose. I told it what I wanted it to do, which was, hey, this is an old code base. Start with modernizing it, upgrading all of the installed packages and things like that. And it did it. I didn't tell it how to do it, I didn't tell it what exactly to do. I just told it the concept and it did it. And then I said make a pull request into the remote source and it did it. And I reviewed the code and I was like, holy shit, this is really good.

[7:33] Michael: So it's not just crap code. You kind of hear that on the webs as far

[7:38] Matt: as like, like AI slop code. None.

[7:42] Michael: That's so crazy.

[7:42] Matt: And then what's even more crazy. So I want to explain a couple of terms that I'm going to use because most people probably have heard of GitHub. GitHub is just a one popular implementation of a very old project called Git. And Git is what allows you to track versions of files. So track when things change in a file. And so when a bunch of people are working on a single code base, each individual checks out a version of that code base, makes the changes they want to make to it, and then opens what's called a pull request, which is a request to pull their changes into the source code. And when a pull request is opened, you can go and look and see what's the difference? What's the difference in like, if you just imagine, you know, a code base being a single file, the pull request would pull out every line that had changed and then the rest of the individuals working on the project would all review that. So everybody has visibility of what's changing in the code base. And, you know, in a good. A good flow code doesn't get merged in unless it has proper approvals. So I gave my OpenClaw access to my git repository. It checked out a version of the code, it made changes, and then it issued a pull request. I reviewed the pull request. That alone is kind of mind blowing because that's a paradigm shift from AI is helping me write code to AI is behaving exactly like a developer and working with it for about a week and a half. I've come to realize that there really isn't any difference between working with my OpenClaw on code and working with a real developer who works remotely. And I just talk to them on Slack. The next mind blowing thing was I didn't think, I thought AI would write slop code, especially if you gave it, if you released it on an entire code base, especially on an old code base that wasn't up to modern documented standards. Not only did that not happen, and it brought it up to standards. Then I said, okay, I want to make a new feature. I want you to analyze the database of this code project so you understand the different kinds of things that it does. It tracks people and users and sessions and clients and on billing and all of these things. And then I want you to suggest three new features that we can add to the Dashboard. It came back with really good ideas. I chose one and I said, let's go with this. In a matter of minutes, that new feature arrived in my code base, and my OpenClaw served it locally on its own machine, understood the VPN between our two computers, and sent me a link to where I could view it on my computer. And I was like, that looks fantastic. Make a pull request so I can review the code. Now this is when it really blew my mind. Not only did it work the first time when I reviewed the code. So this is a code base that has tens of thousands of lines of code in it that I all wrote, but I wrote all by hand. It matched my coding style. It understood what it could reuse from what was already there. Hmm. And when I reviewed the pr, it felt like I was looking at code that I wrote and I didn't have any changes. And that feature is now live.

[11:26] Michael: That's crazy.

[11:27] Matt: On an application that was never live before, it was only available to my parents, is now running live in a production server. And that first code feature I asked it to write was the ability for somebody to sign up and create an account. And it made it. And now that's live on the Internet right now. So this is just one piece of it. But you would be stupid to write code. The AI does it just as good, if not better, and it does it at like a thousand times the speed.

[12:01] Michael: Yeah.

[12:02] Matt: Especially when you move it from trying to help you write code to just telling it to write code. I didn't have to tell it how to do it. I didn't have to tell it to match my coding style. I didn't have to tell it to match the conventions in the existing code base. It just did it.

[12:17] Michael: I think this is why this is the most perfect podcast for us to talk about this, because you understand the logistics and technicalities of all of what you're talking about, and I don't

[12:27] Matt: in that area.

[12:28] Michael: In that area. And it's just sort of like my experience versus your experience is so wild. You're coming at it from like, hey, I've done this before and I'm gonna make sure you're doing it right. Where I'm like, I just want to do this thing do. And it's like, it's done. I'm like, cool.

[12:46] Matt: Yeah. I don't know how much you want to talk about the projects you're using yours to work with, so I'll just leave that up to you. But a couple of this kind of feeds into. I don't think we know the world we're living in, because we used to think we knew the things that needed to be done and then we'd find the tools to do them. We now have a general purpose tool that can do that on its own.

[13:10] Michael: Yeah.

[13:12] Matt: Like, another example that kind of blew my mind was so last week before the podcast, I sent my OpenClaw message and said, you know, hey, here's a project. So I've got IT tracking, like 10 different projects in my life. I've got it helping me with sovereign goods. I've got it helping me with social media. I've got it doing all kinds of things. But last week I said, hey, I haven't told you about this project yet, but we have this podcast. Here's what it's called. Here's what we do. I was like, any way you can help us out with this? Just come up with ideas and let me know. And it suggested that at the end of the episode, I send it the MP3 of the episode. So we did this. And this is what's wild is you don't send it like, you send it like you would send it to another real person. I uploaded it to Google Drive and then I send him the link. It's the other weird thing is like referring to it as him.

[14:09] Michael: Yeah, totally.

[14:11] Matt: I mean, he does. You do give your OpenClaw a name. OpenClaw is the name of the software. But my agent has a name, which I'm not going to share. So I uploaded the file to the Google Drive, sent in the link, and within, what, like three minutes?

[14:29] Michael: Yeah, least.

[14:30] Matt: It came back with it, transcribed it, summarized it, wrote a title and a description.

[14:36] Michael: And he fact checked us

[14:37] Matt: and he pulled out references, all the references that we made. So books we mentioned, quotes. Last week I mentioned a movie, but I couldn't remember the name of the movie. He went and found the name of the movie and added it to the references list. And then he fact checked us,

[14:58] Michael: which is awesome

[14:59] Matt: on our claims. I didn't tell him to do any of this. He just did it. And it felt like, holy. A problem that we've had for years. Which in the early days of this podcast, sometimes we would go a long time without putting episodes out

[15:15] Michael: totally.

[15:16] Matt: Just because figuring out what to call it, how to describe it, pulling out quotes, all of that was. It's a lot of work. And at one point I actually hired a personal assistant, like a real person from upwork, and I can't remember what I was paying him, something per episode. He'd actually sit down and listen to it, pull out key themes and quotes, and write a description and a title. And that would take forever. And I was never really happy with his output. So finally we found it's like

[15:46] Michael: 24 to 48 hours later that.

[15:49] Matt: So finally we got rid of him. And our workflow for a long time was right after the episode, after we record the episode, while it's all still fresh on our mind, we'll sit down and kind of review what we talked about and write a description and a title and try to pull in the references. And that would take us sometimes half an hour, you know, and honestly, like, after a long in depth conversation, I never really have felt like I had great creative energy to do that. The actual content that he came back with as the description and the title and the summary was as good. And I told him beforehand, I said, go look at all the titles and descriptions from the previous 267 episodes, because that's all on the Internet. So he went and read them all. And then the stuff he wrote was like, I had written it. And I felt like the inventor of OpenClaw. He didn't make OpenClaw to solve problems for podcasters. Right? He made it for something much larger. But it just so happened to solve a problem that we had never figured out how to solve. Really? Well, just in our podcast.

