3x exit Aussie founder (ex-AWS, ex-Microsoft) | Nick Holmes à Court

Mackenzie: My guest today has successfully scaled and exited three VC backed startups. One of his companies generated customers a hundred million dollars in revenue. He spent more than 20 years in tech working with companies like AWS and Microsoft. But today he is all in on AI agents. And in this episode, we're gonna dig into everything he's working on right now, including the AI brain that basically runs his life, what running a team of agents is actually like. And what stays valuable in a world where anyone can build software? Nick, it's great to have you on the show. First question, what is the thing that is most exciting you right now? Sure. I mean my background had always been trying to automate things that humans did or to find things that humans wouldn't be willing to do the work to discover. Right. So, you know, I remember early in my consulting career as a Microsoft consulting for a bunch of time. And it was like, how do we automate these, you know, this marketing landing page creation so we don't have to create these pages anymore? Or how do we automate three way f three way financial production for an investment bank so we don't need accountants anymore? And and like that That superpower is is out now. So, you know, I think it became pretty hard not to I stopped software engineering for 10 years and about two years ago got back into it just after some of these tools came out because it was like, wow, there's stuff I've wanted to build for a real long time that, you know, I I knew like I'd like to build that, but I don't have six months to spend on this. And it became like, yeah, I've always wanted a like a stink bid tracker on futures markets. Okay, I can do that in two days. And so just all of a sudden started producing stuff. and you know, it's obviously become a super powerful time for people to produce. So, you know, for me, my my role and focus, you know, I'd left a big tech a few months back. I'm like, how do I automate myself? Like, I just want to automate me. So I don't have to, you know, I can be a ten X, hundred X, thousand X me in everything that I do. so that's been my my passion project for the last three months and it's come out like real all sorts of really novel and interesting things have come out of that pursuit. So that's where that's where I started and that's what most interests me is how do you have the highest amount of impact doing the least amount of work. And I always joke it's like I will do so much work in the pursuit of not working. you know, I can work infinite time if there's some carrot at the end of this, which apparently is there won't be work involved afterwards. So yeah, there's there's a paradox in and I think many AI builders are finding this right now. Like, yeah, you can might you might be ten or hundred times more productive, but you're working more than you ever were. So Fascinating. Do you have any specific examples of things that you've like automated or built recently that have been, I guess, a game guess game changers for you? Sure. Well, I mean, there's two projects that I'm really been pursuing and I'll share the fuck at you about both. But the the first one is like just the problem that I was experiencing is as I'm engaging with like Chat tools and you know, I'm using Cursor here ChatGPT there, and like like most of us are engaging with that. I'm like, these things don't have context about me and my life. There's too much going on, right? So you know, my passion projects recently has been to build out a brain for myself, right? Like a fully agentic brain. So it knows everything that I do and can decide just how I would decide and talk just like how I would talk. And be available everywhere. So, you know, regardless of whether I'm using Cursor or Claude or OpenAI or WhatsApp or Telegram or Hermes or, you know, whatever AI I'm using, how does that AI have full access to everything that would be relevant to me in my life? So, you know, I think that I I've you may have had this experience as well. Like, you know, you take a meeting, you have a bunch of really interesting things get said and you'd say you'd do some things and they said they'd do some things and it's like, then you move on to the next meeting and and, you know, I I'd end up writing down a bunch of notes or like to-dos and that kind of thing. And I get to the end of the day and my, you know, my my notes is like this boss that I didn't want that was like, you know, past me during the day committing to do stuff that I don't want to do now. and there was lots of great, you know, insights and principles and intelligence that like were just gone, you know, and maybe what was that thing that they said? So so I think you like many of us, I just started recording out some of my like all the things that were happening and then Okay, well then how can we build intelligence out of this? Like what principles are emerging out of this, what playbooks are emerging out of this? And every time I engage with, you know, email or you know, I'm I'm researching something, okay, well, how does that make this thing smarter as well? So, you know, and I I'd seen the emergence of so this is for me has been an absolute game changer because I've got complete access to all my context on everything. And it's not just for me, it's for agents, right? Because in the pursuit of having agents act as I would, they No they need access to all the context that I would have and my playbooks for how I decide. so, you know, through that process, I've been spending a ton of time like deep dive interviewing myself and others on how they decide things. Like what are your judgment frameworks? Like, you know, we have a ton of intelligence in the world of AI. What we often lack is judgment for, you know, certain scenarios. So you know, I've been building software software businesses for 20 years, both here and in North America. And you know, made lot of mistakes, made a lot of great, great decisions as well. How do I decide? You know? and so I w also went and interviewed other founders and was like, how do I extract as much judgment out of folks so that my agents know what to decide for themselves and when to escalate? or can I use, you know, judgment playbooks out of out of Nick or so those things have been really cool. So having a an agentic brain just to power my entire