Stop Following the Lego Instructions

What if teaching kids to complete the Millennium Falcon set is exactly what's making them unprepared for the real world?

In this episode of KP Unpacked, KP Reddy and Nick unpack why AI reading drawings is a feature, not a company, why reindustrialization in Detroit changed how KP thinks about hard tech, and why the Lego analogy explains everything wrong with how we raise kids today. Original Legos were a mixed box of bricks with no instructions. You built whatever your imagination created. Modern Lego sets are Millennium Falcons with step-by-step instructions. Kids complete the set, lose their mind when a piece is missing, and never learn creativity. Sound familiar? College degree, job market, no pieces, losing their mind.

KP takes that analogy into AI: reading drawings is spell check, not a bestseller. Everyone's building tools to "read plans and specs" and the head of pre-con 10 minutes from YC is telling his team these founders have no idea what they're doing every time they leave. The hard part isn't reading the door on a drawing. It's knowing whether you need three hinges, the right finishes, or the shim dimensions based on decades of inference. Then KP shares takeaways from Detroit's Reindustrialized conference: own your building, run your own machine shop, stop outsourcing prototypes to vendors who put you at the back of the line. Antonio Gracias (early Tesla, SpaceX investor) said it best: stop making three SKUs for mass production. Make 15 form factors, release faster, do more interesting things.

Key questions answered:

• Why is AI reading drawings a feature, not a product or company?
• What's the difference between object detection and inference in construction drawings?
• Why does every stakeholder look at the same door on a drawing and see something different?
• What do original Legos teach kids that Millennium Falcon sets don't?
• Why are college grads losing their minds when pieces are missing?
• What should we actually be teaching kids instead of following instruction manuals?
• What happened at the Reindustrialized conference in Detroit?
• Why should hard tech founders own their buildings and machine shops?
• Why does outsourcing prototypes to manufacturers put you at the back of the line?
• What did Antonio Gracias say about nimble manufacturing versus mass production?
• Why do fewer SKUs and more frequency matter more than cost efficiency?
• Why is gaining understanding the actual goal of using AI tools?

If you're building an AI drawing reading tool and calling it a company, wondering why hard tech funding requires a completely different playbook, or trying to figure out what creativity and imagination actually mean in an AI world, this episode will challenge every assumption about tools, skills, and what we're really solving for.

Listen now.

Transcript

SPEAKER_03 0:00

And I there he is. How are we feeling this big KP?

SPEAKER_01 0:04

Let's go back to the chiropractor.

SPEAKER_03 0:06

That's what I asked exactly how you were feeling.

SPEAKER_01 0:09

Yeah, I hurt my shoulder. That's how I'm feeling. I'm feeling in pain.

SPEAKER_03 0:14

Did uh does it help going to the chiropractor?

SPEAKER_01 0:17

I've got a few people. Uh this one seems to be doing a pretty good job. It's kind of interesting. She has older equipment that has more time, patience, and experience than a younger person I was going to that had all the high-tech stuff. So I was just thinking about it as I was rushing

Shoulder Pain And Better Care

SPEAKER_01 0:34

back here to get with you. I was like, that's an interesting dynamic. You got the older experienced person with the less relevant, less new tools, right? And then you had the new young person with all the latest tools. Which is like what we're seeing every day, right?

SPEAKER_03 0:51

To reflect on one of your posts from earlier this year, you you posited the question of would you prefer a young contractor without the decades of experience but with all the best tools? Or an old contractor with no no great tools but has decades of experience and on on the you know on the job knowledge, test of knowledge. And it's an interesting question.

SPEAKER_01 1:15

Yeah. So I think I'm experiencing it. She seems to be doing a better job. Cool. Also, not in a rush.

unknown 1:21

Yeah.

SPEAKER_01 1:22

I don't know if that is a standard of care dynamic or an age thing. Like she's just very patient and takes her time. She was running late. There was another kind of customer waiting. She didn't, she wasn't bothered by that in a way that, like, hey, I need to make sure you're set before I move on to the next person, versus maybe the person that has an app telling them it's time to stop and move on to the next client.