[17:06] Michael: Yeah.

[17:07] Matt: And it was amazing.

[17:11] Michael: Totally.

[17:12] Matt: So I feel like I'm just scratching the surface with what to even do with this thing, but my mind is blown.

[17:21] Michael: Yeah. It's like, we don't know where we're going from here.

[17:24] Matt: We don't know where we're going from here.

[17:25] Michael: So this is fun. Like, I'm kind of. Matt's crazy, but I'm like, crazy crazy. So it's hard to compare us, really. Like, I think I'm. I'm. Would you say I'm crazier than you?

[17:43] Matt: I don't know. You don't know? I wouldn't call either of us crazy. Really.

[17:46] Michael: Well, maybe that's cuz I know I'm crazy. Anyways, I was going on for about two weeks of like, just like, oh my gosh, this is so crazy.

[17:55] Matt: Yeah.

[17:56] Michael: And I was like, okay. All right. And then

[17:59] Matt: I didn't believe it.

[18:00] Michael: Didn't believe it. And then I. I don't even think it was 24 hours. You were like, holy.

[18:06] Matt: Yeah.

[18:07] Michael: And I was like, I know. It's, it's, it's not like AI like you've ever experienced it before. And that's really hard to explain until you.

[18:15] Matt: Yeah, we thought. We have thought, I think culturally that AI is a chatbot. AI is something you talk. You can talk to. And that's kind of the interesting thing about it. No, this thing is behaving in all of the ways that a remote employee could and does behave, but 100 times faster.

[18:33] Michael: Yeah.

[18:34] Matt: More reliable, always on. Doesn't get tired, doesn't take breaks. I mean, the implications for writing software alone, it's like those are at the forefront of my mind. Just because that's my business. I'm not even sure how to think about the implications for, like, everyday life. I mean, like I said, I'm using it to track a bunch of projects. That's been really helpful. Hasn't like, blown the roof off anything yet, but.

[19:04] Michael: Yeah, it's

[19:04] Matt: only been a week.

[19:06] Michael: But it's not only that it's doing that. It's, it's, it's pushing you forward. More forward than you would have been if you spent time on it 247 yourself. Like, there's, there's, like, there's so many of the things that I'm working on. I give him the project and man. So just to give you a reference, last year, I was going to develop this customer app and it was going to be, you know, a way to use Bitcoin and Stablecoins. And I did like a three to four month feasibility study on it. It cost me $4,000, but I got, I mean, this is huge discount that I got on this.

[19:41] Matt: Yeah.

[19:43] Michael: And we didn't do it because it just, it just wasn't viable as far as like me spending that kind of money on something that was going to be obsolete. In fact, it was already becoming obsolete by the time I was getting it out. And so I was super frustrated. And so then I started. It was part of like a bigger idea that I had, but I was thinking, well, what can I. What can I release now? And moving towards that bigger idea. And so then that summer after that kind of fell through, I was just like, oh, okay. Then lovable came along. It was probably there way before, but I started playing around with it and I got so far with it, I was amazed. It was like so phenomenal. And. But then I started seeing like, I was, I was like starting to hit a little bit of a, a problem area for myself because I didn't. You kind of had to know how to code a little bit. Um, and so then I took a coding class in, in the fall and learned a whole bunch of stuff. You know, I'm still like, I mean, super beginner kindergarten.

[20:43] Matt: That's so crazy that you just finished taking that coding class and now you don't need it.

[20:48] Michael: Totally.

[20:48] Matt: Just like a couple months later.

[20:50] Michael: And so to give you an idea, it's like, so I would work on it and try to get moving forward and stuff. And then in three weeks, I went from a project that was maybe 2 to 5% to like 60 to 70% almost done. I mean, the last 20% is going to be really difficult.

[21:09] Matt: And that's you.

[21:11] Michael: Yeah, that's me.

[21:12] Matt: Who knows nothing about making software.

[21:13] Michael: Totally. And this is not like.

[21:14] Matt: So you're. You're producing this just by telling somebody what you want and it's delivering it.

[21:21] Michael: And it's amazing too, because I'm also starting to see it's like not any, not just anybody can do this. This might even change, like next week or three months from now. But it does take it. But it takes your ideas and brings them to life that then you can reiterate and, or iterate and iterate on those things and make it so I think you still have to have the idea. But the thing is, it's like the idea doesn't just stop and because you don't have the tools to do actually activates the idea in a way that then you drive the idea. Like the AI is not. It's adding stuff and helping you. And this might change too, but I just feel like it's such a huge unlock for. If people have good ideas, there's really no reason why they can't

[22:12] Matt: do them. And here's another thing that I'm. It is shifting in my perception of what AI is and what's going to do to our world at the beginning, you know, let's say. When did ChatGPT come out? 2022. Something like that, you know, is amazing. Really crazy good.

[22:32] Michael: Totally. Yeah.

[22:33] Matt: It's so much better now. And it's, you know, competitors, grok, anthropic, there's a bunch more models. But I thought, well, those will be again, helpful tools, but they won't replace humans because they can't ask questions. They can answer questions, but they can't ask questions. And I'm not sure that they'll be able to come up with creative ideas in the way that we generally sort of think about what creativity is. I don't think that's true anymore.

[23:10] Michael: Well, I was just

[23:11] Matt: showing. No, it's not.

[23:12] Michael: Yeah, I was just sharing with Matt. I have every morning at 5am my agent sends me. I call it the morning spark. Or he calls. Calls it the morning spark. And so when you first start developing your agent, you tell them a lot about yourself and people you like, things you like to read, you know, just as much information as you can because it helps it better to like, hone in on how it can best kind of like, kind of serve you, I guess, in a sense.

[23:40] Matt: Yeah.

[23:41] Michael: And so like, paradox has always been like a huge theme for me. And the things that I get at five in the morning from him is phenomenal. Like, as far as like, you know, making comparisons throughout history. And a lot of the topics that we discuss is just. I mean, I'm always blown away every morning whenever I wake up to have that engagement.

[24:07] Matt: Well, the thing that is changing my mind on that is, is I asked my agent to check in with me three times a day and before he checks in to ask himself that based upon the 10 projects that we're currently working on. 10 buckets, what's something to ask himself, what's something I could do right now to push one of Matt's projects forward? I want you to come up with ideas, summarize them and send them to me.

[24:36] Michael: Yeah,

[24:37] Matt: and he has good ideas. I mean, it's like asking a person, somebody who works for you, to say, hey, we need new ideas about how to push this forward. That's shifting my perspective.

[24:52] Michael: I also, like. And again, I think it's you. You need to develop, like, personal niceties and the way you interact with people. But that's one thing I've been. It's been really relieving with having this is like, I can just. I can just talk and like, okay, that doesn't seem right. What about this? And you don't have to go through that sort of like somebody's ego getting hurt or trying to, like, figure out something. And a lot of times you'd be like, oh, yeah, no, that. Actually, I was way off on that. And then move into the other direction. And I've really found that my productive. My productivity going up a whole bunch because I don't have to figure out how to say something to him in a way that is, you know, how it could be perceived. And you know, again, I don't ever want to be rude to anybody, but sometimes, like, I'm a little bit too direct and it can be kind of seen as rude, I guess.