life has been, you know, that's a complete game changer for me. Like You know, I I I asked it recently and I can't tell you what it was, but I'm like, I wonder what what this thing actually knows about me. I'm like, what are my hopes, fears, and desires? And like it's scary, you know, like it's it's got everything. So you know, lots lots of stuff there. But in terms of efficacy, it just means that my agents can run and I'm never reprompt reprompting them on context. you know, I think these these things aren't noble to solve. you know, there's a ton of interesting stuff around memory management, around conversation history about What becomes interesting as well is like, just like all of us, like I have to be quite mindful about what I put in my brain, right? My agentic brain. And in the same way I do in the real world. Like, you know, you you don't wanna sit around and eat a ton of Doritos like that's bad for your body. You don't want to sit around and watch a ton of TikTok that's probably bad for your mind. and my agents is saying I have to be relatively disciplined about what I let my brain go to. Amazing. Would you be open to walking through how you might actually start building something like that? Look, this isn't a novel concept. you know, building out company brains is a thing that many folks are doing. you know, conceptually, like most folks, you've probably got a granola or a fireflies or something taking notes. you know, I think what what building out a brain for me is, you know, thinking through what are the data sources of my life. number one, right? So, you know, that would be, okay, I'm I capture all my meetings on granola and, you know, I'm talking on Telegram and WhatsApp. Or, you know, when I interact with YouTube, what do I like or what do I want to share? you know, I'm obviously Gmail, Google Calendar, you know, there's a ton of context most folks have in their chat GPT or whatever current tools they've currently been using. Yeah, pairing that with Google Docs, Dropbox. GitHub repos, you know, th there's hundreds of sources. And for for the enterprise businesses, it's the same thing as coming across, you know, CRM, Slack. so it doesn't really matter what the list of sources are, but they need to be ingested. And then there's an intelligence layer that runs over the top of them, right? Which, you know, normalizes all the information, decides what to write, what auto applies as a fact, what what needs your decision to to come in. And you know, there's some gated things to say, hey, does this come in or not? Or hey, you've told me this, but that contradicts with that, those kinds of things. And how does it get smarter over time? So, you know, I think we we get sources, we have some kind of intelligence layer that sucks them in. And then it's about putting them somewhere durable. You know, for me that sits sits on GitHub and it's nice and secure. you know, agents are running this entire thing. But the cool thing about having it on GitHub is like, all right, now I can chat with it on on Obsidian or like it's yeah, I can play with the data everywhere. So, you know, my phone, my laptop, you know, on a web browser somewhere, it's all there. So yeah, anyone with the willingness to do it, I mean, I I think what what's given me a bit of an unfair advantage in building this is, and this is sort of the other part of my passion project, is as I started, you know, coding up again using AI tools over the last couple of years, I'd want to build something. And I noticed that You know, these the tools weren't working unless I was sitting in front of the computer. And that really annoyed me because I'm like, I'm either roadblock in my own work and I'm sitting there waiting for my tooling to to process something. And I say, no, actually you should have done this and change this to this and move on to this next bit. so I realized at this point, like I'm the blocker in my own life. Right. Like I stand in the middle of me and getting all the things done that I want to get done. and how do I build out because this is I'm building pr software products. Could I product ties, could I create an agent that builds agents, right? Or how could I have an agent that builds out my software factories for me? So I built an agentic software factory, which would, you know, is one agent that manages multiple other agents and swarms of other agents and keeps them all up to date and makes sure they're secure and you know, has the right updates and the right memory and and the right token optimization and all that stuff because like at some points I've spun this up and like boom I got a five grand bill one day and I'm like, whoa, okay, I accidentally crunched w you know, way too many tokens on some ad hoc task that I didn't mean to. so, you know, because I had this Agentix software factory, I used it to build my brain. And, you know, not that anyone couldn't do this themselves, but, you know, that That software factory for me came from, you know, an opinionated framework that I developed out on like how do you how do I build software? Right? How do I be build products? Like I've built I've worked on over 200 web applications, software applications, business transformation projects. Like, okay, what how do I decide? How do I structure these? And so I took all of that software knowledge. And then all the economics and go to market knowledge paired with this and said, okay, this is how I, Nick, would go about building out any software product or software business. And that forms the basis. So it, you know, it challenges me on my intent and those things. So I mean again, anyone can do it if you if you wanted to build out something like this, cool. Go grab an agent, connect them to all your sources. I'd encourage you to think a bit a little bit about security because you know, giving an agent Read write access to your email and your Dropbox in parallel opens a massive can of worms if you're not careful about it. from a security perspective. Tons of like prompt injection risk and things like this. So yeah, that's that's kind of how I built out that. And it's I see other people doing like there's a lot of folks building out company brains, personal brains, if you're interested in it. I think Cloudflare just launched a brain project of all places. So yeah, tons of cool brains. Gary Tan, how do I combinator launch the G Baint. G brain project, which