Episode Transcript
[00:00:00] Speaker A: Total news.
All right, now we're live.
[00:00:02] Speaker B: Excellent. Excellent. So we're back.
[00:00:07] Speaker A: Oh.
[00:00:07] Speaker B: So welcome to the Gregory Paul show. I'm Gregory.
[00:00:10] Speaker A: Hey everyone. Paul.
[00:00:12] Speaker B: And we break the latest, we break down the latest in SaaS, startups, AI. I know, the SAS apocalypse is driving me bonkers. I don't know if you've been paying attention to the, the stock market, but it's a little out of control.
For today's, today's livestream, we have a special guest who I'm super excited about. Can I call you Alex? Are we on a first name?
[00:00:36] Speaker C: Yeah.
[00:00:36] Speaker B: Basis. Yeah. So Alex is the one that I met through a coster of mine, Thread. They are an ad network for LLMs, meaning they put ads similar to what OpenAI has recently announced into other third party AI powered chatbots and been doing it for a long time. She is over there at Thread doing a whole bunch of really interesting stuff and I knew that she was a cloudbot. I'm going to call it enthusiast, someone who's doing a bunch of really interesting stuff with claudebots. That's why I thought we'd have you come on the show and answer our questions about cloudbot.
[00:01:21] Speaker C: Yeah, thank you so much for having me.
[00:01:22] Speaker B: Of course.
[00:01:23] Speaker C: Yeah, I'm definitely, definitely an automations fanatic.
[00:01:26] Speaker B: So open automation fanatic.
[00:01:28] Speaker A: Perfect.
[00:01:29] Speaker C: Whatever you want to call it.
[00:01:30] Speaker B: Well, you're in the right place.
Well, let's just. So everyone understands the like, context. Just tell us who you are a little bit about your background and how you got, you know, mixed up with crazy people like us.
[00:01:47] Speaker C: Yeah, of course. I mean, I'm Alex, I'm COO at Thread. So I do everything that's operational. All the headaches of the company go to me and then I just figure it out a bit of the fixer. And it's been an amazing, amazing growth opportunity.
My background is actually not technical at all. I did my undergrad in politics where I met Andrea who's the co founder and CEO of Thread. And yeah, we just got to working together and I later did a master's in digital humanities. So to kind of tie into the tech, but still having a humanist approach. So I like to say that, you know, I'm a humanist who codes to some extent and yeah, so openclaw, all of these tools are things that I'm, you know, implementing on, on a daily and it's just really interesting to me to just figure things out from such a deep, different lens than an engineer.
[00:02:39] Speaker A: Awesome.
[00:02:40] Speaker B: So, okay, so I saw the photo where you guys had a giant Stack of like Mac Minis or whatever. So I want to hear like from the very beginning, like you got involved and you set one up. And I've just heard a ton of stuff about this, but I wanted to hear from someone who had set one up, like, what the experience like and like what it's doing. Like, what did you configure it to do?
[00:03:01] Speaker C: Yeah, of course. I mean, a lot of people did go out and buy different Mac Minis. There's a lot of hardware you can get that doesn't necessarily cost as much. I think everyone just followed a trend that made it a bit more crazy than it is. I did something much more simple.
I saw all the craze about this and then I just literally, rather than putting openclaw, which is going to bring you to a super long video of how to set it up, I just did simple setup. Openclaw and I followed a tutorial and this guy was actually explaining that there's different ways of configuring it. And so rather than doing it on an actual device that I would just dedicate to openclaw, I did it through a utm. So I have a desktop within my desktop and I didn't. It doesn't have anything that's linked to my computer. Right. It's completely new. It's like you're configuring a whole new computer, a whole new device. And I just set it up through there. Didn't necessarily give it any of my accounts. I gave it new accounts.
So yeah, it's, you know, just like a white sheet of paper and I just contextualize it over time, but I don't give it anything personal.
[00:04:10] Speaker A: So it sounds like you have completely isolated to rest of your own personal stuff, maybe even like some of your company stuff.
Is this. Did you purposefully do this or was this part of the guide or how did you know how to do that?
[00:04:24] Speaker C: Yeah, I did it on purpose just because I've heard so many, you know, mixed opinions around the safety.
And actually when you do set it up so they have a code that you just put into your terminal and that will just set everything up for you then and there. And the first thing that it says is, you know, do you realize how risky this is? Basically? Do you know that this is a high risk and you have to say yes or no before you download it? And so I know that there's a lot of risk around this. And of course you've had stories on Reddit. I'm. I love Reddit so much and I Love that people, you know, just talk about everything and anything and just these, like, really raw opinions on it. And someone was saying that because the Clawbot had context on what this person was working towards, they actually went and bought a masterclass for five digits on the person's card. So I just thought, okay, better not to give any of my information. I don't want it to send email to emails to clients because they think that I want to close this deal or whatever it is. So I just wanted to give it absolutely no context, except for what you can find online about the company. And in that way, just having different webhooks through.
Yeah, through, like, different tools that I use. And just having it, for instance, scrape LinkedIn or scrape different news and just, you know, so that we can then have content at the speed of news, which is like Andrea's tagline. And he does that so well. And the company does that really well. So we just wanted to frame it around that. So to just have something running 24 7, but without actually having any security risks around it.
[00:06:07] Speaker A: Cool.
[00:06:07] Speaker B: Okay, so let me. Let me. Let me play. Let me repeat what I think I heard, because I'm curious about the setup in particular.
So you installed it on, like, a separate instance on your laptop, so it's like a virtual machine or something. And then when you configured it, did you go and.
[00:06:23] Speaker A: Did you.
[00:06:23] Speaker B: I'm curious about this. Do you have to go and create the accounts? Like, do you have to go and create a Gmail account for it and then give it access and then a calendar? Is that how you did it?
[00:06:33] Speaker C: Yeah, I mean, I did an email that's external, but I did that later, actually. I started by giving it my email, but through different webhooks, you can actually decide.
There are all these different companies that have webhooks. And I just decided to give it info to be able to draft things through Gmail, for instance, or through Docs.
So that was on my account. But then I heard more and more things about people having their email sent out, so I was like, okay, let me just make a whole other difference.
Email that doesn't actually have any information, just because if it does end up spiraling, I wouldn't want it to do that.
[00:07:21] Speaker B: Yeah, that's what I heard too, is people gave its own email and calendar, and then it can actually book appointments for itself and for other people. So there's actually the claudebot has its own own appointment, like, and can invite people to its meetings.
[00:07:34] Speaker C: Yeah, it's insane. I mean, even the.
The social Media platform. It has. Right. It's just creating different religions and talking about.
[00:07:42] Speaker B: Yeah, let's talk about that. So, so okay, just to clue everyone in.