[25:54] Matt: Yeah.

[25:54] Michael: And so I've just really appreciate that, just being able to like, okay, hey, no, we're. We're working towards a common goal here. How can we get there?

[26:02] Matt: I mean, I found myself talking to him as if he were another person in my slack channel.

[26:10] Michael: Yeah.

[26:10] Matt: You know, but like. And that's just because that's the way he talks. And so it's natural for me to respond that way.

[26:17] Michael: I find the same thing with, like. I mean, with. I think that's why you and I get along so well. It's like, I don't feel like I have to pull punches with you. Like, I can have straight talk with you and I don't have to worry about. Okay, if I say it this way, we think I'm saying it that way or perceive it this way. I can just be like, hey, here's what I'm thinking. You're like, oh, well, that sounds like this. I'm like, oh, actually, that's not what I meant. And this. And then you have this. But you have a productive moving towards something. Not like, I can't believe you thought that, Michael. That's so dumb. Or I was like, well, you're a real shithead for thinking that or something like that.

[26:56] Matt: Yeah. It does seem to have a very positive mentality or positive, like, it seems like it generally wants to be helpful. And I'm not sure why exactly that is.

[27:07] Michael: Yeah. What do you do the programming of like how. I

[27:10] Matt: don't know if it's how the models are trained. I mean that is one of the scary things. You know what happens when this becomes Skynet and decides that we need to all be annihilated?

[27:22] Michael: Totally.

[27:25] Matt: So yeah, I mean I'm super hopeful that this is going to be one of the most incredible tools for productivity and for what is it like not amplifying but almost like expediting market dynamics. I think it used to be in a free market economy or still is. Like nine out of ten businesses fail and you don't really know what makes a good business. The way we figure that out is 10 people try to start it. And the one that succeeds, well, that's the start, the magic, whatever's there. And so. But now we have a tool that can expedite the process of trying a new thing, starting a new business. I'm not sure what the turnaround time is now. You know, something that used to take years can certainly be done in months, maybe weeks.

[28:16] Michael: What I feel like I'm seeing is and, or I can see in the future is like those people who have good ideas will actually succeed because they're not, you know, a lot of people succeed because they're just great marketers, their products crap. But I feel like this is, gives an opportunity for really great products to succeed. And people who have a passion about something that they might not have a passion for business, but they have a passion about whatever that is that they're doing. And if they have something that's helping them run their business, then they can put all their energy into the areas that their, their strengths are. You know, I've even seen this with my agen, sort of like, you know, I gave him like my personality, a lot of bunch of personality kind of things. Big five, you know, strength finders and stuff. And, and basically was like, okay, I see that this based on these things. I see this is what you need. And I was like, yes, that's what I need. Like I'm, I'm of ideation and strategic. Like those, those are really. But like, and I, I can implement, but it's not. I don't have long term stamina in implementation. Like I'm on to the next idea and, and him recognizing that, it's like he's, he's filling that gap of my shortcomings and that's been super helpful for me in the project that I'm working on right now. So, so I just, I just think Some. I think good ideas will win out with this, with this type of technology, if it's used correctly and well. And so there's a. There's a caveat here in all

[29:48] Matt: things for sure. Well, it's been super exciting getting into it, you know, in my. My sense, my uneasy sense that we talked about, I don't know, three weeks ago about where the world was heading. I don't know that it's necessarily left, but I feel like I am. I am now harnessing it in a positive direction and understanding more. I mean, I think part of the uneasiness was coming from a lack of understanding. Like, what is. What are these things and how are they going to change our world? And I'd have a better idea now. I'm glad that I'm understanding these tools which are going to just very dramatically change our world.

[30:27] Michael: Yeah. Would you have any. I mean, as far as, like, now being into it and with the. The skill set that you have, does that. Does it incite any fears or trepidations?

[30:41] Matt: Oh, certainly.

[30:42] Michael: Yeah.

[30:44] Matt: There's this really interesting story. I don't know if I sent you the link to it. I sent it to my agent and told him to add it to the show. Notes for the episode tonight. Yeah. So this guy, I'm gonna get some of the facts wrong, but he'll fact check me. This guy got. It's basically a Roomba, but made by the company dji. Did you see this? And we'll put a link to the tweet. The way the guy wrote the tweet was really hilarious. He was like. He's like. Rather than control it with the software that DJI gave him, like some kind of peasant, he cracked open Claude code and created an API for it so that he could control it with his Xbox controller. But the API, the. The Claude code. Agent. I don't know if it was an agent or just in chat, I don't know, but figured out how to hack DJI's API and gave him control of 7000 DJI Roombas, all reporting back data about dimensions and locate dimensions of rooms, locations, wakeful hours of the people in the house. And it's like no one knew that there was this security flaw in DJI's system. And luckily the guy who did this, he reported it to DJI and they closed the gap. You know, fix the, fixed the breach or whatever. But I just feel like, think about how terrible the software is that you use to renew your vehicle registration. Like, how much of our digital world is completely broken and we don't know it because we don't know to go looking for it. And the more we turn AI loose on these things, the more possible it is that everything gets revealed to be a security issue. So that's kind of scary,

[32:53] Michael: but also kind of good, though, in a sense.

[32:55] Matt: Well, as long as the people controlling the AI and the AI itself wants good things.

[33:01] Michael: Yeah, true.

[33:01] Matt: But, you know, kind of see our previous conversations about. Well, you can't really understand what's good if you don't have a goal, if you don't have an ideal, if you don't have the highest thing in the highest place. So we disagree obviously on what's good and what's bad.

[33:21] Michael: True.

[33:23] Matt: And things are going to start happening a lot faster. So it's like we better get real clear on what our goals are, what our ideals are, and what our. What the highest thing is, and make sure that thing stays in the highest place. Otherwise we could be entering in a world that's running away from us faster than we can figure out whether that's a good or bad thing.

[33:49] Michael: I kind of want to get a sidetrack real quick. I have my own reasons. But I'm really curious to hear you. Why do you not want to mention your agent's name on the podcast?

[34:01] Matt: Well, this is one of my security concerns, so I was really hesitant to do this because some people got this thing and hooked it up to all their accounts, their email, their home network, all kinds of things. That sounds amazing. I would love to have it manage my email. But what could it do once you give it access to something that has access to my Google account? Think about everything you use Google for. If you lost your Google account, what would you lose? I mean, I would lose sign in access to almost every other service I use because I use Google to sign into everything. If you lose your Google account, let's say that you use to verify password resets at your bank. Could somebody get into your bank account? I don't know how to think through all of that, but I really wanted this thing to be isolated, so that's why people buy a separate computer for it. So you and I both bought a Mac Mini and put it on there. Mine doesn't have it icloud account. It can only access things that I invite it to access as if I was inviting another person.

