.png)
KP Unpacked
KP Unpacked explores the biggest ideas in AEC, AI, and innovation, unpacking the trends, technology, discussions, and strategies shaping the built environment and beyond.
KP Unpacked
You Know Nothing (and Neither Do We)
The #1 Podcast in AEC.
This episode marks a first: KP sits down with Nick for their debut recording together on KP Unpacked. Funny enough, you heard their second episode (on the MIT AI Report) before this one because the MIT discussion was too timely to hold back. But here’s where it all started.
Expect raw takes, sarcastic banter, and unfiltered truths about startups, venture capital, and the built world.
Highlights from KP & Nick’s first rodeo:
- “You know nothing” - why KP tells teams to stop looking for answers from the boss and start listening to the market.
- Henry Ford’s “faster horses” quote and why KP calls it a dumb take.
- AI + robotics roll-ups and who’s really the customer when law firms and AE shops adopt agentic tools.
- The death of niche features and why vibe coding and no-code kill most startup theses before they begin.
- The illusion of being venture fundable and how KP decides when he’s out, even on good ideas.
- Why venture theses get stale fast and why copying YC or Sequoia is a losing move.
Resources / Links mentioned in this episode:
Sounds like you? Join the waitlist at https://kpreddy.co/
Check out one of our Catalyst conversation starters, AEC Needs More High-Agency Thinkers
Hope to see you there!
yeah, uh, excited to be a guest host.
Speaker 2:Thanks for having me on well, we'll see how you do am I the?
Speaker 1:am I the full-time? Have I been full-time promoted? We'll see how you do um, now I'm super, super excited to get into a couple of topics. For those who don't know me, I'm KP's partner at Shadow Ventures Anything you'd use to describe me.
Speaker 2:Yeah, and I'm KP. Hopefully you already know me if you're tuning in, but if you don't, I'm KP ready. Yeah, so we're mixing up things a little bit. Nick and I, I think, have some pretty cool conversations a few times a day.
Speaker 1:I think the cool thing about venture is it's just everything changes every day, so we figured instead of we would take our talks outside of our text threads and daily conversations and just like throw them out there for you guys to chew on Might as well. Just start recording all of our meetings.
Speaker 2:Yeah, exactly Good to do that Efficient way to do this, except for audio quality.
Speaker 1:Kids screaming in the background, dogs barking.
Speaker 2:I already had to text our nanny to not bother us during this conversation. I'm sure that's very relatable. Does everybody else is having to text their nanny to listen to that?
Speaker 1:Yeah, I feel like we can talk about a lot of stuff. Actually, I'm going to throw you a curveball. Okay, through uh with you today. That I think we're both, you know, pretty well versed on, like both in terms of the portfolio and um, just in terms of what's happening, happening and trending in venture. But the curveball I want to throw you is something that you said in a, in a meeting that I participated in.
Speaker 1:Maybe I was a fly, I was a fly on the wall, um, and I think it's a really it's just an. I think it's just an interesting topic. So let me find the, let me find the actual quote that you said so okay, here it is the number one thing that you need to know is that you fucking know nothing. Yep, that was a direct quote in a meeting this week. Do you recall the meeting? Yeah, I do. And you said to follow it up, you said we don't decide anything, right, the customer decides everything, the market decides everything. So I thought so let's talk through that, like what. I mean, you don't have to give the full context, obviously some private conversations.
Speaker 2:Sure, we'll talk about that, we just won't name names.
Speaker 1:We won't name names. Yeah, we'll just not name names. Like what did you mean by that? Why is that Like the way? So I'll give some context. The way you said it was that it was a non-negotiable for a team and it's a company you're running and it was a non-negotiable for the team that they had to essentially buy into this principle. And you said it very emphatically, like clearly you mean it, it's something that you know. Like someone disagreed with this. They need to leave the room. They might need to leave the company, right, what? Where did this philosophy come from? What does it mean? All the above?
Speaker 2:Yeah, I think it's well. My lashing out the other day and cursing was I get tired of being asked what should we do? Right, that somehow I have the answers. And I do have some answers, but I don't also think that I have all the answers. So when we start talking about kind of deep dive on product, on a feature, when you start digging into the weeds of like hey, we want to make a decision on and it could be as simple as the most simplistic world is, do I want to put this button to the right to the left? Is it green, is it blue? Right to the left, is it green, is it blue? All those things right. I think my frustration the other day was this idea that I have the answers and then the second piece of that is that the team thought they had the answers.
Speaker 1:Wait that you KP have the answers.
Speaker 2:Right, okay, hey, tell us what to do, tell us what you think we should do. Like it's kind of weird, I'm not the customer, right. And I think part of the frustration is, you know, when you look at businesses, you know that you build a culture. There's a lot of cultures that people come from, especially the AC industry. Is it about satisfying the customer or is about satisfying your boss? So if you grew up in an engineering company, it's about satisfying your boss, because they're the professional engineer, they're the registered architect, they're the principal of the firm, and so your internal customer is that. Your customer is actually the internal customer, which is probably your boss. So when you look at a lot of people that come from very parochial cultures where it doesn't matter what the customer says, it matters what your boss says, it's like a hard break.
Speaker 2:When I was in engineering, I would literally have bosses tell me I don't care what the client says, here's what I'm telling you to do. And so I think a lot of times in this industry, when you pull people out of industry, like in this case, there's definitely like hey, what do you think we should do? What should we do? And I think my point was I'm not the customer. You know we're not doing quarterly reviews with pay bank. This is not what this is. Quarterly reviews with pay bank, this is not what this is. You know, this is a startup and foundationally, I don't have to agree with you, you don't have to agree with me.