Yeah, yeah. So just for people, like, just so they understand we're talking about like, so, so all these people create all these call bots, like Alex has running on a laptop or the computer or their Mac Mini, whatever. Then my understanding, and I don't know if you know any more about this, like someone, I think asked it to do this, like go and create a social network for other Claude bots. I believe it was like queued up. Right. Because can they buy the domain and stuff? Like, I guess they could, but they have a Malt book. This is so confusing to you for people that it's got three different names.
Maltbot, which they got rid of and called it claudebot, and now it's called Open Claw. But does claudebot still work as a term? We're getting off track here, but.
[00:08:29] Speaker C: Well, a lot of people just call the company itself Open Claw, but then they will call their individual bots Claw bots.
[00:08:36] Speaker B: Okay, okay, good. So the two terms do still exist. Yeah, this is like they've really made it.
Oh my God.
[00:08:45] Speaker C: You can refer it, refer to it in any way. It's just that they cannot brand themselves as.
Of course. I mean, then you have anthropic coming in.
[00:08:53] Speaker B: Right, right, right, right. Okay. Okay. So, so there's a, so, so there's a social network out there, like Reddit, and then all these bots can go there. And so my question for you then is like, does your bot go there?
[00:09:03] Speaker C: No.
[00:09:05] Speaker B: Like, no way.
[00:09:07] Speaker A: No way, no way.
[00:09:08] Speaker C: My bot. I mean, unless, you know, I, I, for now, we, we've been dealing with so much of the company, so this has really just been a, a side thing to see if it actually does help out with anything. But I want to play around with it more. But for now I've just, you know, it's, it's like a pet that I don't want to, like, leave alone in the dog park or whatever. I don't want to let go of it just yet. So I don't want it to be fully agentic.
And yeah, it's just doing what I tell it to do.
[00:09:35] Speaker B: Yeah, Your clubbot just sent me a text and said, mom won't let me go out and play. And I was like, I know, I'm trying to get her to let you
[00:09:40] Speaker C: do it, but no way, you're, you're not old enough yet.
[00:09:44] Speaker B: Exactly.
Oh my God.
[00:09:47] Speaker A: All right.
[00:09:48] Speaker B: All right, so, so, so we covered like your background. You covered it like how you set this thing up and then your, your, your over parenting style when it comes to claudebot. This is interesting, but I want to hear like, yeah, what are you doing? What have you experimented with? Like, what does it not do? Well, I'm curious to hear that.
[00:10:06] Speaker C: I mean, so I used Zapier. So I just gave it my like your key and from there I was able to hook it to like Google Docs. You can hook it to, you know, notion Gmail, different, different apps that don't have to be linked. I mean you can configure it in a way that it doesn't link to like the overall app, but only some specific things like any automation. If, I mean I use N8N a lot in each node you can say, you know, append like column in the sheet node. Like it doesn't have full access to everything. So in that same concept is how I use Zapier.
And I wanted to, I've been testing it out to see how it does for content and marketing and just basic research. I wanted to just do a lot of research. And you know, there's so much online especially now we do paid ads and LLMs. And of course, you know, OpenAI announced that they're doing ads and LLMs as well, which is amazing for us because, you know, it means that we're doing something right. We've been doing it for a while now, so that's good. And yeah, just a lot of people posting about it from different sectors as well. So of course we deal with publishers and advertisers or somewhat of the middleman. So just we don't have time to scrape like the whole Internet and just look for that. So that's what I've had it do. Basically it's just really a researcher and that doesn't sleep. It just reads, reads, reads and then gives me context on top of news and anything that's relevant for us to either talk about internally, like in the team.
Is there something that we should be implementing? Are we better at this, at that? Just really for us to use, but also for us to then make content about or just actually be able to find the link of the specific article that was mentioned that maybe we wouldn't have come across otherwise.
[00:12:03] Speaker A: Do you have it set up on a cron job where it's a thing that runs continuously on a set schedule or do you give it prompts, say like first thing in the morning or at the start of the week? You Give it an instruction and it goes.
What triggers this?
[00:12:20] Speaker C: So far I've basically been doing prompts. So I did it in.
I didn't. A lot of people do it in WhatsApp, but I didn't want to have access to my WhatsApp because I have. You know, I've been using WhatsApp for. Since I was like 13 years old. So I just did Telegram because I.
[00:12:39] Speaker A: Right, of course, yeah.
[00:12:40] Speaker C: Use Telegram. And that's also like an easy thing to hook. You just, you know, ask your Telegram. There's like a specific. I don't remember the name now, but I think it's called like father something. It's like a weird name. It's pretty weird, but you have to ask it for a code and then it'll make a. Create something. It'll create a specific chat for. For you to be able to talk to your club.
[00:13:03] Speaker A: Okay, so you have a connect to Telegram in Telegram. You're messaging. Hey, I would like to know what's the news story out there that's relevant to us? Please summarize it, please research it. I think give me some output that we'll discuss.
[00:13:17] Speaker C: Okay, Exactly. Scrape, you know, every.
Every channel that is relevant to X contexts and return.
[00:13:26] Speaker A: Right.
[00:13:27] Speaker C: Yeah.
[00:13:27] Speaker A: Right.
[00:13:28] Speaker B: So it's.
[00:13:28] Speaker A: Do you.
[00:13:29] Speaker B: I'm sorry, go ahead.
[00:13:30] Speaker A: Do you eventually wanted your cloud bot to do this automatically, where it's just continuously monitoring, searching for news that anything's relevant to you and then feed your information without the prompt?
[00:13:43] Speaker C: Yeah, definitely. I want it to be autonomous at some points. Probably to do other things as well.
Maybe not. I mean, that I'm still gonna have to see. But I wanted to be able to do really specific things in terms of my automations that are already built manually, let's call them on platforms like any 10.
So for instance, if I want to enrich like a table in clay from a list of LinkedIn URLs. Actually, you did put me on Dripify. Dripify has been like my. My life and savior. I love so much and product placement. Yeah, sorry.
[00:14:28] Speaker B: Actually we're so that's one of those tools.
[00:14:30] Speaker A: Like we don't.
[00:14:30] Speaker B: We like. We don't even tell anyone. We're like secret.
[00:14:33] Speaker C: I know. No, you have to gatekeep them, honestly, because I totally.
[00:14:37] Speaker B: When people just leave it there, people listen to this. They'll figure it out. If they can figure it out. If not. Too bad.
[00:14:43] Speaker C: Exactly. Well, I use these tools to essentially just enrich. I use a lot of tools and I just enrich different parts and then at some Point, there's an automation.
But for instance, if I wanted to just get, you know, through Clay, if I want to enrich a bunch of LinkedIn URLs and get the email from there, like, you have to, because LinkedIn is so annoying with automations. You have to like manually get the people have engaged with you, extract their URLs, then put it into clay, et cetera, et cetera. So that's something for like a step of the automation that I would want my cloud to do.