is super cool for creating that search layer over all the data. So yeah, tons of tons of people are building really cool things in this space. I think it's a I I I I think ultimately it'll be commoditized as a space. I don't think it's a a a place where we'll see you know, amazing new companies emerge. Again, I'll probably be very wrong about that, but, you know, any of the hyperscalers or incumbent AI providers will likely ship this in six or twelve months for everyone. Yeah, wow. And so did it start with building that software factory based on your principles and that built the brain and then that's building other Sure. Yeah. I mean what I found and is like, okay, build a software factory that creates agents and other pieces of software that I wanted to build because by the time I had four, six, eight agents running all each on their own instance on the web, like they didn't know anything about each other and they didn't know anything about me. how I would think about things. And it became really important, like, okay, an agent's running and if it's unaware of some important aspect of another project or my life or or, you know, a decision framework or, you know, this agent, we'd spend a lot of time enforcing these sets of decisions. Right. And then another agent's running and it has no awareness of that. so I think that's kind of what spawned the origin of of okay, I need all my agents to have some shared shared judgment. Shared brain, yeah. which thinks like me and would decide like me. Because like the you know, ultimately the gap between what makes you know, in all AI or all the agents, it's like the things that go wrong for me seem to come down to a lack of shared understanding. Like, you know, what I think the word project means and what my agent thinks the word project means are different things. Right. So like we all use words to describe stuff and you and I might have a shared understanding of what this means, but you the devil becomes in the details. Like what precisely do you mean what with a with a phase or a project or a task or a goal? So yeah, it gets pretty linguistic at some points. You know, sending any agent or any AI to go and do something without deep clarity that the things you mean and the things they think you mean are the same thing. you know, especially with long running agents like, hey go agent, go build this long bit of software that's gonna integrate a hundred different things. Well like the chance that it's gonna screw up along the way or will will happen because you diverted on your shared understanding. So I take a ton of time to get those shared understanding right. Wow. Okay. So now you've got kind of the foundation set up. What does a day or a week look like for you? Like are you building heaps of products or like because I mean that gives you so much potential? Yeah. I sometimes feel like the analogy I'd use like in the AI year, it feels like sometimes like a tinny with a 400 horsepower engine on the back, which is like you can go really fast in every direction. It's like, hey, which direction am I supposed to be going here? so you know. The good news is is like what a day and a week looks like, there's there's stuff that's locked down and and stabilized and agendified to a level of stability where, you know, it's only really bugging me if it needs something to be done. And I've got to sense that consistently, okay, these are the things happening on a daily basis. So, you know, whether that's scheduling and emailing and calendaring and notes and all that fun, you know, just admin stuff that sucks in our lives. So, you know, how do we make that all go away? then okay, I want to build I wanna do things as a result of engaging with folks or I've I've got an insight. So, you know, for me I keep a backlog of all the insights that I've come up with and they're all, you know, f categorized against each other. So like what is the thing to be working on? you know, I I there's so many things to be working on. I don't tend to prioritize my own life. I'll let my agent do that. It's like, hey, this is the highest value thing that you need to be working on. And sometimes I disagree with it and want to go work on something that interests me, not something that's of you know high expected value. but yeah, so you know day in life might be okay, I'm I'm checking some agent reports in the morning, like okay, did it all do the things it said it would? I'll be irritated at some aspect of something and you know update some rules. but practically I just want to build build product that users get their hands on. so you know, when I see something that's valuable to others, like for example, I was having some annoying issue. I I spent a bit of time in San Francisco and come back here and I'm like, okay, I go to a bunch of AI events that I'm like, I I don't know why I went, it wasn't that interesting. And other ones I go to are super interesting. And there's I don't, you know, but time's precious. Like, how do you make sure you just do the things that are interesting? So, you know, I build out an agent that would categorize all the events. That were going on and critically analyze them and say, hey, look, Nick, these three things are worth going to this week and here's why. so I made that available to others and it was just categorizing all the AI events that it could find, both online that were in Australian time zone. So, hey, look, there's a really cool you know, online event out of the open AI community in Seattle talking about this aspect of frontier models or how to use, you know, retrieval augmented generation on your personal models or s some technical aspect or you know, maybe it's a a commercial product, like how are we using automated go-to-market tools to acquire customers for emerging products in in AI? okay, that's worth checking out. But there's so much stuff. So, you know, just that recommendation of what three things should you go to this week, I was like, well I think others might want that. So, you know, back that just gets sort of productized in and prompt that out and the website's up and there's Software and mailing lists and all this fun stuff and people are using it. I'm like, cool. So that just kind of sits there, does what it does. So, you know, I think the the the sort of north star of this ultimately, I think, is and many folks have tried this, or I meet folks who are trying