Speaker 1:We have to have consensus about what the customer in the market thinks, and that's all that matters. What's your response to someone who says you know the Henry Ford quote? That you know? If I asked people what they wanted, they would have told me what's the quote.
Speaker 2:Yeah, it's a dumb quote. They would have told me like faster horses. If I would have asked customers what they wanted, they would have said faster horses.
Speaker 1:Faster horses. So what, like? I think there is a.
Speaker 2:Nick. The second part to that is what did he give them? He gave them faster horses, Like at the end of the day. I mean, I'm not saying you take everything a customer says at face value, but you have to understand what is it? Because they may not be able to frame. You have little ones, I have a little one. They can't communicate what their needs are, they just cry. Sometimes we have to decipher what the actual need is right. So I think even in the Henry Ford quote, yeah, people couldn't, you know, they couldn't envision the technology right. So I heard a great quote the other day. They said they promised us flying cars. What did we get? And someone said helicopters, helicopters. Have we got helicopters? Like sure, it's a flying car, you know. So I think that the fallacy with that quote a little bit is I think that's what henry ford did get. He did give people um, is is faster, horses, just not horses yeah, yeah, yeah.
Speaker 1:Usually when I hear objections to that theory, it's usually around the idea that the customers have no idea what they want. Right, yeah, and so essentially what you're saying is like it starts with first listening to them and asking them, and without first having that conversation and understanding the playing field, you can't even start to either deliver exactly on the problem they're trying to solve, or innovate and deliver something that you know a steve jobs would deliver or that a henry ford would deliver, that is vastly different than what came before it and maybe would not be exactly what they requested, but still is solving their problem uniquely, um, you know, in a way that we could not before previously, as you know, as a civilization, yeah, I mean the first iPhone.
Speaker 2:The phone part didn't work on the iPhone, like your phone, your calls would drop constantly, right, and so, yeah, it was kind of funny you had to carry. If you had, if you were the first iPhone use customer, like I was you actually kept the second phone. It didn't work at the time. Yeah, yeah. So but I also think even you know that's why customer discovery is just such a life cycle process. It's just constant. It's just a constant process because sometimes you can't deliver what the customer needs are with what's available in terms of technology and innovation or at the price point.
Speaker 2:I don't think, you know, I always love talking about Steve Jobs because, like people like to use him as the epitome of everything that is right in the world and everything that is wrong in the world. And, you know, I don't think there is a meeting where, like I think, people will spend $1,500 on a phone. Like that is not what, you know, the vision was necessarily. I think, over time and iteration, you start to understand what the customer's thinking is and how far you can stretch their imagination into what's next. Yeah, I mean it's interesting. Like you know, our kids are going to be the most documented kids ever, my older kids. I have boxes and boxes of pictures and video cassettes that'll never. I don't know, I think about this a lot.
Speaker 1:Actually, it's so, it's so different, like the fact that we'll be able to instantly show them a video, when they're 18, of a video when they were one. Like it's. I mean, it's a, it's a weird concept, you know. I mean it's really, it's really neat. I mean we, I look at all the time and we do, we look at, you know, baby, baby videos and photos of my four-year-old right, like, but yeah, I, I, that concept will be, you know, the documentation as a result of, you know, improvements and essentially in chips, right.
Speaker 1:You know it didn't exist on the first iPhone.
Speaker 2:Yeah, Right, my first phone that had a camera. I didn't understand.
Speaker 1:Yeah.
Speaker 2:Why I was like, oh, this will be good for me to take pictures of my receipts for expense reports. That was my, that was my use case. I would have never thought like how we take pictures now, like I would have never thought. I never understood it.
Speaker 1:And, by the way, I'm sure you're aware of this trend that people are running away from Like creatively, everyone's tired of, like the perfected iphone photos, like people are buying you know um old polaroids and you know old, older style cameras to get the vintage feel, basically to do anything that's differentiated from what technology has now commoditized right, right and yeah and so um we can all take a perfect picture now, yeah, that's just, that's just boring.
Speaker 2:That's just boring, right? So you know, we've got. I think we're on our third what's that called a instamax or something in stack. Yeah, like on our third one, because they also are really good at breaking yeah, um, all right.
Speaker 1:So to close, this thought out, so let's, let's go back to the, to the quote um, the lesson is to listen to your customers and don't ask. Don't ask your boss, don't expect anyone at a startup to tell you the answers. What other takeaways are really important for this context?
Speaker 2:yeah, yeah, I think that's the point, though. Right, like use your experience, intelligence, creativity to infer things, right, but you can't just say like I've been doing this for 20 years. I mean, I could say that I've been doing this for 30 years, I know everything. I can say those words I'll probably I might get it. I might, I might be successful, but but I probably won't. I think that's one of those things. I've had a few calls today with some AEC executives that have these strong. They built some internal tech. You've seen this before. We think we can spin it out and get it venture fundable. And then you try to explain to them you're the only customer for this product. Have you talked to the market? They're like no, we know transportation engineering. Why would we go talk to someone else?
Speaker 1:Or they've talked to their buddies, their team, their colleagues, their peers at other firms and they've pitched them the idea, and the idea always sounds good, sound, you know, it always sounds good in theory and their buddies aren't going to shoot them down. So it's it's never. It's never true market feedback. Right, and I mean that is the, the unwillingness to, to pursue true market feedback. I think is like I think it's maybe one of the bigger problems. It's my, it's one of the bigger reasons. Startups and you know, at the stage we're investing in, it's maybe one of the bigger problems. It's one of the bigger reasons. Startups, at the stage we're investing in, it's one of the bigger reasons they fail. They don't actually want to go to market.