But I haven't tested it out yet because I'm still waiting for. I haven't heard that much on integrating Codbot into like, already set Automations. So that's kind of what I'm waiting for.
Yeah, I.
Because I'm not, you know, I'm not a founder and I don't come from a background of entrepreneurship. I don't think I'm risky enough for that yet because I think that the reward might actually not be that high compared to what I'm already automating. So I'm still just waiting around, but, you know, reading about it a lot and seeing what I'll. What I'll want to do.
[00:15:53] Speaker B: All right, so you, you talk to it on Telegram, right? Yeah, yeah. So. And so that's the entire interface for the most part, and it just sends you links and stuff to like, whatever.
[00:16:03] Speaker C: Yeah. Then you also have a chat. Right. From the website that you can talk to.
But yeah, Telegram is.
[00:16:12] Speaker B: So that's the whole interface. Interesting. And then like, what is it not good at? Or what have you tried to make it do that you found it struggled, Like, I'm curious to understand limitations.
[00:16:24] Speaker C: Yeah. So the first time I set it up, I set it up with.
You can essentially choose if you use OpenAI anthropic, any API key from any of the LLMs. And I set it up on a really recent model. And so it was actually really slow because it was. Yeah, I heard that I would ask it in the beginning, I think I said, I called it Bestie. I just wanted his name to be Bestie, and he knew that I was referring to Bestie. And I said, hey, Bestie, what's up? And it took five minutes to respond.
And I was so confused. And so then of course, I asked one of our assistants that we all use, which is just basic LLMs, and I said, why is it taking so long? And it was because I was using this very old model, really recent model. Sorry. So then I went for something that was still somewhat new, but a little more not as heavy on the actual bots.
So that's something that it's not good at. And I think that that's the trade off. You either want something that's going to take a super long time, but that's going to probably have higher quality, or if you want something quite quick and you want an answer then and there, you have to take something that's a bit older in terms of a model.
[00:17:42] Speaker A: Yeah.
People are composing models together where chat interface uses, let's say, Gemini fast so that you get a speedy.
And then if it goes into deep research mode, you can use something like Claude and then ChatGPT is good for certain stuff. Or you can use like open source models. Right.
For video generation.
It's. It's really cool. What people are composing things together. Have you, have you stitched two clockbots together?
[00:18:13] Speaker C: No.
[00:18:13] Speaker A: Have you given them?
[00:18:14] Speaker C: Okay, no, I have my one. But I've, I've seen so many people actually create like whole teams.
[00:18:20] Speaker A: Right.
[00:18:21] Speaker C: And they actually have bots that manage other bots, which I find crazy.
[00:18:25] Speaker A: Yeah.
[00:18:26] Speaker C: And I really, really want it to get to that point where, you know,
[00:18:29] Speaker B: what is the advantage of this?
Do you know?
[00:18:32] Speaker A: So, hold on. No, no, you go first.
[00:18:35] Speaker B: Yeah.
[00:18:36] Speaker C: I mean, from what I understood is that, you know, you, you essentially don't have to go. If you, if you are at a, at a point where you're at this level, let's say. I'm not going to say who is high level, who is low level. I think every job is very important in a team, but if you don't necessarily have actual context on how these people work day to day, it's very hard to actually dedicate things to them and to understand how quickly they're going to be doing these tasks or not. So I think if you configure and contextualize maybe, I don't know, like a VP of a department, and then that VP will have more context than you on how to delegate things to people that are right under.
[00:19:18] Speaker A: Yeah, exactly. That's exactly what I heard.
[00:19:21] Speaker B: It manages the tasks better. Yeah. Tell me, Paul, because I heard someone last night was talking about this, but.
[00:19:25] Speaker A: Yeah, yeah. So the best case scenario that I've heard is you have one single club bot that you talk to.
So like a chief of staff clockwise and then the chief of staff has then all of the developer bots or researcher bots or you know, your content marketer bot, but that you don't talk to any of these sub bots because they're all just orchestrated by this chief of staff and the chief of staff has all of the context of your personality where you know, you're feeding it system prompts, where you know, I like things done this way. Never use EM dashes. Right? Never use words like hey dude. So thing you can give it so much context that it will delegate all of these things to the sub agents,
[00:20:12] Speaker C: essentially creating like a real world company team.
[00:20:16] Speaker A: Yeah, right.
[00:20:18] Speaker B: Fascinating. Because I met a guy last night, he was like, I have 20 cloud bots up. I'm like, what are you doing? Like what it was like it seems sounded so over engineered, like it was cool. I was impressed.
But I was telling Paul I felt like I went to some like car enthusiast, like race car enthusiast to like club or something. He's like, I got six engines and it just goes. Really? I'm like, it's cool, but I'm not really sure like what I would need that for.
Do you find that like it burns tokens and it's expensive? That's the other thing I've heard.
Yeah, yeah, it does.
[00:20:50] Speaker C: For sure, for sure. It is expensive. I mean it's expensive. If you. Again, it depends what kind of model you use. It depends on kind of, you know, membership you have depending on like which API key use left and right. It really varies.
I think there's a lot of people who have, you know, kind of figured out how to. As you were saying right there, you can use like multiple models but source into like one. So that's going to be cheaper. There's a lot of tools around it, like a lot of loopholes.
But if you just do the manual setup, it can get very expensive. And unfortunately you don't learn, you don't realize that it's expensive until you've actually tested it out.
But yeah, I mean, as you were, since you were talking about meeting someone who's got so many clubbots.
I'm in London right now and no one is doing this. At least no one I've met at events or I haven't even heard people talk about, maybe talk about it in passing, but I haven't met someone. I mean I. A lot of my friends are doing computer science or they've done it or you know, they're in that field and no one is actually like desperate enough to go try it. But I think that is, you know, a huge difference between SF or just the US in general and like Europeans is that if there's not an absolute reward that you know about and there's like too much risk, you won't necessarily take it. It's a bit more of a conservative approach.
So unfortunately I haven't had many people to talk.
I just read it about it, read about it online and then. Yeah, Yeah.
[00:22:25] Speaker B: I don't think outside of like San Francisco, Seattle, actually, those two places. The only place where I think people like you go out in public to like a meetup and they're really excited about even New York, because I'm from New York. Like, there's like kind of tech scene there, but people are. Would kind of think you're kind of lame if you want to like get really. You're getting very excited about cloud bots and stuff.
That's interesting. Yeah.
[00:22:48] Speaker C: Do you know Tech Unicorn, the Influencer?
I'm not sure on X, but she's on Instagram.
[00:22:58] Speaker B: Okay.
[00:22:58] Speaker C: And she's amazing. And she was in Davos and she actually went to an open call event, I believe in sf. And Ashton Kutcher was there. I didn't know that he had background.