SPEAKER_03 1:49

Yeah.

SPEAKER_01 1:49

I don't know. Well, it was interesting though.

SPEAKER_03 1:51

Some news in the industry industry this this week. I think yeah, no, you had a couple topics on your mind. AI is getting better at reading drawings, apparently. That's all everyone wants to talk about. Have you seen evidence of this, or is this is this just hype?

SPEAKER_01 2:08

Oh my god, everybody wants to talk about it. And the the thing is, like, I've been working on how to read drawings for 18 years, like in terms of using AI to read drawings. I've been on it for 18 years, right? And you know, you can call

AI Reading Drawings Gets Loud

SPEAKER_01 2:26

it object tech object detection, computer vision. I mean, heck, we were doing stuff, you know, 10 years ago with some built-in tools in AWS, right? So I don't think it's about reading the objects, right? But I think, you know, Devin was helping me with something the other day, and I was like, you know, reading drawings, so object detection is not the same as inference. Linking objects is not the same as inference. And ultimately, like back in the BIM days, we used to talk about diet design intent models and build intent models and all that. And and you know, the idea being is if you look at a set of drawings, it's not about seeing the objects. We we know the ob the objects are not hard, right? Because these are generated, these are vector generated from CAD drawings. Nobody's hand drawing anything. Sure, do you have some overlapping text everyone's like, yeah, sure, right? There's some set of computer vision challenges, but it's about inference, it's not about seeing things. And I think that's what's so interesting is like people aren't getting that. And I posted something, I was like, wait, so not a lot of engagement. But one person was like, hey, that's kind of funny. I was like, I was like, you know what contractors have been doing with architects drawings for since the beginning? It's called inference because it doesn't tell you what's

Inference Is The Hard Part

SPEAKER_01 3:54

it doesn't tell you everything, right? It's like just because it shows a door, just like it doesn't mean anything, right? Yeah, the inference is what means something. And so, you know, Devin was kind of coming up with this image for some investor pitches and stuff, and and I was like, look, if uh if all of us look at a a door image on a set of drawings, we all look at it and have a different inference. The guy installing the door is like, oh, I need three hinges, I'm gonna need the right shims to keep it up. Like, that's what that person's thinking about. Like, how many screws am I gonna need? Right? The architect's like, oh, what finishes? And I'm being a little bit like, you know, architects, what cut what brown is that? Is that the right brown? Like, whatever it is, right? But everybody looks at that same thing with an inference based on their point of view and the work they have to be done. And so I think it's just so funny, like everybody's focused on reading drawings and specs, and it's like it's not about reading, it's about interpretation. Yeah, and the problem with that is I'm an engineer, don't tell me how to read my drawings, don't tell me how to read my drawings, don't tell me how to interpret my drawings. I have a way of looking at things, I have a point of view because so much of inference is kind of interpretation plus experience. Right? So, how do estimators look at drawings? Counts, labor. They're really but what they're really doing is like what information do I need to put together to go ask my subs for a bid? Yeah, which which some estimators that's very little. They just swivel chair the entire set and say, Hey, give me your bid. And they're you know, collating all these bids, right? They don't actually do anything. But I think I think that's what everybody just so it's just interesting to me, like the the narratives. It's like these are all people that have never built anything. I was talking to this contractor, I went to the World Cup game Monday down in Santa Clara. Yeah, so much fun. What did you play? It was Jordan and Algeria.

SPEAKER_03 5:56

Oh wow, that was actually a really good game. Yeah, well, so much fun. There's a really fans are so much fun, too. They all I think the the Algerian goal was like the 80th minute or something like that. And there's I just saw the video of their fans going crazy.