Speaker 2:No, and I think the other thing too is if you ask your employees and you're the boss, like, hey, isn't this a great idea? Of course it's a great idea, you're the boss. These are very parochial markets. And then I think, when you know, when you find these founders that have a ton of industry experience and are wanting to do a startup, it's all like what I call the pedestrian ideas. It's all these problems that they saw at work that they now think is a product. But the problem is, everybody else has seen those same problems. It's not really, you know, and they have chosen not to solve them necessarily. And I, and I think I think it's a huge problem. You know, I think when you have a lot of engineering brains in this industry, it's like I already know, I already know, and it's like maybe, yeah, yeah, the um, there's this.
Speaker 1:There's a segue into uh, I don't actually think this is one of our topics today, but there's a segue into the trend that you've talked about in previous shows. I'm a listener First time, long time listener, first time caller where PE is converging with VC. But basically the concept I have in mind is you go out, you build agentic workflows to plug in to existing stale, traditional workflow-heavy businesses. Think of an accounting firm, think of law practices, think of even, to some extent, ae firms are a big part of the conversation, right, like computationally heavy workflow, heavy industries that you know. Ai changes the cost structure of right, and you think about where the opportunity is.
Speaker 1:The concept of going out to acquire distribution right, you're going to make an acquisition before you actually organically sell into the market. Right, there's an interesting thing here where you don't actually like you're, by acquiring distribution, by making your first acquisition, you're actually not having to sell the product. Right, you don't actually know product. If you know, if the, if the end customers that you've built built this agentic tool for, if they actually care about it, if it's meaningful, it's kind of, uh, you know you're, you're building internal tools and, like, you're validating it yourself. Right, you're saying this is good enough to change the cost structure of this traditional business, but I think that's an interesting thing. That is not talked about is that you're acquiring distribution. You're going out to buy these practices and in some ways, you're circumventing the risk that the product does not get adopted. What do you?
Speaker 2:think about that idea. I think the idea, though, is, if you look at these models, you're buying companies and industries that already exist, so it's not really that you're on the front end offering anything different. It's really more on the back end, you're operationalizing it, so, as far as the customer goes. So the question would be like in that model, who is the customer right? Because, in many ways, the end customer their experience maybe gets a little bit better, but the end customer maybe isn't the end customer. So, if you're basically going in and buying a law firm and deploying agentic AI, isn't the customer really the lawyers right? So I think what you need to think about is you know, isn't the customer really the lawyers right?
Speaker 2:So I think what you need to think about is you know, that's the constant thing that can just like keep you awake at night is iterating who the customer is right, and I think these roll-ups and these strategies like buy a bunch of accounting firms and roll them up and put agentic AI on the backend, I don't think the end customer is the customer. I think it's really the partners of that accounting firm or the law firm. They may indeed be the customers, because the end customer doesn't know any better. They're still getting their drawn plans and specs right. They don't really know anything. So I do think maybe that's the shift in thinking you have to approach. If you remember, like at Summit two years ago, three years ago, our friend Tim Sheehan, the CEO of Greenlight, he made just one big point the entire time Distribution is all that matters. Yeah.
Speaker 1:Yeah.
Speaker 2:It's like we did a deal. I think it was like JP Morgan or somebody, and he was like we didn't even know if we're going to make money on the economics. Distribution is all that matters, right, and there's something to be said to that, becauseto-market motion, the customer feedback that you're, you know and to take it back to the quote that you're begging your team to figure out.
Speaker 1:It's like somewhat of a way around that, but the actual measurement of listening to the customer is listening to how effective the tools are, listening, noticing, observing how they're getting adopted, and the metric is more about adoption and uptake internally, and then you know. Ultimately, the metric is like how do your costs change? That's the promise here. The outcome is like do your costs improve? Are you, you know, are you operating more efficiently as a firm with AI?
Speaker 2:Yeah, Think about this way In the past, where I've done a lot of these corporate spin out deals and I did one for an engineering firm 2,500 person engineering firm and the funniest thing we built they built. We together built a product in-house and we had more adoption outside of the firm than inside of the firm right of the firm than inside of the firm right we couldn't sell the internal customer at all. We were able to sell the external customer better and there's some politics and culture. But I think in your point around whether it's the AI roll-up strategy, the end customer is the lawyer right, the partner lawyer. Your competitors may be a paralegal or a recent law firm. You know law school grad is probably your competitive aspect, so you have to outperform paralegal or a recent grad because your customer is probably really the lawyer in the firm. Yeah, yeah.
Speaker 1:Cool, all right, curveball over. Yeah, yeah, cool, all right, curveball over. We can, we can, we can take it uh to the traditional uh approach of the show, which is discussing your viral sometimes linkedin posts. Yeah, I don't know if this one went viral. Yeah, it seems that I'm. I'm pulling it up now. It seems to have some good feedback. Yeah, um, all right, I'm gonna read it, I'm to read it to you and then we'll talk about it. Okay, sound good.
Speaker 1:There has been a macro trend to vertically integrate businesses. I was asked by a large engineering firm CEO how did we go so far away from the master builder concept? My response, number one buildings got more complex. Two the world got flat. It was possible to be a specialist expert globally versus a generalist expert locally. Ie, a design structure you know that was super tall, that you know where you could service and build this super tall building globally, versus a full service engineering firm that can do everything locally. Number three risk aversion and lack of single point on a project. Hence the owner's rep emerged as essentially the point of contact for the owner.
Speaker 1:I think this over-specialization and fragmentation created even more silos Between design and construction. It was not typical for there to be 100 plus companies executing work. I think AI and robotics is going to drive vertical integration back to the master builder. The new master builder is the inference and orchestration model managed through MCP. There's a lot to unpack here. Let's start at the top. So, like the three points, like the vertical integration concept, the master builder, we went away from it. Let's talk about the three points. So first, buildings got more complex.