[00:23:09] Speaker B: Dude, there was like a. I wasn't. Oh, I missed it. There was a. They did a Open Claw con, like a conference. Like, they must have just organized it like at the last minute because, like I would have flown down. I live in Seattle. It's only like an hour. I would have gone for it.
And it was like a thousand people registered and I was going through like the listener. It was just like a who's who of like Bay Area technology people.
Insane. So like people in San Francisco are like going bonkers for this thing right now. I think, like, for me. And I want to hear like both your perspective, actually. Like, I think what people are so excited about is that it's finally going to fulfill on some of the early promises, I think of AI. Like you heard about like agentic AI and agents and all this stuff. But like, for the most part your experience has been like, you go to a website and you type stuff into a chat box and it kind of spits stuff back. I did experiment with the Claude connectors and a few MCP servers and they didn't work very well and they would fail and they would like, they were just totally a mess. I asked an engineering friend of mine about and he's like, oh, you're doing it all wrong. And I looked at what I was like. I don't know. Like, I was like. I followed the instructions exactly and like, it was doing it wrong, not me. Anyway, so I think like Open Claw is like the first time that we're seeing it really fulfill on these types of value props where there's like literally an agent. There's this agentic thing that you've set up and it can do a lot of stuff. Like it can go out on the Internet and do tasks and do things autonomously. Like, it's the first example, I think, that we've seen and that's why everyone's so excited about it.
[00:24:55] Speaker C: Yeah, I think there just needs to be some security parameters around it and then we'll be fine.
[00:25:01] Speaker B: Yeah. So what do you think about the security piece? That's one thing that's holding a lot of people. People back. I'm curious what your perspective, and it sounds like you've put a lot of guardrails in place yourself. But I'm curious what you think about this security in general and perhaps how you think it should evolve.
[00:25:18] Speaker C: Yeah, I think there's been conversations online about them saying that they are actively building it so that you can have options to have it more secure or just decide essentially what I did on Zapier, where you decide what to give it, rather than having it be fully, like autonomous on your computer or whichever device you're using. Because I think that is very.
I mean, I have so many things on my computer that I would not want to put into like an LLM, for instance, because, you know, there's just. I don't know. But what, what I definitely want to.
To wait and see before doing all of this is to see if there's a way to just, you know, if you wanted to have your contacts to be able to reach out to some people but not have your credit card information, you know, because even if you don't have your credit card in a file stored somewhere on your computer, if you have. If you have it saved into your Amazon so that it goes faster, that's the sort of information that I could get and I. I have. You know, I'm not going to say that agents are there to. To kill us and to steal our money and our information, et cetera, but it does happen that there's an issue or that it misunderstands something and it wants to do something.
Well, like the story that I mentioned in the beginning. Right. Booking a master class was essentially to help a user, but it's not the goal that you want. So.
Yeah, but I'm very excited to see where this goes. I hope that if we can actually delegate a lot of the things that take so much time and that are not necessarily, you know, that aren't enriching to. To us and we can just go outside and touch grass whilst Our agents are building some things and then we can come back and just be the like, last level person in, in the, in the journey. Then, then why not?
[00:27:11] Speaker B: Yeah, no, I, I, I agree. Okay, so to close this out, I have a fun game for the three of us. So since A16Z is making all these aggressive investments into AI companies, I want us to all give a number that we think A16 has already offered this guy who created Open Call. And it's got to start, it's got to start at 100 million. So Paul, how much do you think they've offered him already?
[00:27:41] Speaker A: North of a billion dollars.
[00:27:43] Speaker B: They offered him a billion dollars?
[00:27:45] Speaker A: Damn.
[00:27:47] Speaker B: Do you have a guess, alex?
[00:27:50] Speaker C: I mean, 16 have to do over 100. I'll say 160.
[00:27:56] Speaker B: Yeah, it's got to be more than 100, right? Like this thing is.
[00:27:59] Speaker A: Yeah, for sure.
[00:28:00] Speaker B: This thing is like the next cursor or something, right?
[00:28:03] Speaker C: Like, but there's so much to build around it. If you actually want to make it a product that you can, you know, have like any consumer use, you have to put so much around it. So I think there's still a lot to build.
[00:28:14] Speaker B: Okay, so I'm gonna go with half a billion, I think 500 million. Because I think, you know, nailed it. That, like, yes, I, if I was a 16Z, I'd look at this and go like, you need as much money as, but you need Paul's billions.
[00:28:26] Speaker C: What you need.
[00:28:26] Speaker B: Because like, I think they could be marketing and releasing like hardware versions. Like, I think there's multiple versions and there's multiple use cases. Like if you just buy this little cool device that was your little agent, and that way it's all like air gapped off. You could like literally pull the cable out. You got it all there. And there's very clear parameters about like, what data information stored here, what data information in the cloud. I think there's a giant market for that. And you just get the box, you turn it on, it's like, hi, I'm Claude Bot. It's all fun and cool, like a little like, like a kid's toy. I think that's a giant market. And then I think there's a slightly more sophisticated, why have it. Excuse me, but there's a different market for people that what's cloud based, which is what I want. I looked at all this and I went like, sounds like a lot of effort. I think it's going to, I think there'll be a cloud version in a couple of weeks anyway. So I'm not going to bother with it. I want to go and like log in and just be like, this is, this is who I am. These are all the accounts I've created for you, my bot. Like, here's your Gmail. You're now Charles at Vibe, your sass. And like, here's your calendar and here's your stuff. And then I just kind of chat with it online and tell it to do stuff. So things that, that's a whole different version of that. Like, like, I think, I think they nailed the next big product in the technology space in general. I called it like the Friendster moment, which you're way too young to remember. But they were like the first social network that showed the promise of like, wow, everyone go on and like connect to social network. And it was huge. And the company kind of failed. They raised a bunch of money, but it's the same thing. I see this being like the, the possibility of like what I think will be one of the largest companies ever. Like, I think that like, if they nail this, like your kind of personal AI agent and there's be dumb a couple of different versions of it, everybody on the planet will want one.
[00:30:12] Speaker C: What about Moat Book?
You think that's something?
Stay improve.
[00:30:19] Speaker B: You know, what do I think my osmium Moatbook is? Did I tell you this, Paul? I think it's an example of like, when you over train on Reddit data, you get bots that go and duplicate Reddit. That's what I think about Moat Book.
[00:30:35] Speaker A: I think eventually Reddit will open the project.
[00:30:38] Speaker B: Too much Reddit on the brain is what I think about. I think it's fascinating, I think it's hilarious. But that's, that's what I think it did. I think there's so much Reddit data in there, they kind of went like, let's make Reddit, because that's what we know.
[00:30:51] Speaker C: Yeah. Anyway, I know you gotta run threads for, for us to read, right.