SPEAKER_01 6:10

The crazy part was from a time of possession perspective, Algeria had the ball the whole time. Did they? That's what was wild about it. So the score I don't think was indicative. I mean, it's super interesting. You look at stats around possession, right? Do you need to possess? You know, if you can shoot on goal and score, does it matter how long you possess the ball? Totally, and maybe by possessing the ball so long you get tired. Like, I don't know. But so I was with these construction people, right? And we're just like talking about different stuff, and we went to the office ahead of time, and the the head of pre-con there was just like, man, and these guys are like maybe 10 minutes away from YC's offices, and he was just like, Do you know how many of these kids come in here doing customer discovery on us, showing us how they can read drawings? And he's like, Every time they leave, we're like, What are these people doing? Like just wasting our time, they have no idea what they're doing. And so I so I think it's it's it's an interesting time because there's you know, which by the way, also I think LinkedIn's algorithm has gotten really good. Oh wow, interesting because all I'm doing, all I'm seeing is feeds of people talking about how they're reading drawings. Wow.

SPEAKER_00 7:25

Like, I don't think that's coincidental that everything in my team is about AI reading plans and specs.

SPEAKER_03 7:30

That's a shift because I don't recall reading stuff or spending more time on my LinkedIn timeline and then getting an alg an algorithm switch of just seeing that sort of content. It's usually just like the most current and up-to-date posts.

SPEAKER_01 7:46

Yeah. Interesting. So anyway, I think it's I think it's really interesting because nobody's really asking, you know. So I think it's a good it's a good point though.

SPEAKER_03 7:53

Like, and I, you know, my my head goes to you're so you're coming on like, hey, I'm an engineer and don't interpret my drawings. Yeah, I I can read my own drawings. I don't need you to read my drawings. Every everyone, every you know, every independent stakeholder that you listed out, contractor, the architect, the sub, they're all thinking that way too in their own jurisdiction, right? So, but I but I think there's something interesting where, okay, like if we if we can if we can add their context and experience, like what we've talked about a lot on this podcast this year, of like, you know, an additional skills file added that takes into account what they likely have interpreted in the past and how they think about things, then you're you can at least get get an interpretation that's close to how they would think about it. It's not going to be perfect, yeah. But that might help a person not as knowledgeable that's in the home office, or even you know, an associate of the person in the field, an associate of you, you know, you're bringing up Devin. Like if you know, if Devin doesn't have your point of view exactly, like that might help him get further educated when he's making his, you know, when he's doing he's doing feature design. So like there is, you know, there there is a need for that sort of product, but yeah, if you're not like the the work, the bulk of the work is interpreting. So like if you don't have all that context, then it's like kind of irrelevant.

SPEAKER_01 9:19

Yeah, I think it's like it's a feature, it's not a product, it's not a company, right? And I think my point about that is spell check is really helpful, grammar check is very helpful. Doesn't mean you're gonna write a bestseller because of those two things. And I think that's like there's there's been this over-indexing towards somehow reading plans, AI reading plans and specs is this, you know, is the end all be all to everything. And it's like it's a feature, it's like spell check, it's a so what in many ways, right? I'm not saying, you know, and so I think that that's that's just my

Plan Reading As A Feature

SPEAKER_01 9:53

feed was full of all that, and then people bantering about it, and oh, you can't do this, and you couldn't use it for pre-con. And I'm like, who cares? Like, nobody cares. It's just like I don't understand it, but you know, I'm sure we'll you know, we don't get any uh listener mail. Maybe these people, maybe we've run them all off, but I do think to the point of experience and you know what are you gonna do with the drawings is really what starts to matter. Because I mean, 15 years ago, I could use computer vision to count light fixtures, to count doors and windows and dimension them. That's not that's not a new move. I think what's how you know what it might be is that it's a new move for people that are new to technology that it's just gotten so much easier to do. So they're like, oh, look, I can count all my windows and doors. I mean, I don't know. Yeah, but but I but I think it's in the world of like doing the hard things and creating real problems. I think you really have to have a deep understanding of like who's gonna use it, why. And I think I wrote something the other day I can't remember about if you look at IPD. I'm speaking at this design build conference in November. They asked me to come talk about some things, and so I was writing up some stuff for them. That's it's wild. They're asking me for my presentation now, it's in November, and I was like, I'm not gonna do it because do you know what's gonna change between now and November? Like literally everything, right? I was like, so I'll give you some points, but you know, one of the things with BIM, IPD, design build, design assist is another method.