Speaker 2:Yeah, I mean, if you think about buildings these days, hva systems are more complex. Think about like maybe the elementary school that you saw was like cinder blocks, pretty simple systems, same floor. So we started dealing with much more complex systems. Now we have AV AV is a standard For a while there. I think they still do it Cat 6, low voltage security systems, automated systems there's just so many new systems that go into a building besides just concrete and steel.
Speaker 2:So buildings got more complicated and then people kind of started making them more complicated. You know, we started realizing like oh, we don't want all our elementary schools to look the same. You know, we're going to design this one to look farmhouse style because it'll be better for the experience with the kids, right. So we went away from those ideas. So design got more complicated, the products and the needs got more complicated in terms of how we look at buildings. So you had that right.
Speaker 2:But I think when I was a kid, my dad would go to saudi arabia to work on a project for like a year. Right, he was gone for a year to work on one project and would fly in every once in a while, not like we do now, right, but? But when I was a kid, like the idea that you flew somewhere for work was not like how we do it now, and so what it meant is, if you're an architect and you're the local Atlanta architect, you did everything. You did schools, you did hospitals, you did everything that was needed as an architect to serve that community. You didn't get to decide like, I'm just going to do schools, I'm just going to do airports. Same thing would happen with construction and engineering. So when you have to serve a hyperlocal market to make a living, you kind of have to do everything. Well, airplanes showed up, buildings got complex, you hired specialty consultants for everything you know. You look at a job now there's an indoor lighting consultant and an outdoor lighting consultant, right, you have landscaping, you have sprinkler, I mean there's so indoor lighting consultant and an outdoor lighting consultant, right, you have landscaping, you have sprinkler. I mean there's so many consultants on these projects now.
Speaker 2:But it used to be that, actually the AC, if you think, if you talk to the old guys, right, they'll say, man, back then you had to know everything, which means you could coordinate everything. Because you had to know everything, because there was like this master builder concept, and so I think you know when the world became flat and people could fly around and use telephones and you know email and email comes along, you can email drawings around that's when people realized if I just wanted to be I mean you can make a living as an outdoor lighting consultant and build a nice company because you'll operate globally you can be the outdoor lighting consultant for stadiums and just make a living doing that. You just zip around and work on every stadium project and build a nice business that way. So I think that's been kind of.
Speaker 2:The overall dynamic is that everybody got very specialized and could find niches. Think about what's happened with venture. Right, we're thematic, we have stages, we have. You know, everybody has a flavor. If you go back to the beginning of venture, it was just like whatever.
Speaker 1:We're going to invest in semiconductors.
Speaker 2:Right, what stage? What stage Like? Who cares Like? That's not the point.
Speaker 1:Yeah, you know now we?
Speaker 2:yeah, I think we've. Just, you know, like most industries, they become kind of over-industrialized.
Speaker 1:Yeah, the last piece. I think AI and robotics is going to drive vertical integration back to the master builder. The new master builder is the inference and orchestration model managed through MCP. First, clarify what MCP is and when you say inference and orchestration model, give people some context there.
Speaker 2:So what happened in the labor model? Right, Everybody got specialized but nobody was in charge, right? So then owners were like hey, I'm dealing with this architect, I'm dealing with an interior designer, there's 100 people. I don't know how to manage all this. Somebody needs to be in charge. Hence the owner's rep showed up and said and generally early stages of the owner's rep were like retired architects and engineers and contractors, you know, maybe there were the project executive for schools at a school system. They retire. The school says, hey, you're pretty good at this, why don't you help us manage?
Speaker 1:this, the natural consultant, motion Part-time. You're an expert in the market. I'll help you out and advise you.
Speaker 2:You've been doing this forever, would love to have you come help me on my next project. So that's kind of the origins of a lot of these folks. But then I think what happened was when that business started getting bigger and bigger, the talent level went further and further down, right, because you didn't have enough. There weren't enough retired people that wanted to go work to do this stuff, so the expertise kind of went down, right. So then, once again, then we went through an industrialization cycle, right? So, oh, we're going to build this industry, we're going to hire people straight out of art school and make them nothing wrong with art school, just don't manage my building project, right, and come in, you know, and I was talking to someone today and I said it's one thing to say, hey, we hire people that don't know how to read drawings. It's one thing to say that, right, okay, we can't find the right people, whatever.
Speaker 2:The problem is the point of view that you don't need to know how to read drawings to manage a project. That's the problem, right. It's one thing to say like, hey, you know, you don't, you're not born reading drawings, right. But to say that, like, it's not important, it's like you know in VC, right, it's like you have to understand finance. You don't get to just understand VC, you actually have to understand finance and math. Right, it's like I'm a VC I don't know math. Like, is that really going work? You know um, so I think, and by the way.
Speaker 1:There's probably vcs out there that don't know math, um, but you know they hang around for what it's worth. Vc, vc. Math can be very simple.
Speaker 2:I mean, I'm a vc yeah, but I think there's some people that just think they can hang around the hoop at yc and do deals right.
Speaker 1:Yeah, there's like the hustle culture, VC that doesn't run the return math. They just get caught in the hype cycle and the momentum trading aspect of VC for sure.
Speaker 2:Yeah. So I think, when we think about why ownership showed up, it was because it got very fragmented. So now fast forward, right? So if I'm an architecture firm and instead of hiring an outdoor lighting consultant, probably just ask chat GPT, here's my drawings. Please design outdoor lighting for me. One there's no stamping of outdoor lighting plans, right? So there's no regulatory reason to send it out. It's just that we don't have enough people around. You know, I don't want to keep someone full on staff that does outdoor lighting and I barely use them, yep, so now I can use. So I think, if we look at all the specialization, how much of that specialization, specialization can now be handled by AI? So therefore, I can take it back in-house.