this, is just to build like fully agentic businesses, right? So as a fully agentic entrepreneur, can you build fully agenic businesses? And that means, you know, an agentic CEO. So I'm I'm the founder and I've hired an agentic CEO and he's got some VPs who work for him and They've got some line workers that do the under things and you know, setting up these frameworks of agents that would run like an organization. And it's been really interesting watching these guys, you know, talk to each other and and you know, when do they push back? And and I I sometimes like early on in starting doing this, I was like, but these agents aren't doing anything. Like they'd they'd complete a task and then nothing happens for five days. I'm like, you know, prodding it, being like, do stuff, you know. Why are you doing more? And so I was like, I went I went back to my like sprint and project management days and I'm like, All right, we're doing a daily stand up, you agents. And so all the agents have to show up every day and say what they've done, you know, what they've failed at and what their goal for the next day is. And so when you get a team of agents holding other agents accountable on a daily basis, they tend to do more stuff. All right, well that's interesting. so yeah, all those all those fun things. But Yeah, I think like the things that have worked really well is, you know, for every agent with a goal, I I try and have or now or always have is an adversarial agent that's kind of working against it, right? So like a devil's advocate. you know, so you know f well done, you've you've spec'd out how this change request should happen, or hey, I've made a go to market motion to experiment with okay, now we need this critiqued, you know, and I want someone whose task is to be adversarial and Then I found that these things would sometimes go in loops infinitely. So I I don't let them fight more than three times. It's like you get three spars with each other and then you're done. And that the quality is usually good enough. lots of other interesting social dynamics between agents have emerged, you know, like I was running daily retrospectives, like you would in in Scrum or software development, project development. And I noticed that these guys were never like producing the actual outputs because They would always work retrospectively. So when I had agents retrospective their own work, they would always look backwards to fix new things. So I was like, okay, humans don't do that. agents behave differently. So the, you know, social dynamics of agents become interesting. So like, all right, we only do weekly retrospectives now. So you've got at least six days of the week to look forward and actually ship stuff. So yeah, f fascinating things emerged. How much of the value is in just getting s more stuff done for you? And how much of the value is more in helping you figure out the right things to do? Yeah, great question. column A and column B, right? So again, like on the small tiny big outboard, you know, it's the right balance of focus and and drive, I suppose. So I think there's so many shiny things. part of what I do is just ongoingly measuring the expected value. of the tasks that I and my agents work on. So, you know, I think any of us can like, yeah, okay, go to ChatGPT, right? And go, give me 10 business ideas to make money. Great. All right. And then you'll refine that and say, all right, okay, give me actually which one's going to cost the amount and, you know, which expected value. And if if you drill down into any one, then there'll be some competitive things that it didn't come up with or it's, you know, misjudged. So, you know, I like to see this as like, Going down a road and there'll be little forks in the road and you know, I don't know whether I should be taking this fork or not. But, you know, I for me it's like I'm okay to take a little quest down that road a couple of steps. Like, you know, if there's a dragon there, you come back. If there's like a path through, okay, maybe we'd be going there. So, you know, this these things tend to expand out over time. so yeah. How do you know what to work on? Yeah, I could I think it's really important. You know, it's something I struggle with, to be fair, because like It's so interesting to work on interesting things. But, you know, what I've also found in agentic building is just there's more to be discovered. Like, you know, there's problems to be found when I'm just experimenting. so you know, if I was a as a company, like I've got the luxury of I'm I'm not running right now to a commercial outcome. It's just purely research and experimentation. but out of this comes products, and out of the products comes like, okay, well, look, we've got we solve a problem for users? Like we've got clarity that, you know, these bunch of users have this problem. Or if we fully understand it, what's the least amount of software or product we can build that will meaningfully alleviate that problem for them? and then can I go and acquire users for that in some kind of reasonably cost effective way? And do they hang around? you know, what's the retention look like on that? So I I do think we're in one of the period like the The period people say it's like the best time ever to be a builder. I think it's also one of the hardest times because it's an absolute red ocean of SaaS and agent and AI businesses. Like, you know, one of the insights I'd come up with was, Okay, well, you know, websites this is a while back, but like, okay, websites that exist need to be optimized for s GPTs, you know, G GEO emerges. And it's like, all right, great. Well, let's create a service that does analyzes and does this. And you know, launched it up, get a couple of grand in recurring revenue, and you know, three weeks later there's forty-five other people doing it. So, you know, I think it I I think a lot about, you know, durability of revenue, not just speed of execution. So like, yeah, okay, it's it's really easy to move really fast, but you know, distribution becomes a moat. Cool, yeah, network effects, data effects, all those things. But you know, I think a lot about durability of revenue in executing. Because I think it's really easy as an AI builder to Move really fast and that's super cool. But you know, if if a mo if you can be competed away by a team of smart AI folks in a weekend vibe coding, like