First Principles In Project Delivery

SPEAKER_01 11:30

If you take all the first principles of all those things, like what are they trying to do? It's all these different methods, but what's the first principle? So the first principle is get all the expertise in the room as early as possible and as often as possible. That's the first principle of it. Now, the dynamic of that, like design assist, I went down this rabbit hole, which ended up going into case law and litigation. I was like, wow, okay, this is interesting. I've never dug into some design methodology and found so many lawsuits. Being a design assist is like, hey, we'll get all the contractors in the room up front and they'll help us with design. Well, it turns out it's created a lot of liabilities because you know the the architect's kind of taking advice from the contractor and then signing off on the drawings. And if something goes wrong, the architect's still liable, but our contractor told me and I paid them, so they should be liable. So create, but if you take if you take all the lawyers out of the room, you know, and you say, like, what's the best way to handle any problem just in life, right? You just get everybody in the room and say, Hey, let's go solve this problem.

SPEAKER_03 12:46

Yeah.

SPEAKER_01 12:46

And I think there's there's just some basic stuff like that. I mean, clearly I'm solving for some of these things, but I just I don't know. I just feel like sometimes there's the point is being missed. And I know these days first principles gets used over and over again, but there's a reason for it, right?

SPEAKER_03 13:03

Yeah, I've I've heard people talk about this dynamic just even more broadly in AI, which is like you know, there's tool-shaped objects out there that people get excited to use. And I think like a common example is uh just how people how most people use LLMs throughout the day. People are excited to use them, they're using them all the time. And yet for a lot of the basic tasks, like no one's actually getting more productive. So it's like you you do just despite how how advanced the tools are getting, you still have to keep the center of the problem in mind, which is like, hey, we're just trying to build more efficiently, cost effectively. We're trying to solve problems between stakeholders, right? It's a you know, there's there's many ways to solve that problem, but if we just invent more tools that aren't trying to solve that problem, they're just you know, they're just grow, you know, growing their own decision trees. Yeah, it's not actually helping the, you know, it's not actually helping anything, getting anything done.

SPEAKER_01 14:05

No, and and I think it's I think we're it really started, I think we're I don't know, this is how I use some of these tools, right? Is there's like these ideas, right? So you have multivariable problems, right? Just in life, you have multivariable problems, which just means there's a lot of variables, more than you know of and more than you've thought of, right? And then as a subset of multivariable problems, you have multi-physics problems, which means like how do things work in the world, right? And to me, that's the fun part of all this AI stuff,

AI For Understanding Not Hype

SPEAKER_01 14:39

right? In other words, I don't have to, I can run a 20 variable problem and get some insights to it that I couldn't before. Right. You know, I was talking to I was talking to Barry about this, we were nerding out, and I was like, we should get a we should get like a live data stream for MacCueWeather, right? Live variable. Why not? Like, why not? Right, and because in the past, people would say, well, what's the point? Right, what's the point in having this variable? I'll go figure it out when I show up to the job site. So what we try to do is because of just lack of human compute, right? We have to like fix some variables and say, okay, well, what do I really need to decide? And so I think that's like really cool. And then you start getting into multi-physics problems where you're worried about, you know, mechanical, electrical, all the things and how they interface in the physical world. Like that's super cool too. So I think like the idea, and and I was using this example last night. I said, if you go to 10 civil engineers with the same years of experience, the same world life experience, right? And gave them the same project, they would all come up with 10 different solutions. None of them, I mean, they're all wrong, by the way. Right in a way, like in a nuanced way, they're all wrong, right? Oh, this is true. This this one's too expensive. Oh, you know, there's there's always gonna be a problem, right? So it's never there's no perfect, right? There's no such thing as you draw drawings and it gets built exactly because if it did, we wouldn't need as built, right? And so so then you say that you have 10 people designing the same thing. Well, you would never do that in the real world, yeah. You'd have one person do it, yeah. But with AI, could you have 10 10 engineers do it and then say, okay, well, great, let's resolve the issues, right? Because so so much of the time getting the right answer is iteration. I think that's just how it is. And I think what AI is enabling, if you if you want to tackle hard problems, it's not about answering one question, it's about running kind of simulations against a series of questions and multivariable events to come up with some options to understand what's going to happen. And Barry uses this word like 10 times a day, just with me, is you have you should be using these tools to gain an understanding. That's what the goal is. It's all about gaining an understanding. And if you don't, if you don't get the get to that point of understanding, you're like using the tools wrong. And you have to design the tools to gain an understanding of the basics. Yep, yep.