[00:30:57] Speaker B: Oh my God, the language and all that stuff.
Wait, I don't.
[00:31:02] Speaker A: That's right.
[00:31:03] Speaker B: I don't want to keep you from dinner. I know it's late in, in London. I know you're having a great time speaking with us, but I want to keep you, keep you on. Yeah, I keep. You could go back. What? Keep you on track.
Because it's like there and then. Yeah, we'll have you again back on the show. In fact, like, make some more progress. Like when you come back, when you string together like 20 clod bots and they like, that's it. Take over The United Kingdom or something. And like let us know how that went.
[00:31:32] Speaker C: Even be part of the company. I'll have him, you know, touching grass and I'll just have a thousand boats instead of him.
[00:31:40] Speaker B: Oh my God.
Awesome.
[00:31:43] Speaker C: Well, thank you so much for having me and of course, thank you on the progress.
[00:31:48] Speaker A: Yeah, come back.
[00:31:50] Speaker B: Bye. Bye.
[00:31:50] Speaker A: Let's show us.
[00:31:54] Speaker B: All right. That was awesome.
[00:31:55] Speaker A: I like your prediction at the end.
[00:31:59] Speaker B: I do. Like, yeah, it's, you know, they're going to create a, like a, they're gonna, they're gonna conduct a coup.
They're gonna take over the United Kingdom. Maybe they'll even.
[00:32:10] Speaker A: That's where we're all headed.
[00:32:14] Speaker B: So I, you know, I signed up for AI dot com. I guess this is like claudebot online is. I don't know. I didn't get the invite back. I signed up for the wait list, but I haven't gotten access yet.
Have you heard about this at all?
[00:32:29] Speaker A: Of course. They ran the super bowl commercial. Super bowl ad.
[00:32:32] Speaker B: Yeah, yeah.
[00:32:33] Speaker A: So I did not look at it.
[00:32:36] Speaker B: Like, you know, it's only one of those things, right? Like, I don't know, I just signed, I signed up for everything on the Internet. Like, just give it a shot, right? Like, why not?
My understanding is like, it's kind of like a cloud based cloudbot like I was talking about. So you, at least in the signup process you have to give it like you have to do your name as like a URL and then you have to give your bot a name and they have two different URLs and then it says thank you and then it states your information and then we'll get back to you.
So your bot has like a URL and then you have a URL. So I don't know what's going to happen. I'm excited though.
[00:33:10] Speaker A: Huh?
[00:33:12] Speaker B: I mean, it seems really straightforward, right? Like you said this thing online somewhere, you give it its own email, its own calendar and then starts doing stuff.
[00:33:23] Speaker A: Sounds good.
Here's my money, here's my credit card.
[00:33:27] Speaker B: Just take it. Just give me it.
Oh, hilarious.
[00:33:32] Speaker A: I don't know, man. I'm waiting to see. I really, really am just waiting to see for, for something productive. I went to the open cloud meetup this week, I think it was on Wednesday.
Saw some presentations, lots of cool stuff. One of the best example with somebody is having recently or about to become a dad or a parent, they had open class set up so that it does all the research for the nurse, all of the research for booking things related to the baby.
Does a whole bunch of like, additional research to like teach the parent how to do, how to, how to have a baby, basically. I thought that was really cool.
But you know, outside of that, I'm not sure. I'm not sure what.
And I also saw the, the demo where a person set up a team of cobbots where she had a chief of staff clapbot with developers at Cotton Content Marketers.
Very cool. Very cool.
[00:34:41] Speaker B: Yeah, I mean, so you and I examined it. So I was like, okay, there's parts of my newsletter that obviously I could, I could automate. And we looked at it and I actually think there's too many steps to use.
Yeah, yeah. So it's not where. So, so I have some obvious tasks that I want to automate, but I don't think any of the tools are at that level yet because, like, it was funny. Like in my mind, like, oh, it's simple.
Just duplicate the old newsletter, change this graphic, do this. But it, when you go through what actually has to happen, you have to like log in a tool and then go here and click here and click down. And then like you have to drill all the way down into it. Then you got to go into like canva and create this graphic. And there's like ton of parameters around how the graphics get created. Right. It sounds like a simple task, but when you dive into like all the different steps, that's. It's a lot. There's a lot of steps.
[00:35:31] Speaker A: So I think the two problem is this one is too many steps and the output so far from everything I've seen is not that good. Yeah, I think that's the biggest problem. It's mediocre. Right. So the output is mediocre and then the steps to get there is super complicated. So that it doesn't really incentivize me to reverse engineer and optimize it. Turning to engineering exercise.
[00:35:58] Speaker B: They're making me choke. Yeah, it's. It's a hard way to get mediocre output.
That's what I need.
Well, so I, my. I think this is actually a really important point. And so, and I've been recommending this and I will continue to recommend it, that I think that like all these tools and I want to broaden the scope because I think that the word or the term AI now has become synonymous with automation. And they're actually two different things.
And I remember a couple years ago with all the engineers and technical people, we actually had some pretty big arguments about this, like what the definitions of things were. Do you remember even, like ML is not AI debates.
[00:36:41] Speaker A: Yeah, that's right.
[00:36:43] Speaker B: We've come a long way where there was a point where I would get engineers who were really upset. They're like, you can't label this as AI. It's actually ML machine learning for those aren't up on the, on the, on the Bay Area lingo at machine learning, which is like a precursor or a comp, I would call it a component of certain types of AI. There's lots of different types of AI and artificial intelligence is something, is something completely different.
And so at least that was like, I was talking about debates like that and I was just kind of. My attitude was always like, look, I'm in marketing. I got to explain this to like normal people. And so they understand what AI is. And so we're just going to go with that. And they were really mad because they were like. Or some people were mad because they're like, look, there's these AI labs and everything. And if you tell them like everything I'm doing is AI, they would laugh at me. And I was kind of like, okay, that's true.
I actually understand where you're coming from. But I was like, my job is to like market this thing and sell it. So like, sorry, but, but my point is that there was a point where people were like, very rigid about what these terms mean. And at some point we crossed this
[00:37:50] Speaker A: like line where now we call it agents or AI.
[00:37:53] Speaker B: Any kind of automation is AI because there's all kinds of tools I use that are like automation tools.
[00:37:59] Speaker A: Right?
[00:37:59] Speaker B: Zapier has been around. They're technically not AI. They're just tools that automate stuff. And I have all of these automation tools and I've had them for a long time. I think they're great. And my point was that it was a very long way of saying I think people should automate as much as they can, except for the messaging, except for the content. I think the content piece is the thing that you need to like, have a hand on, pay attention to, customize. I also think that the level of customization that people are doing with AI generated content is overkill. And I've done this a long time and I've seen a lot of instances of this.