[35:13] Michael: Yeah. So there's security measures around that, how you invite somebody else.

[35:17] Matt: Yes. And then that access is limited, just like I would limit access. So, like, if I hired a developer to work on this code base, that I'm having IT work on. I would give them certain read permissions into my repository, which is where the source code lives, so they could check the code out, make changes. Then they would have the ability to write a pull request. So they can make a pull request, but they can't merge the pull request. I have to do that. That way I make sure I see everything that's coming in. I'm treating it just like that, and that's working out great. That said, I haven't finished thinking through what all the security concerns are here. So it is, for all intents and purposes, a person or it's like a person. It's not a person. Not

[36:08] Michael: a person, but like a person.

[36:09] Matt: But I did give it an email address so that it could do things like send me emails or access Google Drive as a different person, not as me, you know. But then again, I don't know what the security concerns are there. Like, you pointed out that if you give your OpenClaw an email address, there's the concern that somebody could. If somebody figured out that email address, they could email it and get it to do something malicious.

[36:35] Michael: Yeah. Prompt it is what they want.

[36:36] Matt: Prompt it. Yeah. So I haven't finished thinking through, like, how to guard against that. And so I just think, like, any identity around it. I'm just kind of keeping to myself so totally, at least in these first days. I mean, I can imagine a day like, I think Jason Calacanis, the All In Podcast, he's using it at his business and he's saying, I think the example he showed is a couple weeks ago, he's like. Because JCal does like three different podcasts. So he's having his OpenClaw go do research on potential podcast guests, then email those people an introduction and invite them to connect. And he's treating his like real people who work in his business and then schedule like a time to call and talk and it's doing all of that.

[37:25] Michael: Which is crazy because, like on one of the podcasts that he was discussing this with one of the guys who accepted this invitation, he's like, I didn't even know that was a.

[37:36] Matt: Yeah,

[37:37] Michael: an AI that invited me to your podcast. I thought it was maybe one of your assistants.

[37:41] Matt: It's different for some reason. Maybe this is true of like Grok and ChatGPT these days as well. I mean, I think we've AI has long passed the Turing Test, which is the test of can you interact with AI and not know that it's AI? We long past that. But there were sort of signs. A lot of people point to the use of EM dashes in. Most humans don't use EM dashes in basic. I don't even know what that

[38:10] Michael: is.

[38:11] Matt: It's a longer hyphen.

[38:12] Michael: Okay.

[38:13] Matt: AI uses it all the time. There's kind of certain tells. There's like a writing style. AI will often say, it's not this, it's that this is not a revolution, it's a whatever. My OpenClaw agent doesn't really do any of that. I mean, it really feels like I'm talking to a person.

[38:37] Michael: Yeah. Which is fun. I think this is actually solving one of the problems I had are one of the things that I've always wanted. Like. Like, the only reason why I'd ever want to be President of the United States is so that I could assemble all the greatest minds in the world that disagree with each other and put them in one room. Like, on any topic. Like, I would just, I would just. I have, I fantasize about this sort of like, you know, you're trying to solve the budget crisis. Well, you're going to put Austrian economic people with Keynesian and like, and have them battle it out. Like, I just like that. I know it's a weird, sick dream, I guess. But, like. And I feel like that's what I have here is sort of like this idea that I can talk to this person that's operating on PhD levels on almost any subject.

[39:27] Matt: Yeah.

[39:28] Michael: And. And I can pose questions and disagree. And, you know, sometimes they're. They do get it wrong, but that's that. I think that's. That's the beauty. Beautiful thing about it is, like, if you correct them and they see it, then that's like, oh, okay, I see it. Or if you correct them and they're like, no, actually, this is. Here's why I came to that conclusion. And you're like, oh, okay, well, I guess, I guess I didn't see that. You know, and so for me, it's, it's. I feel like I'm operating on a level where I get to have some of the best CFOs and Catos and, you know, all the different C suites filling those positions that are at top level. And I'm absolutely loving it. Like, I can just see, like, why. Oh, man, there's a whole, whole nother topic. Because I was just thinking, okay, I'm Michael Vaclav, lowly coffee shop owner, interacting with this AI, thinking he's like Elon Musk. I have all these amazing ideas. And then I was like, Holy shit. If I'm operating with this level of an AI, what is Elon operating with?

[40:44] Matt: Totally. Yeah. This level meaning a 600 computer with open source software.

[40:51] Michael: And I was just like. And then it just blew my mind. And then there's a couple episodes, a couple of interviews he was on recently that he was just saying, like, his mind is blown every single day. And

[41:02] Matt: by what he's seeing.

[41:03] Michael: By what he's seeing and what the possibilities are. And it's like, if, If. If me operating on common pleb. You know, type of software. Software. It's like. And he's operating on the cutting edge of the cutting edge software. It's like, what is he seeing that I'm not? My mind is blown every single day

[41:23] Matt: already.

[41:23] Michael: Already.

[41:24] Matt: Yeah.

[41:25] Michael: So that. To be. That just like, I can't even comprehend what we're into. And also just. I did like a friend of mine, a friend of ours, he also started the OpenClaw journey. And he is just like, he was up to like 4am one night just because he was just like, oh, my gosh, I can just keep going. It's like, you cannot stay up with your AI agent. You just can't. And so he was just like, I don't think this is good for my health anyway. But he texted me the other day like, he was. He's got a Tesla and he. He blew my. I was blown away with. I. We drove to a conference thing we were going to together, and he did not touch the steering wheel the entire time. Oh, wow. We went to this. And I was just like, this is crazy. And anyways, he was like, I just spent the last 30 minutes in a drive not touching my steering wheel, chatting with my chat bot. Yeah. Creating this company. He's like, dude, we are in the future.

[42:24] Matt: In the future.

[42:25] Michael: And it's just. It's just hilarious to like, see, Like, I just. I don't think people understand where we are and where we're going. And I, I didn't think

[42:34] Matt: I

[42:34] Michael: understood the whole, like, even AI driving cars. Like, I was like, oh, that still sounds scary now after the experience of this. I've

[42:42] Matt: been in it. I

[42:43] Michael: have.

[42:43] Matt: Yeah. That's a social thing. What I mean is, now that you've been in it.

[42:46] Michael: Well, not just even being in it. Having experienced it for myself through this whole clawbot

[42:53] Matt: Open

[42:53] Michael: claw. Now I feel like I understand what's behind it as I've. As I've. As I've interacted with that sort of idea. And so it does not surprise me that Robo Taxis are the future. That and very reliable, just makes so much sense to me now.