SPEAKER_03 17:20

Is he is he is he suggesting that to you personally, or is he thinking about that? I guess he's suggesting that to your customers.

SPEAKER_01 17:28

I think we're suggesting it just culturally, right? People say things without an understanding of things, you know, and I think he just, you know, he yeah, he's Dr. Barry Clark, he's an academic, you know, and all that, right? So I think what he his point is you can't truly solve a problem until you understand the problem. And what AI is doing is getting people to start trying to solve a problem without gaining an understanding of a problem, yeah, or using AI to gain a better understanding of the problem. There's a little bit of like slow down and then move fast kind of mindset, I think, that I think it's just too maybe it's just too easy to be on quad on your phone and just train, you know.

SPEAKER_03 18:17

Very interesting. And yeah, I do think, I mean, you alluded to this, but I do think there is a scenario where you can still gain that understanding by playing, playing stupid and just interacting with the model and iterating on its answer, you know, iterating on the answers and growing your understanding that way. Like you don't have to have you don't have to like operate in a really slow fashion and get stuck in thought. Like all of this can be interactive, but the point still stands, like you're not just having a chat to have a chat, you're having a chat to grow your understanding.

SPEAKER_01 18:52

Yeah. But it's interesting. We we didn't have this on the agenda, but I did my you know annual Father's Day post. I don't know if you checked that. I don't I think I missed that. You know, everybody else opens it first thing in the morning on Father's Day, man. I always enjoy your Father's Day posts. So it came up because I have so many dads, and I always tell people like, moms were great too. I'm just a dad, and I tend to hang out with more dads that ask me about dad things, right? So there's a re I don't write Mother's Day posts. I'm sorry, like I'm not a mom. So before anybody gets angry.

Fatherhood Anxiety And Lego Lessons

SPEAKER_01 19:30

So my I get a lot of questions about from dads about like what should I be doing with my kids? Should they go do computer science still? Should they do engineering? What should they be focused on? And and and it of course, right? You should you should be worried about all those things as a father, right? And I kind of made the point like look, is one, I think they're as dads, we have some anxiety about the future for our kids. And I think we we while we focus, you know, yes, I might be sounding chauvinistic or whatever, but I think dads focus on a lot of the more practical things, and one of those aren't necessarily happiness. It's a little more like, hey, you need to get a job and make a living, you need to be able to provide, you need to be able to be independent. Like we we tend to peddle some of these more less emotional things a lot of times, right? And uh, and of course, you know, broad strokes. So I think that's why I get a lot of dad questions like, what's the future? How are my kids gonna make a living? What should they do? And like, what skills should they build and all that stuff. And I just kind of like reflected on all those conversations ahead of my father's day post. I was like, what are we, what are people really asking, right? And I think it's less about what college degree you get. Like that's a choice. That's a that's a choice of how you want to learn to learn, right? Because that's all educational system is learning how to learn. You may or may not learn something along the way, but you mostly just learn how to learn, right? And you know, maybe when you're 18, you're like, forget about high school, I'm gonna go be a mechanic. Great. You want to go learn how to learn by being hands-on, perfect, right? And so I think that choice is definitely a personal choice for a lot of people. But the the thing I reflected on uh and I got fixated on was Legos. So when I was a kid, way back when, when the world was still in black and white, Legos came in a few primary colors in a mixed box of just Legos. It was just Legos, there were not various versions, it was like just Legos, and what that meant is there wasn't anything on the box. You just loaded, unloaded the box, and you started building things. So you used your imagination, creativity to build something that ooh, look, I built a car. It looked nothing like a car, but in my mind it was a car, right? You know, and Legos have been around forever, they've done okay, right. But when was the big unlock for Legos? When they started building sets that you actually built something with intent and a result. Here's the instructions, you too can build the Millennium Falcon. That's when the stuff took off. And so what it did is it set people up to say it's about completing the Lego set to get the result, but it's not really about creativity and ingenuity.

unknown 22:27

Right?