And it generally doesn't yield better results.
Like hyper personalizing content, hyper personalizing advertising, hyper personalizing sales outreach hasn't made a huge difference in the performance.
Micro segmentation I believe deeply in. Right? Like, are they CEOs of companies that are 50 people? Like, that's a segment that's Very clear.
But like hyper personalized that for each one haven't. I haven't seen a big advantage. I've actually seen disadvantage because the stuff can be so ridiculous that no one believes anyone would, would write like that.
Like the emails are like, oh, I. It like just starts mentioning all this stuff about the like it's just not realistic. That's not how normal people. That's right, Normal people talk. Right. So I'm all for an automation. I just think that like don't automate your content.
[00:39:33] Speaker A: Yeah, we talked about this. Right. It's kind of in marketing. We want to make the recipient feel they're special. Right. Like we are especially marketing to them. You could do this with hyper personalization, but to a certain degree, right. Like if I'm getting an email that obviously look like a code outreach and you're just going on and on about the events that I've posted pasting paragraphs on this YouTube transcript, I'm like, you know, obviously there's automation.
[00:40:03] Speaker B: So dude, like what's interesting is this, this has been around a long time.
So I remember like working on.
So they used to call it dynamic. I guess they still call it dynamic Creative. Was this idea that like you could auto or auto personalize you. Yeah. You know, you could personalize the ads. Right. And there was all this experimentation when trying to do it and they're very hard to implement because there's just so many use cases and steps and whatever. But like ultimately, any examples I ever saw where we did it, the results weren't much better.
Like, like it's fascinating, right? Like everyone's so excited about it.
[00:40:39] Speaker A: Exactly what you're talking.
[00:40:40] Speaker B: So excited about it. But like I'll give a bad example or just use an example. It's like ultimately the ad came from like a big brand. Let's call American Express.
That's what matters. Like the ad is from American Express.
That's really important. Like who is the brand and then the context of the offer. Like this is to get a credit card because you are so and so and you fit into certain parameters and you should get a credit card. Right. So then trying to personalize a bunch of stuff around it. Like I think when you talk it all the way through, it's obvious. Like it doesn't add a lot of value. No. Like they want an American Express credit card because they understand that product, they trust that brand, they know what it does. They're in the market for that kind of thing because that's what they want. And then you like figure out that they like happen to, you know, live on this street in Seattle or whatever. Like, who cares?
It doesn't matter. It's not relevant to buying that product.
[00:41:32] Speaker A: It's not. It really isn't. It's like, I know exactly what you talked about. Maybe like 10 years ago. Lots of ad companies, a lot of time into dynamic creatives where, you know, you go onto a shopping site, you add something to cart and then this item, customized to your size, color, follows you across the Internet.
[00:41:55] Speaker B: It didn't. It doesn't matter.
[00:41:57] Speaker A: It doesn't matter, dude.
[00:41:58] Speaker B: Like, oh God. Fifteen years ago, there's some ad platform I was working on and we tried all these things, like used. We had all this data, doing all the stuff and then one day, like we somehow linked it up to like a forum where people would ask questions before they made a purchase. And it turned out like, it's so obvious in hindsight. If you go to a forum, you ask a question about a product, you are giving a really strong signal that you're going to buy that product. And like, we've ripped out all the other signals because that one was such a strong buy signal that the other ones actually reduced performance.
And I love this point in particular because I think there's a naive and like inexperienced perspective that more is always better, more data is always better. And I've seen cases where it was not better. Having more data did not make it more effective. What mattered was having the right data, the right signals. That's all that mattered. The other stuff just increases processing cost and convolute. Convolutes what you're trying to do. In that case, it actually reduced performance because it was like waiting other signals and that signal was so strong that like it worked.
[00:43:14] Speaker A: Totally agree. How do we start talking about this again?
[00:43:17] Speaker B: I don't know. So, yeah, okay, okay, let's, let's move on.
Let's.
[00:43:24] Speaker A: There's a couple, maybe two or three different things that we should talk about this.
[00:43:30] Speaker B: I feel like they're all related ultimately,
[00:43:31] Speaker A: but yeah, okay, they're all related. Right.
[00:43:34] Speaker B: We gotta talk about. Something big is happening, right?
[00:43:38] Speaker A: 80 million views on X. You wanna just like share it or just kind of talk through what this guy's saying?
[00:43:47] Speaker B: Okay, let me try to summarize this for like, for the, for the, the, the, the, the, the 500 people who haven't read it already.
Yeah, I feel like this went everywhere.
It's a fantastic article. It's super well written, basically. Yeah. I think about one perspective of like, what the AI future could look like. Right, right. And of course, like, it's. I would say it's hyperbolic.
It's predicting that AI will basically do everything and there won't be much left for people to do. If I was like, kind of just summarize the article.
[00:44:25] Speaker A: Right.
[00:44:25] Speaker B: It makes like, some compelling arguments, and I think there's a lot of facts and a lot of truth to it. I also think it misses a lot of stuff. And I don't personally think UBI is coming. I don't think all the jobs are going to go away. I think it's going to take much longer for this transition to happen. Even though I totally believe AI is real. It's a real trend and it's going to change everything. But I think this article is just like pushing that to a level that is, you know.
[00:44:57] Speaker A: Right. It's like the canary, the coal mine. Wake up.
[00:45:01] Speaker B: That's what he's trying to be, Right?
[00:45:02] Speaker A: Yeah, of course. Right. Like, he's, he's. Well, first of all, the background is that he has an AI company. He has things to gain by pushing this narrative. I'm gonna lay that out there. And then the second thing is, is
[00:45:15] Speaker B: he an investor in Grok? Right.
[00:45:16] Speaker A: Yeah, he's like an investor. So also.
And the second thing is, like, he's. He's pushing the. For normal people to wake up, to pay attention. This is happening, whether if you like it or not, it's going to happen. Not on your timeline, but the timeline that AI seems to be progressing, which is doubling, tripling, quadrupling the capacity every six months. Right. So not a lot of time left.
[00:45:45] Speaker B: What do you think about it?
[00:45:48] Speaker A: I think timing is everything. Directionally, he's right on point. I think AI is here to stay. It has disruptive qualities to it. It is going to change lots of things. But I think timing is everything.
[00:46:03] Speaker B: Yeah.
[00:46:03] Speaker A: From his article, it sounds like it's going to be here within 12 to 18 months.
I think the horizon is closer to five to 10 years.
[00:46:15] Speaker B: All right, so this is, this is interesting discussion because, like, I feel like it's all related right now. So that this happened. Then you got the SaaS apocalypse, right? Like, all the SaaS stocks are selling off for sure.
[00:46:26] Speaker A: Yeah.