Speaker 1:Yeah. Yeah, so it feels like it Feels like a breakthrough. I mean truly. Yeah, Feels like it has the potential to bring us all the way back, which is pretty wild, actually, when you think about the amount of entities that are involved and highly specialized today.
Speaker 2:So think about why do you hire an outside consultant, competency or capacity, or both?
Speaker 1:Yeah, that's the driver right.
Speaker 2:Yep, so AI, I have capacity right. Technology has given us a lot of capacity. What AI has given us now is competency right, and we haven't seen that before. We've had capacity. So now it's like why don't I have someone in-house? I can't keep them busy enough. Or when I'm really busy, they can't handle all the work. That's why I mean structural engineering used to be part of architecture, but then they said, oh, we're hiring all these engineers and they're sitting around when we're slow and then when we're busy, we have to turn down work because we don't have enough of them around, right? So that capacity and competency thing. So I think if you say the AI model now AI allows us to both have capacity and competency some of these niche consulting things get taken out.
Speaker 1:Yeah, that's why we're here, that's why we wake up every day.
Speaker 2:Yeah. So how's robotics any different? Yeah, Right. So GCs used to self-perform. Now they don't Right, they don't really self-perform that much. There's a few of our friends that do. Maybe they do concrete or something like that, but they generally don't self-perform anything. They manage other people and so, hey, I couldn't keep all my drywallers busy, so I just spin it out and let you know I'll use whatever drywall is available. Once again, I can have the capacity and competency and do lots of work. But what if I can buy or painting right? What if I can go buy a fleet of Okibo's painting bots? That's no different than having work trucks right.
Speaker 2:In fact, I've seen some of these work trucks. I think you might be able to buy an Okibo robot cheaper than a work truck. I've seen some of these dually F350 things.
Speaker 1:They ain't cheap. Yeah right, actually, I actually think that checks out. I'm running the math in my head, that's all right so.
Speaker 2:So you think about that, like, why would I buy a work truck? I can just buy an okibo robot. Yeah, now I'm in the painting business, now I have capacity and competency, yeah, so that's where I think that happens. Now, when we talk about, like the inference and orchestration right, what does a GC do they coordinate? What does an owner's rep or sometimes an architect do? Does they orchestrate? Right, it's coordination, we call it that right.
Speaker 2:So in our world it's more like orchestration versus coordination. So the robots have to have someone to talk to that says, hey, don't go do, don't go start painting because the drywall robots aren't done, right, so that's your orchestration. Probably some big antenna in the middle of the job site with AI and cameras on it saying whoa, whoa, whoa, back off Okibo. These guys have to come in right, just like what the superintendent does, running around. Same thing on drawings and design. Similar thing right, there has to be an orchestration engine. Maybe it's the owner's rep, maybe it's somebody else, but when you think about that, so all of this now starts to map out to and I talked about MCP, so it's model context, protocol, right. So if you know APIs and you know, hey, json software integrates and Google it if you don't know what I'm talking about. But you know the idea that two pieces of software integrate right. So now with MCP we're no longer talking about just data integration, we're talking about agents. So when we think about like an MCP server, Massive quantities of agents.
Speaker 2:Agents, right? So let's say, tomorrow you start vibe coding, right, and you're like man, I built this cool little tool, right? That grabs all the light fixtures off of a PDF drawing. I figured out all the symbology for light fixtures and I built this little agent and it can go out and grab all those things and come up with redesign the spot, make sure that the lighting is designed properly and build a lighting plan, right, yeah, what do you think you can charge for that? Probably nothing, right, probably nothing, maybe a little bit. But instead you plug it into MyMCP. So when it's orchestrated hey, what's going on with the lighting plan? It's going to ping your agent and say, hey, next agent, what the hell's going on with the lighting plan? Here's what it is, right. So. So that's how you have to start thinking about this. So, just like we think about orchestration, these mcp servers are really fascinating. Um, I think I'm going to be launching kind of an open source one in the next 30 days. Um, where people can just build.
Speaker 1:Yes, I didn't know that yeah so you learn something new every week.
Speaker 2:For AC in particular, yeah for AC in particular. Sweet Love it. That's why I'm getting everybody vibe coding man.
Speaker 1:Gotcha Always a grand strategy behind every small tactical thing you do.
Speaker 2:Vibe, code your stuff, plug in and strap in. That's the game, right? Yeah, so you just got to get Barry Barry showing up All this stuff, but so so I think that that, and what I love is the argument about these ideas. I tell people like you can't argue with me about if you can argue, when, but you can't argue if. I mean, look at Ahmed at Lumina, right, Look at where he's at now, where he's come from. His development and innovation cycles just could be faster and faster. Right, Because a lot of what people forget, like AI, you're going to write a lot more code.
Speaker 2:I was talking to a guy that's building a humanoid robotics company solo out of his 650 square foot New York apartment and he's basically 3D printing things at scale and buying like little actuators and all at scale and building mini humanoids Insane, and it's part time. That's insane, Right. I was just like dude, like, quit your job, We'll give you some money, Go's insane, Right. I was just like dude, like, quit your job, We'll give you some money, Go do something. Right, it's like, because those are the type of guys you want to back. You know a guy that's like side hustling building humanoid robotics robots out of a 650 square foot apartment.