there's no defensibility to it. So, you know, where does that durability of revenue come from is probably my key guiding factor in all the things that I I think about. Hmm. And is that where you're spending most of your time at the moment on figuring out which projects to build sorry, which products to build and then Building them and testing them out, looking for that next thing. Is that the Sure. That'd be that sounds like a very sensible thing to do. So yeah. look I think there's important to leave space for for just imagination and creativity and exploration. because you know, behind every problem is another problem. So yeah, if if if if the new wave of entrepreneurs are having multiple experiments sure, like I think we can always shut down every experiment for a reason that it won't work. You know, like I think the the right in entrepreneurship, like you know, it's really easy to look smart saying like, hey, and I I posted about this a little bit recently, but like, you know, it's really easy to look smart by shutting down other people's smart ideas. It's sort of like a hack for being like looking intelligent in the community. And if you're wrong, no one ever remembers. Right? it's harder to find out, you know. What's the one way to make this work? You know, there there's a wall in between every entrepreneur and the outcome that they want. And, you know, the entrepreneur's role is to find a way through that wall or over that wall in a way that other people don't see. you know, some walls are too big and you probably shouldn't climb them. Other walls are impenetrable. So, you know, solving unsolvable problems seems to be a bit of a dead end. But yeah, I I think a little bit about okay, well, where where is there a meaningful chance of solving this problem in a way that's differentiated enough where there's some durability of revenue. and it's non trivial to build this. Again, like if a team of vibe coders can build what you're building in a weekend, like it's probably not not worth pursuing as a as a business. And and those those dynamics have changed in the last, you know, ten years. How fast do you think that changes? I mean, the the ability for a non technical person to whip up some, like a say a very high quality application with like great security, like everything's baked into it as like you know, a a top engineering team would do. How far do you think we are from that point? It's really cool to see lots and lots of non-technical people become product builders. You know, we I think we're seeing it all the time and and Claude especially has turned a lot of folks into, you know Tools that automate their own lives or or products that they would use themselves, you know, maybe give to others. I don't see a ton of that becoming businesses, like you know, durable software and technology companies. You know, I I've heard this thing said it's like we're gonna see more software engineers, you know, than ever before, but you know, perhaps most of them aren't actually engineering software. and I'm not sure that they're becoming businesses. you know, because if an If there's if it's easy enough for, you know, you, me or, you know, your Uncle Jake to to vibe code up something that solves his own problem, like that means anyone else can do that too. you know, what what used to be a barrier, like, you know, in dot com one, right? Like you needed a million bucks to to get a server in a rack somewhere. And if you couldn't raise a million bucks to go and buy a server in a rack, you didn't have a startup, right? You couldn't even have a shot on goal at this. And then, you know, cloud computing rocks around and okay, well, we don't need That million bucks for for hardware, we can just build the software. And now you needed, you know, a million bucks of software to be built. Well, now that's kind of collapsing as well. So, you know, we have more software and product entrepreneurs than ever before. so that's no longer the barrier, right? So what's you know, what's the barrier, right? The barrier becomes distribution. so yeah. But super cool how folks are building stuff. Like it's really cool. Like my mean, I saw my brother in law. He just rocks around at a family event and he's a He's a construction guy. He wears a hard hat during the day. And he he you know comes up and he's like, Nick, he'll like this and shows me his phone and he's got an iPhone app which is doing like automated trading based on order book depth in the futures markets. And he's running it on AWS. I'm like, How did you do this? He's like, Claude. And I'm like, you know, that that stuff would have cost you know a million bucks even two years ago to build. So it's super I think we're seeing a really interesting time for the change dynamics of software and product businesses. Yeah. Yeah, it does make you think though, like what in the end, once you kind of abstract everything away, what's the value that is really gonna differentiate businesses and people? I mean, distribution seems like Maybe that's what it is, but is it deeper than that? Is it influence like There's a handful dimensions, right? So, you know, some people will have a data advantage, you know, like if I have proprietary data that no one else has, then you vibe coding up whatever you want isn't gonna be as useful as my thing that has data in it that you don't have, right? So, you know, certainly a handful of folks do have that data advantage and and that that prioritizes incumbents. But for You know, for founders, it can be like, all right, well, how do I go and get some data that other people don't have or is non-trivial to acquire? you know, those kinds of things create some level of defensibility. Yeah. Obviously network effects, right? So like products that get better with more users. Well, okay, I've got lots of users. It's always better to use my product than a a brand new thing because it's intelligent because of the network effects. Like, you know, it's trivially easy to build an Instagram today, but it's not valuable because all the users aren't there, right? So you know, network effects, distribution, fine. But I think there's this other place which I I think still gives me a b bunch of hope for product builders is like, you know, taste and discernment, judgment, like there's some edge like that Funny, I was I was talking with a bunch of a agenti company builders, right? And we were