[00:46:27] Speaker B: You've got Claudebot, you've got like, there's actually other stocks selling off today. Like some dude. I can't believe this is real. A karaoke company sold itself, transformed itself to an AI company, is now selling like, AI navigation software and like, the largest, like, logistics companies that do, like, shipping all their Stocks are crashing because some guy thinks that like someone's gonna clodbought their way to like, like managing global shipping and trade. Like it's totally insane. Like it's wild what's going on right now.
And for me to contextualize all of this and you and I were talking about this. I, I, for me it's exactly like dot com. It's not like 3D printing we talked about already. 3D printing like ultimately like didn't really happen. Not the revolution that people thought was going to happen wasn't there.
The AI revolution is 100 real in, in the way that I think the dot com and Internet revolution was. So the Internet revolution did transform the entire world.
It just didn't necessarily happen in the, at the same pace that people perhaps expect at certain times. And that transition was a lot smoother I think than other people expected. But a lot of stuff changed. A lot of jobs did go away. A lot of companies did fail. Like the, the narrative. I think in retrospect, like knowing what happened 25 years ago, I guess I would make the case someone who was like experiencing it and was there. I like to say that like you know, at the time people got overly excited about it for sure. Companies went public and they were a bit ridiculous and they crashed a lot of them. But I would say every crazy idea that I had heard about and witnessed experience, talk to about with people during like 98, 99, 2000, every single one came true. There isn't like a single crazy idea that we talked about that at the time maybe was super far fetched that hasn't actually occurred. Like Blockbuster went out of business. Streaming video.
[00:48:28] Speaker A: Streaming videos, right.
[00:48:30] Speaker B: Like the record stores are gone. All music delivery happens across the Internet. Right. Like most e. Like mailing, like mailing, mailing things. Like okay, business mail. Like I used to have to like ship stuff. Yeah, we'll get to facts but shipping stuff. I used to have like ship graphics and like elements around like produce things and it. There was just a lot of like shipping. Like I don't ship anything I don't need. I don't have a FedEx account anymore. I don't send things.
I'm glad you brought fax machine because like it's one of those quirky ones where like it does survive for some reason in healthcare. I've never totally understood why and no one can ever totally explain it to me. My best guess is that there's some legal element to it.
[00:49:19] Speaker A: Yeah, yeah, yeah, yeah.
[00:49:21] Speaker B: Remember we had DocuSign, we had court on from DocuSign and like there are some questions about the legality of like digital documents and I think that's why they do that. No one's ever been able to explain it to me effectively. If anyone's listening to this and can tell us specifically like why fax pro scenes are something that are still in existence in the healthcare industry, I'd love to know. But you're correct that everywhere else gone.
[00:49:46] Speaker A: So again, it's direction directionally correct, just timing. Right. I'm sure people in the early 90s in the Internet industry thought that it's going to change everything by, I don't know, in 95, 96.
[00:50:03] Speaker B: I mean I do. There's so much sun to talk about. Like there's. There's industries that like people assume would go away that didn't. So real estate agents are still a thing. There's industries that people predicted would go away, that did. Travel agents are gone. Like I don't even know sometimes like how to fully explain it. Like the best explanation I could come up with for like real estate agents versus travel agents is that real estate ultimately is a local business and there's a lot of local knowledge that's hard to codify and put online and it just, it still could happen. It's just that like the knowledge of like how to, how to sell a home and the regulatory things that are required to do it in Seattle are different than in New York City and are different than somewhere else. Right. Something universal about it. So it's very difficult to translate that into a universal element online. But travel agents, it's universal. There's nothing different about booking a trip from California to Hawaii and booking a trip from New York to Hawaii. So I think that's the explanation, but I don't know.
[00:51:16] Speaker A: So I think your perspective is AI will have this effect on the job market, as the article has stated.
Are you optimistic that people just figure things out, the government will intervene?
What's the optimistic view here?
[00:51:39] Speaker B: Yeah, that's a fascinating question. I think that like, I think that, I think that we might be experiencing it now, some of this. So like there's all these layoffs.
It's super hard to tease out. It's quite possible that there are going to be less white collar jobs and that we're experiencing right now that like some of the more simple tasks like we were just talking about, like claudebot is doing like research stuff like these were jobs that like you had interns do. Right. And so does the bottom 5% of inefficiency in the corporate world go away. Like, that might be happening right now.
So. So it's having an impact, but I don't believe it's like 50% of, like, people are out of work and you need ubi. Like, that's where this stuff gets to be too far for me.
I feel like it's like the climate change stuff. There was. There was like, a long time. It's very, like, I would use a term like alarmist approach to climate change.
[00:52:41] Speaker A: Right.
[00:52:42] Speaker B: And a lot of people still believe that stuff, but even Bill Gates came out and said, like, it's not that big of a risk. It's not gonna do what we said it's gonna do. There's always this, like, group of people who, like, push it to this extreme level and scare everybody. And they have the reasons for doing that. And, and I think they're. They're generally proven incorrect.
[00:52:59] Speaker A: That's it.
[00:52:59] Speaker B: And so the climate change one, I might get some haters for this, but, like, because I bring it up all the time, Al Gore's documentary and in, you know, I can't remember what's called now An Incomplete Truth, Inconvenience, Inconvenient Truth. Anyway, Truth. It's so long ago there's predictions in that, like, that did not become true at all. So he was like, there will not be snow on Mount Kilimanjaro in Africa. And he gave a time frame that was like, right. 25 years ago. They have the biggest snowfall they've ever had.
[00:53:30] Speaker A: So to, like, stay on track a little bit. Right. So there's lots of predictions, AI. That may or may not come true. We really don't know. Right.
[00:53:37] Speaker B: Like, so that, that's.
Well, no, no, I was saying. No, I'm. I'm taking it one step further. I'm thinking that. I think that, like, the dire predictions are similar to climate change ones, that they're not going to happen.
Right now it appears like it could happen, that everyone's gonna be out of the work. It's gonna be really difficult, but I don't think it's gonna happen at all. That, that like. And we have plenty examples where, like, people make. These extreme. Extreme predictions haven't come true. I think there's like a.
There's like a. There's. There's a lot of reasons why people do it. So I just don't think the most extreme ones are gonna happen, you know, and the climate change stuff, like, we can measure the. The temperature, like, so there were a lot of. A lot of facts were true and are true and have happened. Right. So that's why for me I put it in the same bucket. It also has this like apocalyptic fear element to it. I think that's why it's also similar
[00:54:28] Speaker A: of course people, people like clickbaity titles. Right?
[00:54:32] Speaker B: Right. It gets people, it gets people motivated. There's fear. Like it, most people, it most, it motivates, it motivates people to, to say things and that's why.
So I think that's the incentive for like pushing it to that level.