Speaker 1:That's pretty wild yeah yeah, yeah, it's like, it's like the great there's. I mean, there's a term in AI research called unhobbling. It's like different milestones that you become unhobbled, that you unlock over time as a user and so when you think about the physical world, like humanoids have been largely hobbled because of you know many reasons, but, um, one is like just pure expertise and time and, like you know that you would have to devote to learn how to, how to, how to build, how to code a humanoid robot. But this guy's able to do it in his free time in his new york, in his new york apartment, because of the knowledge that he's able to find. You know, leveraging the llms today. That's insane, I mean, yeah, it's insane. It's able to find. You know, leveraging the LLMs today. That's insane, I mean, yeah, it's insane.
Speaker 2:It's insane, right. And so you know, and I think we've seen it like on the venture side, you know I probably have 25 meetings a week with founders, like high velocity stuff I do. It used to be like, hey, five are interesting. Yeah, how much deal flow are you seeing from me these days? It's not because I'm not meeting with people, it's just like not that interesting, right, it's like, hey, like, oh, that's a cool niche product You're going to do. You have a Revit plugin that does lighting design for hospitals. Somebody can build that. They can build that themselves, right, you don't need to go buy niche software or like a feature software. You should build the feature yourself. And so I think you know, for those that are startups that are listening in, if your product feels like a feature and you think you're going to grow into a product better, either do it quick or just be done. You know, open source your product, maybe just open source your product, maybe just open source your product.
Speaker 1:Yeah, the traditional vc advice if you go to you know a lot of the accelerators, the ycs of the world is like build a simple mvp right.
Speaker 1:Yep, leanest product you can possibly build for as cheaply as possible.
Speaker 1:Doesn't need to have perfect design, doesn't need to be comprehensive in any nature, but you iterate with the MVP over time and you know, gets you to, wins you a first market and gets you, you know, a certain amount of customers that get you to the next phase.
Speaker 1:I think that strategy today is really hard to pull off. Like starting with that first market and iterating and iterating and iterating, because the competition at that at that phase now the amount of people that are building really tiny point niche solutions in a in a really minimum, minimal amount of time is like, I mean, we're just that's the you know deal flow question. Like we see a lot of small niche products that are actually solving a real problem or interesting right, but man, like so saturated and the ability to build it is like you don't need really, um, you don't need, like, a background in computer science. Like you can learn how to code, you know a lot faster. Like you do need to code but you don't have to have this like extreme background and knowledge on how to build really complex software.
Speaker 2:Yeah, no. I talked to a group today that I said you know, I was like look, do you think you're venture fundable? Let's be, let's be honest with ourselves. Should I invest in you or in lovable Right Like you? You tell me if you're venture fundable and people are like well, the, the bar keeps changing, the goalposts keep moving. I'm like yes, 100%. I was like and I love using this example that I charged Blue Nile like $5 million, some insane amount of money, to build their first e-commerce site in like 1997. A $99 a month Shopify account has has more functionality and they were venture funded to sell jewelry on the internet. Would you ever venture fund a company today to sell jewelry over the internet? And it's like no, like that at that moment in time, sure, today, absolutely not. So I think the same thing is is you know what, you know why?
Speaker 1:because shopify exists right, because shopify exists they come out, they commoditize the ability to go build that, that in that infrastructure, to sell goods over the internet for any individual right.
Speaker 2:So why am I gonna invest in you know, in your software that someone can code and rep replit pretty quickly and I think, look, I'll just like vibe coding.
Speaker 2:Someone can code and replete pretty quickly and I think, look, all this like vibe coding and no code and all that. I think in our industry is just more and more applicable because people buy software that they're using 10% of. They're using three of the menus on there, but they were sold to Bill of Goods. You can do all these things. So they're building, they're using 30% of it. Once they realize that, like I can just build my own stuff that actually serves a hundred percent of what I need, or they're using nothing, oh, we're using a bunch of spreadsheets. Great, I can vibe code something, get it up to speed, and I can have my own software right and I don't have to do a spreadsheet anymore. And I can have my own software right and I don't have to do a spreadsheet anymore. So I think it's a very rare situation where companies in our industry are using 100% of the software they purchase.
Speaker 1:Yeah, second topic you ready, yep? The VC thesis lifecycle is experimentation, pattern recognition, institutionalization of the pattern by others, then failure. If you don't experiment and identify the pattern yourself, then you are at a huge disadvantage. Borrowing other people's VC thesis is not the way I actually really like this one. I think there's a lot to talk about here. But yeah, unpack it a bit. What are you meaning by that?
Speaker 2:I think you know. Um, once again, much like the previous talk topic, venture capital has been kind of institutionalized. So people say, oh, it worked for yc, let's build a similar yc model, right. And then there's, there's a thousand clones out there and so, whatever that pattern was the the early I haven't dug into the early days of YC, but they clearly saw some pattern through some level of experimentation and said, hey, this works, what we're doing works. But then once everybody knows that it works, they go off and build it, and so once everybody has it, it no longer matters anymore. You have to move on to the next thing. So the YC model is not going to work forever. That's just not how life works right.
Speaker 2:So I think it's really tough, especially for us being kind of smaller allocators, to kind of you know actually I don't know maybe it's smaller allocators it's easier for us to be creative, and maybe it's easy for you know General Catalyst and Andreessen and Thrive to be creative because they're very large allocators. Maybe it's like most markets the monkey in the middle is the one that struggles. Right, because we can be. Well, we can do something. We're talking about something this morning like, hey, let's just go do that. We can just, we can just do that. Right, we have an investment committee right here on this call yeah, I mean, I think we, I think we um.
Speaker 1:I think the you know, the question for us, or the reason that you know we maybe don't have the luxury of, of, uh, of, of waiting, is that you know, for us, creativity is survival, right, like we have to be. We have to be super scrappy and hustle to find, to find a niche, to find founders before other large firms, you know, recruit them and chat with them and we have to show. You know, we have to compete with everyone else, right? So, like the big firms, even you know the types of firms that have the luxury of really large AUM and balance sheet in the right deal, we're still competing with them. So, like, how in the world are we competing with a founder's fund or a general catalyst in a deal? And the only way we do that is, one, reach them earlier or two, add value in a unique way that others cannot.