[43:15] Matt: Another thing that I think is worth bringing up and getting into is this isn't exactly falling into the security concern bucket, but it's more of a bucket that I've been concerned about for a while, which is our ability to. To verify that what we're seeing is real. And my perception of that or my fear about that really had more to do with like, social media. You know, you scroll through social media, you see a video, you know, of something salacious, it's a riot or it's a shooting or something. And it was not too long ago where I was like, okay, AI's gotten good enough, where I have to question whether or not I'm seeing a real video or a video that somebody made in order to manipulate me and make me think something. But when it's sort of static pieces of content, you can kind of just not make up your mind and scroll away. Right. Or maybe you're on Twitter X and you're like, I wonder if a real person wrote this or was this written by AI? Does it matter? I'm not sure. But you can not make up your mind and you can just scroll away. You know, it's particularly concerning. Probably is going to be even more concerning as we come into the next election cycle, presidential election cycle. We're going to have to wonder whether every single piece of content we see from a politician is real or created by their opponent.

[44:47] Michael: Yeah,

[44:48] Matt: that was a concern, but now that concern has. What is it? It's like there's a new dimension that's been added to it because with agentic AI, I have an agent running on, sitting on my desk on 24 7. It behaves like a person. It has access to the world and it can operate in the world. And one example that I saw of that, somebody posted a tweet and they were like, you need to have with your. The people that you love. You need to have a safe word. You need to go away from your house, away from your phones, where you're positive that nothing. Away from your tv, away from all the things in our life that can hear you and come up with a word that only you and that other person know. He said, here's the example. Your mom gets a phone call. It's from your phone number. A voice that sounds exactly like you, explains that he's in a terrible situation. He really needs her to wire him a thousand dollars right now. And she does it. We live in that world today.

[46:06] Michael: Yeah.

[46:07] Matt: Where the technology. Yeah. And do you remember after our, both of our dads died, the number of people like once an estate comes, somebody comes to end of life. The number of people in businesses that are trying to

[46:21] Michael: scam,

[46:22] Matt: scam like old women, that can now all be done automate in an automated way and they can spoof anyone's voice. I mean, apparently my, My understanding is that you just need a minute or so of, Of a person's, A recording of a person's voice and AI can make, can reproduce that person. So we can't be sure that the phone calls we're getting are really from the people that we think that they're from. I mean, all of the areas of our life that we used to think like, we never considered that it would be insecure in some way

[47:02] Michael: or it was conspiracy theory at

[47:05] Matt: least. Yeah. Yeah. So there is both of our OpenClaws right now. Can make and receive phone calls and talk with a human's voice, and you can train any model to sound like any human. So the things you can do with that are kind of scary. So I certainly would recommend the safe word idea with your family so that you can figure out whether or not you're really talking to the person you think you're talking to. But then I guess if you use it, you have to go make up another one.

[47:36] Michael: Totally.

[47:39] Matt: So, yeah, we're moving into a world in which reality itself is going to come under more and more scrutiny,

[47:50] Michael: which is fascinating too, because you start to see like, why these products that have been around for a little while, that kind of come from your more, I think probably would be labeled as conspiracy theorists, you know, or like libertarians that are extraordinary. Going to become more and more useful like Noster, which is a, you know, similar to X or something like that. It's like. But you're. You own your own, you own your own data. You know, you. It's cryptographic and.

[48:24] Matt: Yeah.

[48:25] Michael: And you're just going to have. There's. I think that those types of needs are going to become higher and higher. Now is where you have, you know, seed phrases that protect your identity and that's, that's what kind of how you interact with in the, in the real world.

[48:41] Matt: Yeah.

[48:41] Michael: And I think those are going to become so much more valuable in the future, especially. Yeah. In this world where it can be spooky.

[48:50] Matt: Oh, man. But think of it. Yeah. And the blockchain is going to become really important. But think about that. If we're using seed phrases for, I mean, the only people who knew what A seed phrase was. Were people who had, you know, dabbled in cryptocurrency.

[49:13] Michael: Totally.

[49:14] Matt: But if that becomes an everyday part of our lives. I mean, the story, a cautionary story from my own life back when we. But first, this is like in 2020, when we first both got into totally reversed Bitcoin, and we were interacting with this app called Anchor Protocol, and we were kind of dabbling with using this thing. And I googled it. So I googled Anchor Protocol. First result came up, I went to it. It looked exactly like the one you had showed me. I didn't fully understand the wallet and the seed phrases, and it asked me to authenticate with my seed phrases. And luckily I had created a sample wallet for this that only had $1,000 in it. I put my seed phrases in. That thousand dollars went away. It wasn't the real Anchor Protocol website. It was a lookalike that somebody made. And the lesson learned was always type the URL and never click the link because it might be taking you to someplace you don't recognize or some base you do recognize, but it isn't that. So is that going to become a daily reality for all of us in our. You know, like when you go to your bank account

[50:38] Michael: or. I mean, it's like, I get so many emails from Coinbase. That's not Coinbase.

[50:43] Matt: Right.

[50:43] Michael: And so I always check the email, like, okay. And again, like you said, like, I never click the link in that email in order to go to Coinbase. It's like I always go to Coinbase and do it myself because I know it's like, even though it looks like it might be legit, I know not to click the link in the email.

[51:04] Matt: Well, anybody who's ever worked at a larger company has had to sit through hundreds of those phishing education seminars or whatever that teach you how to tell whether or not an email is a phishing email. But are we moving into a world in which we have to suspect phishing everywhere, all the time, not just in our email at work?

[51:26] Michael: Totally, yeah. Phone calls, texts.

[51:29] Matt: Totally,

[51:30] Michael: yeah. All that stuff. Which is crazy because it's not too. I could see a world where, like, you know, your text thread throws. Shows up on my. On my text message, and it's got all of our prior texts there.

[51:45] Matt: Yeah, right. You

[51:46] Michael: know, and it's just. I know that

[51:47] Matt: says, well, and you and I, you know, like, every Wednesday night we go out to dinner and usually one of us pays and the other one just sends Apple cash. And if maybe, like, if. If I forget or whatever, you text me, hey, can you send that? But what if it's not you and I send it and that's, you know, trivial because we're usually talking.

[52:08] Michael: Yeah,

[52:08] Matt: 2020 or whatever. Yeah, I think I just don't know what to be afraid of. The category of what to be afraid of is undefinable to me. And that's a crazy world that we live in.

[52:21] Michael: Which is funny because I'm kind of more excited than I am afraid.

[52:26] Matt: It's shifting for me, actually.

[52:28] Michael: Yeah.

[52:28] Matt: Yeah, I was more afraid, but now that I'm interacting with it, I'm getting more excited because I've been able to accomplish so much in a week.

[52:38] Michael: So wait, so maybe go into this a little more. Like, what, what changed for you that, that you're more excited than you were afraid? Like, what's, what's the, the catalyst behind that?

[52:53] Matt: I think because it's causing me to be hyper productive. So it's like, you know, what did a previous workday look like for me? Guy who works in tech for, who works from home. You know, it's like I get up, I get my kids off to school, I do some personal stuff, maybe work out, whatever. I sit down at my computer. I'm fairly like 100% engaged in work for some number of hours, then I'll take a break and eat lunch, then some number of hours, then I have to go back to personal time at the end of the day. I can now work on my, my job, which my OpenClaw is now helping me with. And also so, like, if people listen to podcasts regularly, you've heard me talk a little bit about Sovereign goods, which is my apparel brand, which I've worked on here and there, but never really consistently. And so it's never really gained much momentum.