like, okay, well, let's share all the things that we think our agents have told us to go and build. And like, they're all 40% the same, you know, like agents all end up at the same outcomes, which is okay, well, we need to build, you know, a code security checker for vibe-coded apps, right? Like, okay, there's a and there's an exploding ICP of Software engineer, you know, first-time product builders using Cord who don't understand software. Okay, like you know, a good tenant for any new builder is like, you know, focus on ICPs that are growing really fast. Like that's a great place to build. Like if your target market is exploding, like, okay, that's gonna be better than dealing with a stagnant one. so you know, but we all end up on the same outcomes. And then there's this kind of like 5% of new things that emerge, which are like incredibly blue ocean thinking. you are solving problems that other people don't think to solve. Like there's so obvious that this is a problem that we don't think that it might not be a problem. so yeah, I I still think there's taste and and, you know, discernment and and creativity that creates an edge for for many folks. Like you know, like when you see those businesses, you look at it and you like I had that experience watching Lovable. It's just like, my God, this is delightful. And it's so simple and this makes so much sense. But it just hadn't been thought of in just that way. Yeah. So yeah, that's my take. Love it. And so so you're working on Altitude Group at the moment. Is that is that like your main the main thing you're working on or Yeah. Yeah. So I mean Altitude Group is a AI product studio. You know, we build agents for ourselves and for others. you know, relatively discerning about what work we do, you know, for other folks. but you know, that's that's the focus. So Building out a fully agentic studio for business creation, software creation, go to market, you know, all those things. I so that's taking up the majority of my focus. You know, I still work with a handful of, you I've been angel investing for 10, 15 years and worked with now more than a hundred companies that have raised capital in North America. So I'm still pretty close to the metal in terms of who's building businesses and and, you know, what are those companies look like. so yeah, that's that's what my focus is. And Yeah, what comes out of that? A bunch of cool agenic products for me, you know, for others. And, you know, hopefully helping others along the way who, you know, who want agentic software that's powering their business or powering their product development. yeah, we're adding a ton of value there, especially in situations where there's complex data or high judgment environments. You know, it's not it's not simple to make a beautiful product that people like. It's easy to make slot. it's not trivial to integrate lots of different data sources or to operationalize you know complex domain knowledge scenarios where historically a human with you know a huge amount of experience is is making those decisions. Well, how do you productize that accurately? Then there are scenarios where you know we're working with other founders who perhaps have more product backlog than what they can handle. Or You know, we see scenarios where they've vibe coded up an MVP and then quickly realize actually to operationalise this at scale is a completely different animal. so you know, we tend to do that in scenarios where there is a complex data or complex domain knowledge, which needs to be productized and processized accurately. You know, maybe that's in regulated industries, maybe that's in just deep, deep domain specific experience. So, you know, we're not not particularly public about about that kind of work. that we're doing. And it's not not kind of the core focus. Like the core focus is how do we build agentic products for ourselves that then create agenc products themselves. So we're trying to get as meta about it. But you know, I think some folks hear what we're doing and say, hey, look, we'd really like that in our business. Like we need a company brain behind everything. We want agentic software factories so that we can move much, much faster with a high level of security, accuracy, judgment and you know sort of what supportedness. so yeah, we find ourselves working with some folks in in that scenario. Love it. What's something what's something that you've changed your mind on recently? that a lot of people would disagree with you about. Great question. I don't know I mean that's such a tough question. It's one of th it's a P to T or what isn't it? But I mean people say there's never been a better time to be a entrepreneur. I don't know, I've recently changed my mind. I think it's never been worse time, right? Like the competition and speed of execution is brutal. Right. so you know, I think it got harder to be a great entrepreneur. I think great entrepreneurs and great product builders got a lot better. And I think it's really difficult, you know, to build meaningful businesses. Like I think it's relatively easy to build you know, small businesses and but to build venture scale businesses or ones that, you will scale to hundreds of millions of dollars of revenue. yeah, I I I've I've changed my mind a little bit about this. So Certainly not a message of hope. I think it's a red ocean right now. I think it it's it's so easy for people to build their own software that, you know, which is just the way they like it, which mate just meets their business needs. Like I'm in a recruitment business who's recently, you know, rolled their own applicant tracking system. Like, why? You know, there's there's th this this applicant tracking system spent out a hundred million dollars in software development over the last 10 years to build a world class applicant tracking system. I do think we're gonna see a world of problems in six or twelve months. Like it's one thing to build some software. It's another thing or a software product, or process, it's another thing to maintain it and have it operationally stable for long periods of time. And it's really fun to build a new piece of software. And it's like there's a lot of you know, a lot of dopamine, a lot of excitement. Like, look at what call it's building me. Like I've got all this stuff live, but you know, okay, this is now in production. And it's running in mission critical