SPEAKER_01 22:28

It's following the instruction man.

SPEAKER_03 22:29

When you move to the instruction manual, you're just following a script, following directions. There's zero creativity and imagination in that.

SPEAKER_01 22:37

Right. And your your boys will get older, but like when my boys were younger, I'd gotta spend $500 on a Lego set, which is like insane because George Lucas has to get his money, but that'd be done in an hour. And I was like, that time of entertainment was $500 an hour. This is insane. Like, you guys need to move slower. It should be harder. It should be hard. Like, or I don't know, like, but the point is, like, because they weren't focused on the journey and the creativity and all that, it was about completion. Yep. Yeah, complete to finish. So I kind of honed it on my post about this idea that like maybe that's what we've done with our kids. And then what happens is at the end of trying to build a Millennium Falcon, if there's a few pieces missing and it can't be completed, they lose their mind. So you went to college, you went to high school, you graduated from college with a computer, you did all the things that the Lego box told you to do with an end result in mind. And now you're losing your mind because some of the pieces are missing. I I can't find a job. And so what you know, my so my talk has been like it's about creativity, right? Like, you know, my four-year-old, he's nothing but out there playing in the dirt in the grass, right? And the first thing he did is he built his because he got his first Legos for his fourth birthday. He built it. The first thing he did was destroy it. And of course, I'm like, we're gonna lose the pieces, we gotta find all the pieces. He could not care less. Like he was just like, I don't care, right? And then he'd started taking the pieces and putting them back together in whatever form, started mixing them with his wooden blocks, right? And clay and played it, like he didn't care, right? He wasn't like he was unbothered by like what the end result was. And then I realized, like, oh, I'm trying to get him to conform to to the results. And so my my point of my point of my post was just like the only thing we should be teaching our kids is like imagination and creativity, and that will make them resilient. And I and I think Legos are a good example of that.

SPEAKER_03 24:46

Yeah, that analogy is great.

SPEAKER_01 24:48

So I got a lot, I got a lot of fun posts, but I got a lot of fun DMs from it.

SPEAKER_03 24:53

People are like, Yeah, I mean, there's a lot of threads there that you've pulled just in a lot of your commentary over the last couple of years, your books about that topic.

SPEAKER_01 25:04

Yeah.

SPEAKER_03 25:04

Like, except it's for adults. How are you optimizing for creativity and imagination and hum human human-oriented skill sets? But yeah, I think there's I think like I think there's a lot of truth to even to pull this back to where we were, growing your your understanding of what you want to do, what your value is, what your strengths are, what your weaknesses are, knowing yourself, right? And if like we're just following the rule, the rule book and the instructions, and we, you know, we either get a job and we hate it, or we don't get a job and we don't know what to do, like we've we've avoided maybe the most difficult task, which is like knowing yourself, thinking creativity, using your imagination, failing, you know, running into walls, running into pain. Like you know, I I can speak from experience here. Like when you follow that rule book, I think, you know, assuming you had all those things and you got to college, I think that's like a fair assumption here. Is like you had like most of the most of the people that you're interacting with, like had pretty privileged upbringings. You didn't have like too much hardship, right? So what do you do when you run into that first difficulty? Yeah. Like I know I didn't learn how to how to handle that, you know, that sort of challenge early on. Like and and it's up to you, it's up to the individual, right? At the end of the day, your parents can't, you know, you know, they can't go to the job interview with you. I know they I know I know they've tried, yeah. I've heard some nightmares where parents should go to the job. Parents should not be going to job interviews, so let's uh let's not recommend that. But but yeah, like at the end of the day, like you gotta swim. And I think it's like if you can train that into someone and instill that sort of confidence and also just that sort of this sort of creativity and ingenuity needed as they're growing up. I mean, I've got obviously really young kids, and like that's something that's central in my mind is like, can they think for themselves? Do they have any do they have independence? Can they, you know, use their imagination? It's hard to do nowadays, really, truly.