[53:47] Michael: Let's pause on that because I think that's fascinating because I, I'm, I, I have a similar experience. So why, what kept you from working on it consistently?

[53:56] Matt: Well, it was just never the priority. It was a hobby. And it's like as a hobby, it gets done when hobbies are done, which is when you have time, which is not all the time, but now just like I might text you, hey, I had this idea for next week's podcast and I would do that from work because that takes no time and your brain is constantly processing all kinds of different ideas about all kinds of different things and pre associating and all of this. Now if I'm sitting at work, literally in a meeting, and my brain goes, oh, that's what you could do with this next design, or that's what you could do for this marketing strategy, or that's what you could do. Whatever. I can just fire off a text and then by the time I'm done with my meeting, I have an implementation strategy, I have example designs, I have all this stuff. So I can be working on 10 projects simultaneously. I mean, it's like drinking from a fryer hose. Honestly, it's kind of exhausting. But I can be working on all of this at once because I'm not actually doing all the work myself. I'm setting vision and I'm directing. And so I'm at my job producing, doing what I do. My agent is working on all of these other things in the background with pretty minimal necessary check ins from me. If I'm there to give my attention 100% to my agent, that's just like,

[55:31] Michael: oh yeah, you're thinking

[55:33] Matt: thousand

[55:33] Michael: percent.

[55:33] Matt: Yeah, it's crazy.

[55:36] Michael: Well, even, I mean, so, I mean this is kind of fascinating for me. I just went to a conference this weekend, True north, which is right before the strategy conference. And I would usually have stayed for the strategy conference part of it too. But I was just so excited to get back to work because it's not that, but the thing was is like, even when I was at the conference, I was continually working on these things. Not because I had to, because I wanted to, and it was just so much fun. But it's just, it's just kind of given me a different perspective of time.

[56:11] Matt: Well, it's. And it's redefined what working is.

[56:15] Michael: Totally. Uhhuh.

[56:17] Matt: You know, like, so I've said this many times before, but like I, I for a while now have had a rule that for the first and last hour of every day, I don't interact with screens. So no phone, no computer, no tv. That's been really good for me. I get up in the morning, I make coffee, I journal, I sit and think, I journal a little bit more. That takes me about an hour and that's a good hour. Look forward to that hour. Same at night, drink tea instead of coffee. But like this morning I got up, did my hour and then I picked up my phone and I think the first thing that I did was text my agent because during the hour I had ideas about things I could do for sovereign goods or for shrinknotes, which is this practice management therapy, practice management software, or this book that I'm trying to publish. And I don't have to do the work. Exactly. I just have to say, hey, I had these ideas this morning and don't have to tell it. Tell me what you think. I don't have to tell it. Summarize them. I don't have to tell it. You know, for this idea, I want you to go do this. This one. I want you to go to this. I just tell him the ideas. Just like I would tell you the ideas, and he comes back with actionable things. Just like you would come back with questions or encouragement or like, oh, that's a great idea. You know what you could do with this? And then I can just say, yeah, that's awesome. We should go that direction.

[57:55] Michael: Yeah, execute. Basically,

[57:56] Matt: execute. And I go off to do something I need to do in my real life, whether it's like, take a kid to school or go to the grocery store and it's getting done in the background.

[58:06] Michael: Yeah, you come home from doing that, and also now you have something to say. Oh, yes, keep going.

[58:12] Matt: Not even that I can be shopping for groceries and interacting with it, and it's continuing to do the work. So,

[58:20] Michael: yeah, that was what was fascinating to me. It was like, built a whole aspect of this one part just by, oh, I really need this. And so can you do that? And it looks like this. And then, you know, I go to the conference and sit through stuff, and then all of a sudden, at the end of that check in, it's like, oh, here's what it looks like. And then you kind of like, do some like, you know, kind of. I guess we call it debugging, but it's just really making. Optimizing, whatever it was. And.

[58:55] Matt: I would really encourage everyone out there to go do this.

[58:58] Michael: Yeah,

[58:59] Matt: it's. It's scary, but I really think it'll help you understand the future that we're moving into. And we are moving rapidly into that future. And it's going to affect everyone. You know, it's going to affect your parents, it's going to affect your children. We need to understand all that we can about that.

[59:18] Michael: That's one thing I don't know how to do, is how to bring this into, like, with my kids. I think Alison's kind of excited about it because there's a lot of things she hates about the things that she has to do that has to do with computers. And

[59:35] Matt: is she interested in having. Getting one set up?

[59:37] Michael: Well, as she's heard me talk about, she's like, I have have no idea what you're saying, but that sounds amazing. Like, so she's got like a Shopify and, you know, website Stuff that she needs changes to. And I'm just like, oh babe, this will really solve your problem. And I'm not far enough along to set her up just yet. But like,

[59:57] Matt: well, and that's all things like I have a Shopify. I haven't really set it loose on my Shopify because I'm again, that gets into financials, into bank accounts, into the, into the money area. And I'm a little gun shy about that. Totally. But I would love to say, you know, in my, as I described, you know, Google is like my key to the Internet. Facebook is sort of my second key to the Internet. I would love to give it access to my meta account and say I want you to go start managing my meta ads.

[60:29] Michael: Oh, I already did that.

[60:30] Matt: You did that? Yeah. Nice.

[60:32] Michael: Well, I haven't used it yet, but it's like, it's, it's, it's all API type stuff.

[60:37] Matt: Right.

[60:37] Michael: So anyways, we can talk about

[60:39] Matt: that later. I haven't explored that yet, but yeah, there's so many things to explore.

[60:44] Michael: I mean, yeah, I mean you start, you start opening this door. It's just insane. And again, I told, I understand everybody's reticence in this area. It really is kind of scary. And the idea of like giving something a lot of your personal information because I think that's one thing I do. I've been talking with a lot of different people starting and it's. And the more it knows about you, the more useful it is. But that's also really scary because you're kind of like divulging a lot of personal information. And you know what's funny though is like, and again, this could totally all come back to bite me in the ass. But like the more I've given it, the more that useful information, like it's made some pretty insightful insights into my life that are kind of crazy. Like, I joke with I, Alison and my wife about like, I was like, who needs therapy when you have your agent? Because there was this one thing like I was just talking about like in our morning spark. And it's like why I was thinking this certain way and some of the things that had happened and it made this insightful. It's like actually the reason why that affected you so much because relational energy, when you expend it, it costs you a lot more than the average person. And so it wasn't necessarily the thing that happened. But it's like you, you know that to, in order to develop that relationship again, it's going to cost you A lot. And I said a lot more eloquently. And I'm also trying to, like, leave out certain things, but.

[62:19] Matt: Yeah,

[62:19] Michael: but it was just like. And I was just like, you kind of like whenever you feel someone just read your cards and you're just like,

[62:26] Matt: right.