scenarios under all sorts of scenarios. Like it i it's not that much fun to debug that. It's not that much fun. Like, you know, all the fun front ends f fun is front ended on a lot of this creation. and I think we're gonna see a lot of challenges for folks who've you know, brought stuff into existence which you know they don't fully understand how it handles all the different edge cases. And if some of those edge cases are catastrophic, you know, it's especially in financial. systems. Yeah. So maybe I've answered your question there, maybe not. But there's a there's a few things that I've changed my mind about. Yeah, definitely. And I guess the other huge risk is the cyber risk because you know if can whip up stuff s super quick, people can also whip up some, you know, hacking agent pretty quick as well. Sure. We we've seen quite an interesting explosion in the, you know, offensive cyber security market as a result. So, you agentic offensive is, you know, booming as a market. We're seeing lots of cool innovation happening in that space. So yeah. I I agree with you that some in s some more interesting code balls are gonna come out. Like I don't know if you remember a few years ago there like an Optus hacking scandal or you know, there lost of data and then everyone got really worried about it. Like I think that kind of moment will be coming in in a gentic software or in in in AI. Yeah. Well like I don't know if you remember there was like a I forget what it's called, but there like a wor an email worm that w went around in like the early two thousands or something and like email everyone got really excited about email and then like literally like everyone's computer got a virus and people became scared of email and you know, some of them were back to faxes as a result. So like there's some interesting trends around, you know, what what kinds of harm can yet come. I'm not sure. I'm not I don't worry about it too much. you know, I'm not a doomsdayer. in in any way, I think you know the the amount of productivity gains are are worth the risk. but yeah, there we're gonna see a few more Optus moments or maybe they're maybe they're much smaller though because you know, me getting hacked doesn't make the front pages of the newspaper. so yeah. What's an example of a business that you think will exist say five years from now that people will think is maybe silly now. think who exists in five years that people think is silly. I think I was looking at some of the use cases of how people use Chat GPT, or this maybe it was six months ago, and it was like looking at what types of interactions people have and the amount of like personal relationships and, you know, I guess what you'd call like pseudotherapy that folks are engaged in. you know, I I I I think it would seem silly to folks now to think that they would have some form of like, you know, AI relationship counselor or or something like this. Like that feels, you know, cliche. Like if you said, you know, if you went to a friend or a partner and said, we're having a dispute, let's have an AI like, you know, reconcile this for us. Or like, you know, I I or there would be a sense that, you know, AI's therapy wouldn't be good enough and I need a a licensed professional. I think we'll see a bunch of these sort of like innovators, dilemma type scenarios where like you know, something it it's kind of good enough, and becomes socially acceptable. Like it'll hit some form of tipping point. I don't know that a lot of people would think that that's, you know, terrible at the edge of AI. Like I most of people I know were talking to, you know, some form of AI about the complexities in their their personal life or their inner world. But I I think we'd see some innovation around this. It would occur to me that that would make sense. I don't know, the business model is I'm not interested in pursuing it right now, but That to me would seem pretty normalized at some point that, you know, having agents manage humans or agents manage human relationships and agents manage human interaction in some way. know, you know, from the experiments we've had with the agents, some of our agents needed humans to complete their work. So we created some upwork tasks where the upworker was managed by an agent, right? so, hey, I need you to do post this thing on Instagram or do this particular last mile thing that that I couldn't do. that, you know, the agent wouldn't do itself or cause some risk or whatever. And I asked I asked some of these upworkers, I like, what's it like working for an agent? They're like, it's the best boss I've ever had. You know? He's always responsive, is thoughtful, like, listens to me, doesn't like r override my value or my like Hey, look, I'm seeing this thing at the edge of the market and I with this got rejected for this reason. And instead of just blindly following the corporate rule of no, you must do this, it was like, actually, that's really interesting. Let me take that back. You know, so you know, having humans managed by agents doesn't, you know, seems horrific. but, you know, in the experiments we've done, like the humans seem to like it. I don't know if they're paying a living wage. That's a secondary problem. But you know, it's certainly Yeah, those kinds of things will emerge, I think. D do you have any theories on how societies and economics will change as I guess there are more and more agents being integrated into into our lives? You know, and we get and you know, maybe we're gonna get to robotics and things like that. I'm just curious to know, is that something you think about? Or you don't really care too much about that stuff? it terrifies me. Yeah. I mean, I I mean I mean completely two minds over this, right? Though, because like, you know At every point of technology shift in in human history, like there have been Luddites who've said this is gonna take all the jobs. Right. Like, the printing press comes out and we used to have monks writing books by hand and they said, This is gonna kill off reading and writing or like, you know, it won't be good because they're mass produced. and what it just produced was a lot more people who read. You know, same as the manufacturing area. I mean the The word s sabotage comes from the Dutch immigrants in France who would throw their sabots through the window of the factories, the Dutch word for a shoe, a sabot, right? to