SPEAKER_01 27:11

But when everything's funneled to you, right? When everything's just kind of funneled to you. Totally. Um, I think it's hard. It's it's interesting. I think after I left civil engineering to do more telecom stuff in the internet era, my mom still told everybody I was a civil engineer. Because I think she was embarrassed by this idea that I went to college for one thing and I was building web applications and deploying routers, which was not what I went to college for, right? And I think it was like embarrassment for her. Like, you know, how do you go do something that you're not that you didn't go to college for? How dare you? Kind of thing. So speaking of imagination and creativity, had the great pleasure of attending reindustrialized in Detroit a couple weeks ago.

SPEAKER_03 28:02

Yeah. Give us the debrief.

SPEAKER_01 28:06

Yeah, I mean, kuda, you know, thanks to Paul Kwan, General Catalyst, for like dragging me along and getting me in because it was like a sold-out game, so to speak, and making tons and tons of introductions. It was it was super interesting. I mean, I was like absolutely energized after attending. And you know, I'm

Detroit Reindustrialization Takeaways

SPEAKER_01 28:27

glad he kind of was like, Hey, you need to go to this with me. And I did, but it was it was super interesting in terms of it's not really a place that we spend a lot of time, right? These were startups doing things in defense, doing things in advanced manufacturing, you know, physical builders, right? Physical builders, and it was just really, really fascinating to see like who one, all the speakers, you know, and it was interesting, you know, several of our porcos were there, right? Western Chemical was there. Monumental Labs was on stage with Mr. Ohanian, you know, talking about robots creating you know traditional sculptures and stuff like that, which was super cool. But I think you know, these folks, the way they're approaching things, the way they're thinking about things is a really good example of scarcity versus abundance. And when I look at like software development and SaaS, barrier to entry is low, right? There's a it's it's it's easier from a perspective of there's an abundance of resources for you to be successful, starting back with AWS to AI vibe coding now, right? It's there's a but there's a lot of abundance, and so it's a little bit easier, right? It's the lot less barriers, and these physical builders hearing like their stories. I mean, it's the opposite, right? It's all scarcity, scarcity of people, scarcity of this, scarcity of that. And Blake Shoel with Boom had a really good talk about how they're 3D printing their own turbine blades for their engines. And I was like, what are you doing? Like, why would you do that? Right. But it was super interesting because there's like a few takeaways I had. One, there are very few venture capitalists that understand this business. And one of the things that was brought up is if you're gonna go build something, you should probably own your own building, which means you're gonna deploy capital from your VC to go buy a building. And a lot of VCs would tell you, what are you doing? That's not capital efficient. You need to be capital efficient. Yeah, their point is, but we don't we can't rent a place. Totally, that's how it works, right? So a lot of VCs that have not like built these physical businesses will automatically tell you no for a lot of things, right? And I think we've experienced some, I think we've learned. I mean, I think we're good listeners, generally speaking, at least you are. But with some of our core codes, when they make decisions that we're like, are you sure? Like, that's not normal. And they're like, Well, this is what I think I need to happen. We've been like, Okay, well, you know, ultimately it's their company, right? We're along for the ride. Um, and so I think it was interesting. So decisions like that iterating the iteration process, and it seems very counterintuitive. Blake Schultz talked about it, several people talked about the speed of iteration, and that it doesn't make sense. Most people would say it doesn't make sense to build your own stuff, contract it out, right? And their point is if you're gonna contract out getting something manufactured, by the time you design it, ship it and get it back because the supply chain, just time, right? You're not an ideal buyer. You're buying, you're having them make five widgets. They're at best customers are making 50,000 widgets. So you're at the back of the line, you're not a priority. So it might take you three to six months

Why Hard Tech Must Control Supply

SPEAKER_01 32:05

to get your prototypes back. And while that might be good for scale, but in the near term, the general sentiment was you should do all your own stuff. Make make all of your things, have your own machine shop, iterate your own things. Because if you optimize for speed and the speed of iteration, then you'll get somewhere. If you're always waiting on vendors to deliver something back to you, by the time they deliver something back to you, you've already had an iteration and an improvement, right, to deploy.