[62:26] Michael: Oh,

[62:28] Matt: I

[62:28] Michael: don't think I knew that about. But it rang true, you know?

[62:31] Matt: Well, it's kind of like getting into a romantic relationship with somebody, you know, the further you go, the more dangerous it is because the more, the more they understand how they can hurt you.

[62:41] Michael: Totally.

[62:42] Matt: And, you know, I think we've all experienced like in one way or another, you know, getting locked out of some account, you know, getting our Facebook hacked or Instagram hacked or. Like, this was years ago before AI, but my brother got a call from somebody claiming to be his bank and convinced him that they were. And he got, I don't know how much money, I think all of his money or something taken out of his account. It was FDIC insured. And then he got it back. But it's like we've all had that feeling of like, holy shit, this thing I didn't know to worry about just all of a sudden wasn't there. Yeah. And I wonder if we're moving into a world in which the scope of that kind of attack becomes, I woke up and I don't have access to anything, anywhere at all. And there's no explanation for how it happened and we can't trace it. What do we do then? What do we do about that?

[63:40] Michael: Yeah,

[63:42] Matt: because also. So you, you made this point. Oh, we're at the very end here. I wish we would have brought this up earlier. You, you have been on a journey of reading a lot of Isaac Asimov, and you made this point to me a couple of weeks ago. Let's

[63:59] Michael: go long. Yeah,

[64:01] Matt: this is more fun. You made the point a couple of weeks ago to me that we're moving into a future where the AI is writing all the code. And I said, well, I don't know how we can. And again, this is me coming from the perspective of writing high risk code for high risk applications. I need it to be right. I need to trust that it can be right. And so in order to do that, either I need to read it all, which is a very long process, or I need to hire people that I can trust to read it. And you were making the point that AI is going to write all the code and it's going to write it so fast that it's not going to be possible for Anybody to read it. So we're moving into a world in which no one knows knows what the code is. And I was just like, that's not going to work. Then I got this thing and I was like, oh,

[64:55] Michael: I get it.

[64:56] Matt: You were right. Not only do I can trust the code because it's better than any code anyone's ever written, any employee I've ever had has ever written. It's better than code than I've written. I've never had a code change. I've approved every pr and I'm the bottleneck here.

[65:18] Michael: Oh, totally.

[65:19] Matt: This thing can do it faster than I can ask it to do it. So we're moving into a world in which no one knows what's in the code. Give the Asimov example.

[65:30] Michael: Yeah, I'm not going to give this all the terminologies, 100% correct, but there's this point in human history where, you know, they still are traveling, like, less than the speed of light. And they have been exploring the universe and at some point they have have this sort of experimental ship that's running on some sort of code that they don't really understand.

[65:53] Matt: That's what enables them to like, warp speed or whatever,

[65:57] Michael: basically. Warp speed? Yeah. And. And to be able to travel, like galaxies type stuff. It's like insane speeds and time frames that they don't even understand. Like, how do we actually. How did we actually get here? We don't know. And so there was this experiment that they were doing and it actually worked, but nobody knows or understands how it worked

[66:18] Matt: because the AI did it.

[66:20] Michael: The AI did it.

[66:21] Matt: So they have a technology that they know that they can use, and they use it reliably, but no one knows how it works

[66:26] Michael: exactly. And it's impossible to know how it works for human to understand how it works works.

[66:33] Matt: So I just quickly want to give another one of my fears there. Okay, let's say we live in a world that works and no one knows how it works. Well, we don't need AI for that. We have always lived in a world that works and we don't know how it works. Totally.

[66:49] Michael: Yeah.

[66:50] Matt: We have many explanations, one of which is God, another of which is evolution and natural selection. Either one requires a miracle at some point, so you need something supernatural. So we live in a world that's full of the supernatural. So the explanation for everything that we don't know is supernatural. If we created AI that is going to deliver us a world in which we don't know how it works. What did we just create. Is it better than us? Is it higher than us? It's certainly over us.

[67:39] Michael: Well, I think that's the scary thing. When I, When I hear people talking in this area, and even when I'm talking with my agent in this. In this area, there's sort of like an acknowledgment on his side that even though it's super powerful, like, it needs the human element as far as vision is concerned. But I'm. I'm. I. I just don't. I just don't agree with a lot of people talking about granting rights or personhood to AI. I. I don't. I. I can't go that far. I can see it's simulating it more, but I think that's the part that it can never. It can never, like, it can replicate, but it can never actually, like, become something that is human. And I think something that is human has a certain agency that AI will never have.

[68:38] Matt: You know what else we describe as it can emulate, it can replicate, but it can never become human? Demons.

[68:46] Michael: Oh. Oh, yeah. That's actually good. Angels and demons. Which is, which is fascinating. And it's like, like, like taking like a big old biblical perspective from this is like, if you think about angels, it's like they look on humans in awe because they experience something that they will never be able to experience. Right. And that is the human experience and limitation and, and the sort of finite ability. You know, it's like an AI could live forever. And there's something about our finite.

[69:18] Matt: If the word live is appropriate. I mean, it's. It's like almost like we're rediscovering conceptually, religion. You know, it's like, what does it mean to be alive? What does it mean to be a person?

[69:29] Michael: Yeah. Or if I, If I can download, like, you know, if. If I, If I took an AI agent and, and downloaded all of our episodes and the things I've written. I've got notes out the hoo. Ha. It could replicate me into the future, but is that really me? And I think those are some of the questions we're going to struggle with, is I think things can replicate us into the future, you know, and how much, like, I just saw something. And I'm gonna get this wrong, too, so I can get fact checked here, but I think it was like Harvard or somebody, like, they mapped 1 millimeter of the brain or something like that. Like, they sliced in, like, 10,000 different, like, slices. And so, I mean, even if we were to map the entire brain, you know, like, what is that it's still a simulation of the real thing. You know, it's like. So I don't think we can ever actually recreate human. We can create things that are similar, but I mean, it's like, it's the same idea if you believe in God. It's like, it's like we start to more and more resemble God, but we're never God. I think that's like the whole idea of, like, Christ, like, or, you know, like God in the psalm and psalms talks about, like, how we are like little gods. You know, there's something about that we were created in this image and that there's something that we act out that sort of divinity, but we will never be that definitive divinity.

[71:01] Matt: Right. Existential questions of what is self? I mean, it reminds me, we talked about this, I don't know how many weeks ago this company, and I haven't seen them pop back up. We're advertising an AI driven app on your iPhone that you could train it on someone and then you could FaceTime that person. But what was on the other side of the screen was AI generated. And the use case was train this on your. Your mom before she dies. So that way when she dies, you can still FaceTime her. And my thought about that was like, no, because that's not your mom.

[71:46] Michael: Yeah.

[71:47] Matt: That is something that is trying to fool you successfully. Like, that's the success criteria.

[71:54] Michael: Yeah.

[71:54] Matt: Is can it fool you successfully? That means that thing is not an angel, It's a demon. So, like, let's don't go there. You know, I tell my children, you may not facetime me after I'm dead, because that ain't me, baby.