Speaker 1:Um, maybe a third that I would say from personal experience recently is show like a really unique conviction that you know the large firms like they, you know they're in all the hot deals, like if we're, if you're, if we're really convicted in a deal and we really, really believe in it and we show up every day, and that comes through in the conversation we're having with the founder. I think there's a like we've seen, like we get, you know we'll, we'll get it, we'll get carved out and we'll get it. We'll get a spot on the cap table purely by showing that conviction. It's not something you can fake, so like, like for what it's worth. There's been deals where we haven't been, you know, have had room carved out for us, because maybe that conviction didn't come through.
Speaker 1:So, yeah, I think we're actually in a spot where, like, we have to do what your LinkedIn post says, you know, adapt the thesis constantly in order to one figure out where we can actually make money and generate returns while still competing with the best firms in the world. And then, to, you know, get in. Even if we have access and we see the deals that we want to invest in, like, how the hell are we actually going to get in? Get it, you know, get in, get it, get on the cap table.
Speaker 2:So think about this next year. Your first curveball question, right, who is our customer? Who should we be listening to? It's the founder, right. And I would say more and more I'm realizing, like founders, that I can kind of intellectually spar with and have good answers. They have good answers and have conviction around their good answers. They're also quick to say no, I don't know. Man, they're also good at I don't know, but when they have the answer they have a lot of conviction around. In other words, that shapes our thesis. It's that experimentation, it's the sparring with founders to understand what but think about it a lot of times which we get to evolve our thinking. That's the other benefit of being boutique. We get to evolve our thinking quickly and we don't have to ask permission as we evolve our thinking.
Speaker 2:Think about like you go pitch a startup, a VC, and they say, hey, we're a pass because it doesn't fit our thesis. Right, where does your thesis come from? It came from the hypothesis you had three years ago when you raised your fund. That's right. Right, because at the end of the day, I don't think any of our LPs are going to say, hey, I'm really mad, you guys invested in something that didn't meet your hypothesis. I think they're more interested in us just making just lots of fucking money right, like that's what they care about. So I think I mean just like that. I mean we've talked about some of our portfolio companies that today we wouldn't invest in, right, at that time it made a lot of sense Today if they walked in the front door. But I think this idea, I mean I've, I've, I have found myself. I've started catching myself when I tell people I'm a hey, this isn't a good fit. I used to say it doesn't fit our thesis. I stopped saying that. I'm like.
Speaker 2:I don't like. I'm like, I don't like it, I don't see where it's going.
Speaker 1:Yeah, it's a cop-out answer.
Speaker 2:Right, I'm like I don't think you're the right guy. Actually, I like the idea, but you're not the right founder. You're not a good founder fit for me.
Speaker 1:Yeah.
Speaker 2:Right, you're not a good founder fit for me. I don't think I'm a good fit to you. You'll probably be successful, but probably with someone else's money, not mine.
Speaker 1:Yeah and um, there's a, there's a firm called um abstract, the VC, and I I recently actually don't know much, much about them which was, uh, it was interesting to get. Um, I listened to it like a deep dive podcast. Um, invest like the best with the founder last week and it's a really interesting um, his, so his thesis, I think like jives really well with your LinkedIn posts, which was like he was he, so his story is fascinating. He was basically broke, bankrupt, failed as a founder and started abstract with, with nothing. And so you think about like how you, you know, how do you, how do you break into Silicon Valley if you have, you're bankrupt and have nothing, you have no money. And it was through his thesis.
Speaker 1:His thesis was I've noticed a pattern where, if you come from a breakout startup winner, the label would be a unicorn. You know the label would be a unicorn and your shares vest for whatever reason. You leave that. You leave the company, maybe post IPO, post acquisition, or even you know, just a bit early, and you come from, you come from the pedigree of one of those companies and you leave. That is the best predictor of outlier success and venture, because you've seen what it looks like to build an outlier outcome. You've been a part of the team, you did it actually, and those you know so. So the uniqueness of that background, I think coupled with there, was, like maybe some, maybe there was an educational background, like I can't recall exactly, but maybe, like you know, they had an Ivy league background and then they, and then they did that.
Speaker 1:But he formed this really specific thesis like that and basically created a way like think of an algorithm to get alerted and notified when someone would leave one of these big firms and they were in stealth it was clear they were starting a company he would get pinged, he would reach out, would cold email them, cold call them and he'd get meetings because founders at that level, they need to raise money. They're not actively trying to do it, they're in stealth. So he would meet them before everyone else and then he would also, at the same time, he was networking the backend with other VCs, sharing this thesis and showing them really high quality deals, because his, his thesis turned out to be right. Yep, um at those.
Speaker 1:That is maybe one of the best predictors of success. If you've seen a unicorn and you've been a part of the team, you can then go out and do it yourself, and so the quality of deals that he was giving really well-known GPs in Silicon Valley was somewhat unmatched. And they recognized that early right and they backed him. He had Mark Andreessen and a lot of other really well-known GPs in his first fund, which from experience I know that's like extremely impressive. But the thesis was like pattern recognition that that trend and that pattern was happening and you know, then he institutionalized it and now he has like one of the bigger seed funds in the US.
Speaker 2:But do you think if he does his next fund he's going to follow that same pattern?