SPEAKER_03 32:35

Yeah.

SPEAKER_01 32:36

And so I thought I thought, but you know, it wasn't one person saying it, it was everyone saying it, right? It was like, hey, and it felt like an education for VCs that aren't playing in the space.

SPEAKER_03 32:47

Yeah. I think it's interesting because like in that it runs so that advice runs so counter to what you hear on the software side. And even for you know, for a services company, it's like, don't move beyond your specialty. Help keep things as capital light as possible. And then yeah, we have direct experience with portfolio companies saying, like, no, I need to go acquire a business or I need to go own the real estate, I need to go own my factory, which is you know, millions of dollars of CapEx, but it's the only way for them to guarantee the production capacity they need without someone pulling the lease out from underneath them, right? Which like these are critical path things for the viability of the business. And in in hard tech, when you're when you're building in the real world, every every dependency like is fatal, right? Like it, like there, there are, you know, they're just like there's there's so many different, as opposed to software different ways to die and not get your product out into the world. So like you want to control as much as you possibly can, especially things that are in the critical path. So like one of our companies' critical paths is batteries. They have like unfortunately, they have to do their battery packs in-house. It's painful, it sucks, it's expensive. If they didn't do that, they could have an existential crisis on their hands at any given time because they can't ship their vehicles. Right. They have to do it. Yeah. And it's just like there's yeah, like in so in the in in that world, it's such a different heuristic to operate a business out of different capital needs, different financing mechanisms, different, yeah, different understanding from VCs. I mean, yeah, like to your point, like we're we're receptive to all that. We understand it, but it's so different than a lot of our other companies that we funded previously.

SPEAKER_01 34:34

No, I mean, I, you know, and I think like, you know, just to put names on company, you know, Ahmed at Lumina, I think our interaction with him has been, you know, we're we're just as much learning from him as he's trying to learn along the way. But I think, you know, we've been open, right? We've been open to say, like, oh, okay, like that makes sense. So, and it was really interesting too, because I think if you look at making things in America, it was always like, let me make this widget and I'll get China to mass produce it. And one of the best talks was Antonio Gracias with Dollar Ventures, right? Those of you that don't know, early investor in Tesla, SpaceX. He's they kind of, I think the the thing is they bailed out Elon Musk when he had no money at Tesla and kept it going. And he gave a great talk. And one of the things he said is like we have to stop thinking about manufacturing as like mass manufacturing, but rather more nimble manufacturing and speed speed and better,

Nimble Manufacturing Over Mass Scale

SPEAKER_01 35:43

not quantity and cost. He's like, if you think like if we're gonna sort of replicate China and say, how do we build the same widget over and over and over again for cheap? That's not the point. He's like, that's that's generally not the point. We should be thinking, you know, when he held up his cell phone and said, like, look at this iPhone. Why do I get the same form factor for two years? Why can't I get a new one every 30 days? Why can't I have 15 form factors? And so in his mind, he's like, the unlock is by the reindustrialization of America, yes, we can have all the economic conversations and political conversations. But his point was what we have to do is be more creative and do different things instead of narrowing our skews down to three SKUs, which is what we had to do when we were shipping stuff to China and getting it mass-produced. I'm gonna do three SKUs, mass produce it, keep the costs down. He was like a little bit more like more skews, more frequency of releases, and do better, right? Do more interesting things. And I thought I thought that was a fantastic point of view. I mean, literally everybody, the the crowd was crowd went wild when he said that. So but uh it was really cool. It turns out one of his kids is a torch attack.

SPEAKER_00 36:58

Oh wow, yeah, pretty cool. Yeah, all good.

SPEAKER_03 37:03

All right, until next week. That was a fun meandering one.