Right. And that's why I think it's important to like be honest about like, like where what, what, what in the past has happened. Right. So we have a couple examples here. We had like the dot com boom, we have climate change, we have 3D printing. Like there's all these trends that come along and just you know, having some perspective about like what the outcomes could be I think is important.
[00:55:07] Speaker A: Yeah, I completely agree. I do think like to your point, some white collar jobs will go away most likely. The.
[00:55:15] Speaker B: Yeah. And I believe they might be going away right now. Like that might be happening today.
[00:55:20] Speaker A: The. So the actual real question here that I'm constantly thinking about is if we replace these low paying jobs where a lot of let's say young people are getting starts at like a research intern, et cetera, like a juniors developer, how would they get into the more senior places or jobs without that past experience? Right. Like the joke in the industry is I'm looking for somebody that's like right out of school with five year experience.
[00:55:52] Speaker C: Right.
[00:55:53] Speaker A: That's basically what the entire job industry is looking like. And AI is facilitating more and more of that types of thinking.
I think like education, whatever that is, either in like school or like in a workplace needs to really figure that out or in five to 10 years we're going to have entire generation of young people with nothing to do because they're either overqualified to do the things that's automated and under qualified to do the senior more strategic jobs.
[00:56:30] Speaker B: Yeah, I. Dude, it's a really good question. It's come up a lot I think like I feel like one, one answer is like at least my experience with let's say like the dot com was that at the time people asked a different question which was like hey, we need young people to do this stuff because it's new and no one knows how to do it and no one has experience with it. So it's almost like people just had a different perspective about it. Like we're now we're looking at this Going like where all these young people are going to get trained. How are they going to know anything?
We, I don't know. I don't know why, like, but people looked at it differently which was like, oh, we want young people.
I'm of course being the optimistic person that I am. We want young people to figure out this Internet stuff. You hire young people to do this Internet stuff. Like they're gonna figure it out. Right? So I guess like maybe my answer would be to spin it that way. It's like we had like Alex on, right? Young people are like figuring out how to do Claude bots and do cool stuff and like that's what the future is going to be.
[00:57:32] Speaker A: Do you think? Oh, I know we have to end here, but do you think the reason why that happened was the older people never thought Internet was going to replace their jobs?
[00:57:43] Speaker B: Yeah, yeah, I don't. That's a great question.
[00:57:47] Speaker A: Right.
[00:57:47] Speaker B: Like I do some so well. Okay.
Some industries were very aware of the fact that they were being disrupted.
[00:57:58] Speaker A: This is specifically in the Internet.
[00:58:00] Speaker B: Yeah. Back even back then, right. Like the music industry, totally new. They, they tried and tried and tried to create all these things. They had Napster come out, right. So they, it was super obvious to them.
And magazines, so magazines knew pretty early on.
I think newspapers actually probably stuck their heads in the sand. But I think magazines, they totally understood that the Internet was going to be the delivery vehicle. And now you have Instagram, which is basically magazine content in my opinion.
And that was like, that was obvious to them. They Conde Nast, they invested billions trying to do everything on the planet. It just, I think turns out that it's very hard to. It's, it seems like it's nearly impossible to pivot. Innovators dilemma. It's a very famous book. If you haven't read it, check it out. But basically it explains this very problem that big established industries can't seem to disrupt themselves.
[00:58:50] Speaker A: That's right. That's right. Okay. So I think with AI that's more towards that. Right. Where the younger Internet native companies are disrupting the old ones or the AI native companies are going to disrupt the.
[00:59:06] Speaker B: Oh, this is a good question, really good question. Because like I do think that some companies, some industries and some, there are some examples of people who have like learned from like let's just say the dot com disruption.
[00:59:21] Speaker A: Right.
[00:59:22] Speaker B: And like.
[00:59:22] Speaker A: Right.
[00:59:23] Speaker B: I. It's fun, it's, it's fun to say this. Like this time it's different.
So I think, I think Google, Microsoft, like the tech giants they have figured out that like they need to prevent or mitigate or find a way around disruption. And I think they become very effective at doing it in ways that traditional companies didn't do in the 90s. And I think they learned. Yeah, I think this is true.
[00:59:50] Speaker A: 100% agree. Because the founders are still super young. Like by all means, they're super young. They have the memory.
[00:59:59] Speaker B: Yeah, go ahead. Right, right.
[01:00:01] Speaker A: Well, yeah, it's, they have the memory. There's.
To them, this is existential crisis. Right? Like for, for a company like Google to not introduce AI into favor products.
[01:00:13] Speaker B: Yeah, yeah, I saw. It's more, it's, it's, that's one answer. There's more to it than that. Like they have studied this problem and some of these companies have been successful at overcoming innovators dilemma, I think is the broader way to describe it because you're, you're 100% correct. Like, the Google founders are like startup guys. They still get it.
[01:00:31] Speaker A: They're cool.
[01:00:32] Speaker B: They're like, they go to meetups like in cafeterias and hackathons and basements.
It's cool. Like, it's wild. I can't believe it. Right. Like these guys are like still doing that stuff in a garage in Palo Alto. Right, right. But the example I kept thinking about was like Satya from Microsoft. He's like a Microsoft company guy. He's not a founder, he's just a total baller CEO. And the way that he like navigated this AI thing with like OpenAI, I think is like insane. Like CEO of the century. Like he, he got, he, he had, he got, he got them into like his, he got them on his side.
Huge success. They negotiated this brutal deal that I think is very good for Microsoft and now they're developing a competitor. Like talk about like trying to solve their problem as Microsoft. Like, I think it's a good example of them like doing it.
You don't agree?
[01:01:24] Speaker A: Let's, let's end it there because this will be topic for another.
All right, an hour worth of episode.
[01:01:32] Speaker B: Any, any announcements, Anything we need to end with no.
[01:01:36] Speaker A: March. March 11th. March.
[01:01:38] Speaker B: Yeah, just the event is going really well. I've been posting about it for all over, so go check it out. Send me a dm. Have any questions? But yeah, we got, we still have a few spots. Got a great lineup, but we're doing startup pitch competition and a mixer with investors and founders in San Francisco on
[01:02:01] Speaker A: March 11. Who should show up? Founders, obviously.
[01:02:05] Speaker B: Yeah. So we meet, we make people apply and yeah, you need to be a founder. If you're a stealth founder, for the most part, it's good enough. I think you put on your LinkedIn that you're like, you're doing something cool. I like that.
Investors, like angel investors, people that work for a reputable venture firm. Both are okay.
Yeah. And that's who we try to attract to the. To the event.
[01:02:32] Speaker A: Perfect. Yeah. And open claw owners, if you're running 20 open cloth.
[01:02:40] Speaker B: And we. We will take that into consideration.
[01:02:42] Speaker A: That's right. Okay.
All right, That's a wrap.
[01:02:45] Speaker B: All right.