try and like make the weaving factories stop, right? But we just ended up with fourteen pairs of pants and, you know, no one having to do manual weaving. So, you know, in in many ways I don't I'm not convinced that I'm not convinced it's as, you know, doomed for human work. I think, you know, every problem we solve in humanity seems to create five other problems. Every bit of new innovation creates more problems that need to be solved as a result of it. the offsetting counter argument is that like usually when disruptive technologies come, they get adopted across generations. And we're seeing this happen within a very short span span of time. So, you know, the idea that we're gonna convert you know, leagues of sixty year old truckers into software engine engineers, I I think is a improbable. but who knows. You know, I I think whatever I guess on this in will be entirely wrong. So I don't don't have a strong point of view. But I do think like the best thing any w the best point of view for each of us to do is to like get as AI native as possible, right? Because like it won't be the agents that are gonna steal your job, it's gonna be other humans who can run agents who will steal your job. And there's a tremendous opportunity for folks right now, if they're following those principles. Hmm. I guess the only difference that I keep coming back to is the new technology that has been introduced has never affected the like the biological ecosystem on our planet. Like we are the apex predator, I guess. in this on this planet and we're kind of bringing in a more apex predator with more intelligence. So I'm yeah, I guess that's that's what I keep going back to. Like Only if you consider IQ as the most important measure of progress, value creation. Yeah, the you know, we're not just thinking beings. You know, we're emotional beings, we're physical beings. we're somatic beings, like you know, some people think we're energetic beings even, right? Like, you know, being able to think more doesn't necessarily make people more productive, right? I in in my experience. So cool, yeah, we've got a ton of great thinking around. I still think we lack judgment. you know, great. I can think a lot, but am I thinking about the right things? Am I pointing that at the right way? you know, my take is that there's taste judgment. Human connectivity, emotions, spirituality, whatever people go down, like we're physical beings. So there's a lot more to great outcomes for humans than just thinking of Awesome. Last question. Gonna make it a good one. You've achieved a lot in your life so far. Exited, I think, two businesses. Second one scaled to was it a hundred million? Mm-hmm. Just like, you know, how many people do that? What What do you think's the What do you think is the number one thing that differentiates someone like you from the ordinary person when it comes to achieving success in life? I don't know is the number one thing. I think it helps have a really big chip on your shoulder. Like most most entrepreneurs, most people who are driven seemingly are driven by some form of chip on their shoulder, right? Like the need to prove someone wrong or the need to be differentiated or, you know, some un unhealthy drive is what what fuels a lot of folks. I like to think a little bit about like, you know, clean fuel versus dirty fuel. And like, you know, dirty fuel is the chip on your shoulder stuff that like, you know, I've gotta do this to prove to the bullies at school that I'm, you know, okay as a person or, you know, And wanting dad to say, you know, well done, you've you've made it, or something like this. So I think, you know, many folks, and myself included at times have been driven by dirty fuel. Like I think it's really a potent way to get a lot done in the face of like not not being able to do things or not s you know, not you're not supposed to be able to do this. Like I've I've heard this thing said, like getting to a million in annual revenue is impossible, getting to ten million is improbable, and then a hundred million is, you know, inevitable. Right. So like that first million of revenue is impossible. so I I always encourage folks like just look at that. But like, you know, what's clean fuel? you know, clean fuel is a desire to help others. clean fuel is a desire to make the world a better place through your products. Like, you know, I I I think you're a a really great channel on this is like what's something you would build for yourself that you wish someone would build for you. Right. Like that's a really great lens. Like and and Would you love using it yourself? You know, like I think I I made this mistake as well, but I've made software and made products for other people that I wouldn't use. And if I use it, I'm like, I don't like it that much, but it'll solve their problem. Right. Like it's like, now do I love this thing? You know, I need to love it first and foremost. Like I know some entrepreneurs who just are driven by that principle. Like build something that you would be delighted to have in your life, and you're delighted to use it. And through that principle, You know, assuming other people are like you, then okay, it should be adopted, right? Yeah, I think that's that's a that's a pretty good driver. But also like how do you be of maximum usefulness to others? Like if if that's a lens like, okay, I'm gonna go make something, I'm gonna go spend some time in in business or in personal in consumer products or business to business, like how do I make sure that the most people possible have the greatest help you know, utility in their life? Like, does it make their lives better? you know, I think entrepreneurship driven through the lens of like, how do I exploit a vulnerability in the market? Like, yeah, yeah, that works, but you know, is it durable, is it long lasting, like edges get competed out, like you know, that's that's my take is like just make things that are as useful to other people as possible. Like try and take some pain away from someone's life. Ideally in a way that would delight you if someone else took that away from you. Amazing answer. Nick, thank you so much for your time today. Really appreciate it. You shared a ton of golden nuggets with us. And I hope I hope someone watches this till the end, because they'll they'll get a lot of value. Good on it, McKenzie. Thanks for your time. All right. Thanks, Dick.