Speaker 1:I think there's a lot of. I mean he, there's a lot of iterations on it. Now, right, one of the iterations that um that he brought up, that he has, that he did not do at the outset, was brand building. He says you know, vc today is um like he, lose you. You can lose a lot of deals if you're up against, you know, highly credible, you know firms with, with very great brand equity and pedigree, the Sequoias of the world right, very tough to compete with, even if you haven't, you know, even if you haven't in, and they've moved into the spot where they're leading a lot of deals and so that that lead position is hyper-competitive in the best deals. And if you don't have that brand, he has to meet the founder so early and establish so much trust and rapport in a really short window of time that he's recognized that not having a brand is actually hurting him in that time window that he has to build trust. So that's one iteration was like I need to do this now and I need to actually invest in this.
Speaker 2:Yeah, I mean, look at like was it 20 VC? I mean started off as a podcast, built a ton of brand, right, built a ton of brand, interviewing a bunch of VCs. But I would say what he really did is like if you're doing a podcast, if you're doing an episode every day and interviewing a VC or a founder or whatever, and if you're paying attention, you start to see something. You start to see, okay, this is this is what, what is working now and this is the type of people that are making it work. So I'm going to invest in those folks, yeah.
Speaker 1:One tiny iteration he made on the way on the recognized because of the GPs that were in his fund, he had unique access to fundraising feedback from some of the best investors in the world. So he would be like the interface between the founder and their fundraising process and basically solicit feedback from the best in the world, give it to them in a platter and then adjust, you know, adjust accordingly and like no one else was able to do that because they didn't have the access, he did yeah.
Speaker 2:No, I mean a hundred percent, but I think that's the point right, you can't. There is no formula. I mean I even like shy when founders ask you hey, what stage do you invest in? I'm like, what is seed, what is pre-seed, what is? I mean we've tried so hard to like once again institutionalize. This is what seed is, this is what pre-seed is, and I just kind of say like well, we tend to look at stuff that's pre-revenue or unscaled revenue. Try not to like put a label on it. You know, like, because I don't know.
Speaker 1:Has your definition of pre-seed and seed changed over time?
Speaker 2:I just stopped caring about it.
Speaker 2:I think the way I look at it is where in the development of the business do we think that we can understand it the best right and that we can be a difference maker to that founder right, not just through capital?
Speaker 2:And so I kind of look at it and say, you know, pre-revenue, sometimes formation, I mean we've invested in companies where people quit their job and you know we invested. And then I would say, you know, the question is like once people start getting into scaled revenue and they become very metric driven and it's kind of running I don't know how much value we add there. I really don't Right, like I think, hey, we're trying to get from 15 million in recurring to 100 million in recurring, sure, we could add some value, but I don't think it's as much value as we add like pre 10 million in revenue. Agreed, so I don't know. But you know we've built all this vocabulary in venture that I think even the founders are looking for some level of institutionalization. I was actually like playing around with my LinkedIn profile the other day. I almost like deleted venture capitalists and just replaced it with investor because it was just like I don't know, it just felt like, okay, that's like a bucket.
Speaker 1:You're tired of all the VC ridicule.
Speaker 2:Yeah, because I don't have a vest. We don't have Shadow Ventures vests, so it's the other VCs that ridicule me. How can you even call yourself a VC? You don't have a vest?
Speaker 1:Yeah, I expect to run into you at a coffee shop in san francisco one day and you're gonna be like patagonia'd out. You're gonna, yeah, be the prototypical bc yeah, one day I'm gonna catch you yeah, it was kind of funny.
Speaker 2:The other day someone said stuff like you're the only vc I know that it wears earrings and I was like okay I could, I could, I could probably validate that I don't think I've ever seen anyone with earrings. Yeah, I've seen some VCs with some ink. You've got some ink, but piercing is not so much, not a lot of piercing.
Speaker 1:I was never asked what it's like to fundraise with someone who wears earrings to the fundraising meetings.
Speaker 2:Yeah, and who also struggles to keep his beard in trim. Yep, yeah, but no, I look, I think that's a. That's a lot of what I I think the way. That's why I'm kind of excited about like all this ai and robotic stuff, because I don't think there's one way to build stuff, I don't think there's one way to execute on this stuff and I just don't want to you.
Speaker 2:I like that because then it's not like trying to create patterns and sell LPs on some pattern and then exclude founders because they don't fit the pattern. I think there's some deals out there that someone might think like hey, we're not venture fundable, I'm not going to reach out to KP, but maybe they are right, maybe they are, maybe they are venture fundable, but they've talked to 10 other VCs that told them like here's what you need to look like and they've told them no, based on that pattern. So I was wondering is, if I changed it from VC to just investor, do I attract actually outlier founders that maybe have been told no a lot because they don't fit the pattern of seed stage investment, and then maybe that actually attracts a different type of founder.
Speaker 1:Yeah, I still think that I think there's a way to communicate and I think you do this in your LinkedIn posts and your content there's a way to communicate that you think differently, but I think that the average founder is still you know, they're still searching mostly for venture capitalists, even if they're getting no's knows. So, like the you know, as someone who's incentivized to make sure the right founders find you, I think the label of venture capital is still applies, but I think, like how you communicate your philosophy and how and and and what you're excited about investing in has to come through everything else. I think the labels still is okay, but yeah, that's at least how I'd interpret it. Yeah, I mean, I didn't change it.
Speaker 2:so I think I label still is okay. But yeah, that's at least how I'd interpret it. Yeah, I mean, I didn't change it, I left it alone. Well, this has been fun, Nick. We should do it again sometime.
Speaker 1:Episode one in the books.
Speaker 2:Episode one in the books.
Speaker 1:We'll see if I get brought back for a second season.
Speaker 2:Although I'm the one pushing all the buttons on the recording, so we might have to do this all over again hopefully not, hopefully not yeah it was super fun. Thanks for having me later bye.