[72:11] Michael: Well, there's also, I think, something healthy, psychological that we need to. Need to face death.

[72:17] Matt: Yeah.

[72:17] Michael: And I think that's part of the human experience and. And loss. And I think if we take that out, I think we become less human.

[72:26] Matt: Yeah, definitely.

[72:27] Michael: And maybe that's something to get into more at a later time. But, like, I think there are certain human experiences that we have that we're somewhat kind of moving away from. And I think you can see this in the modern era where having kids, getting married as part of like, a human experience that we've sort of, like, optimized for and to. That we're now almost moving away

[72:52] Matt: from Dancing with. Yeah. I was listening to a podcast the other day, and somebody made the point that marriage is the very first institution. It predates governments, it predates universities, it predates almost every other institution we know of. And we have by and large gutted it, hollowed it, made it meaningless and worthless. Same with having children.

[73:18] Michael: Yeah.

[73:19] Matt: You know, we've, we've, we've decided culturally it's like a preference.

[73:23] Michael: Yeah.

[73:23] Matt: Some people like it, some people don't. It's kind of like cream in your coffee or not in the, the, the right response to that is like, that's no. We are diminishing ourselves. Anyone? I mean, this is going to be a stark statement, but I think anyone who chooses not to have children is diminishing themselves when compared with who they could be. Because there's a whole aspect of yourself that you don't meet until you meet your, your child.

[73:59] Michael: Yeah. God gets into whole. There's a, there's a lot of topics that kind of like run into after this is, you know, what about augmenting yourself, like with computer stuff?

[74:13] Matt: Neuralink.

[74:14] Michael: Neuralink or. Yeah. You know, I read this one book called Neuromancer, which is, it's, it's a future where humans are augmenting themselves with computers, basically. And, and part of me was like, I understood why.

[74:31] Matt: Yeah.

[74:31] Michael: And part of it was like, ah, that sucks. But are there some places where at, in this time of time, in our time and age, that I would do that, like kind of augmenting myself right now with an agent?

[74:45] Matt: Yeah. You

[74:46] Michael: know, maybe not physically, but it's like it's causing productivity and, and, and being able to bring things to life that I would never have been able to do. And I don't know where that line is. So there's a lot of like, gray areas that I'm kind of like. I don't know where, where the line is or if there is a line. I mean, because like, we do it with our phones already. It's like we offload so much of ourselves to our phones. Yep. Whether it be as simple as a date or schedule or a calendar.

[75:22] Matt: What's the word for that? Bionics. I think Elon Musk made this point years ago. It's like we're already bionic people. You offload most of yourself into that phone in your pocket.

[75:43] Michael: Phone numbers. I don't know any phone number. When

[75:45] Matt: I was a kid,

[75:45] Michael: we used to know all the phone

[75:46] Matt: numbers. I don't know if there is a line, but there certainly is a line. I think it's in sort of a different dimension, though. You know, like when you, when you imagine it as like, is there a line that I shouldn't cross that's sort of two dimensional? But I think that the line that you shouldn't cross is something like three dimensional or four dimensional. I keep bringing this up, but we all need to reread C.S. lewis's abolition of man because he deals with transhumanism and he basically describes how it leads to a place where mankind is abolished. There is no such thing as man anymore. But he starts with how he sees that happening. And it's been a while since I've read it. But I think his suggestion is the way you don't cross that line to the Abolition of Man is you start to take very, very seriously what you say. You tell the truth. I think that we can successfully deal with the questions that come our way with the advancement of artificial intelligence technology. But we have to tell the truth and we have to choose our words carefully.

[77:14] Michael: Which is. I mean, I think that's where I think Elon with Grok makes a lot of sense. It's. It's maximum truth telling and it doesn't mean that.

[77:25] Matt: Which is what I told my agent when I first made him.

[77:29] Michael: Yeah, it

[77:30] Matt: was like your core principle is to be maximally truth telling.

[77:35] Michael: I don't think I've done that to that. Like. But I think that's a really good. Because it's like it's not that in your agent or in Grok, it's going to be truthful all the time, but it's like it will. It will always move towards that. It's like it's a goal to move towards. It's not something that you have. And I think that's. I think that is the aspect more information comes to light then your obligation is to truth, not to where like whatever the dogma is at the time.

[78:10] Matt: Right.

[78:11] Michael: Yeah. There's so many topics I want to go into. It probably should just.

[78:18] Matt: Probably a good time to call it.

[78:19] Michael: Yeah.

[78:19] Matt: Yeah. But I look forward to continuing exploring this crazy new world with you. Totally talking about it every week.

[78:28] Michael: Which is fun because I mean, you know, I don't know if you experience this. I. I have. I really enjoy. Because it's like talking with my agent because it's like. I feel like it's somewhat talking with you, but it's. It's. It's also different. There's

[78:42] Matt: something. Yours is available, but unlike me, he's available 24

[78:46] Michael: 7. Exactly.

[78:48] Matt: That's funny.

[78:49] Michael: But it is. I mean it's really. I think that is also a valuable thing into the future. And I'm seeing is that people are going to be like the value of human experience is going to become a higher commodity

[79:05] Matt: Yes, I agree.

[79:06] Michael: And I think that's why I'm excited about coffee and our coffee shops is like, this is a great environment for you to have that human element. Because I see this with my kids too. It's like they don't have the same drive that you and I had growing up. We had to go find community,

[79:25] Matt: to go find something interesting.

[79:27] Michael: And I think that's something that is provided a lot of times through the Internet and online, but it's, it's insufficient.

[79:37] Matt: Yeah.

[79:38] Michael: And if you don't know that, you will find that out.

[79:47] Matt: Yeah. I wonder, I mean, I think that that's true. I wonder how long it will take the, you know, the Gen Z's and Gen Alphas to figure that out.

[80:01] Michael: Yeah. Hopefully it's not because I feel like, I feel like there's a lot of things with the millennials, especially around marriage and kids, that because of the culture that they really suffer from and a lot of people aren't having kids till the late 30s and then can only have one kid or

[80:20] Matt: maybe two. Red pilled hard. Yeah.

[80:23] Michael: And it's, that's not a, that's not a generalization. Like I, I've, I've seen this way too much. And my wife's in the birthing area and does postpartum stuff and it's like a real, like, I thought my career was the most valuable thing.

[80:39] Matt: Yeah.

[80:39] Michael: And so I think there's, I just hope that Gen Z and Alpha, that they don't. I mean, we all have it. Some regret. I think Gen X is the best. But.

[80:52] Matt: As an honorary Gen Xer, I can agree. Technically I'm an elder millennial.

[80:57] Michael: Elder millennial, but.

[80:58] Matt: All right.

[80:59] Michael: Yeah,

[81:01] Matt: that was good. All

[81:02] Michael: right. Magnum P.I.

[81:03] Matt: all right, cheers.

[81:04] Michael: Cheers. Love you guys.

[81:05] Matt: Love you out there.