Why GCs Don’t Do Anything (And How AI Might Change That)

It's a little bit of tongue in cheek , but this is my opportunity to ask KP what was the inspiration behind this thing that you posted on LinkedIn ? If you're not following KP on LinkedIn , you should be . Just go to KP the letter K , the letter P and then ready R-E-D-D-Y .

He posts a couple times , three times a day , about his insights , the things he's heard , the things he's learned , conversations he's had around designing , building , developing , owning , operating in the built environment , and all of his posts are thought-provoking and insightful , having to do with the AEC world and beyond , sometimes investing , sometimes startups , like I

said earlier , sometimes designing or building , as we're going to talk about today , and often owning

and operating . So I am joined , as always , by KP Reddy . Hi , kp , hey , what's going on , jeff ? A lot of podcasts recording today . Yeah , so if I get quiet at some point , my voice has left the building . We'll see how it goes .

So , for a little little inside baseball for everybody , kp and I , before we get started recording and all these sessions like , okay , we look at what he's posted , we look at the engagement , you know what's what's hitting , what's not . You know what are people commenting on , what are they not ?

And then , of course , we select a couple of episodes to record , and one of the episodes that KP pointed out before we hit record today was hey , let's talk about this one and I'm going to read it to you here in a minute . But , as we're recording this on February 21st 2025 . So you know , february 21st 2025 .

So you know , if you're listening to this year from now , dial it back a little bit . You posted this about 24 hours ago and it's blowing up . I mean , I think that's . I think that's the technical term for what's happening is blowing up .

We started unpacking it a little bit before I hit record , um , but this is . I think this is going to be pretty interesting . So here's how it goes . I hear lots of answers to this question and I have my own ideas . Why do so few GCs self-perform anything ? If your answer is risk , please be specific . That's it . That's the whole post .

So where did this come from ? What ? What conversation ? Um , what were you hearing ? What were you doing that brought this to mind ? And , and that specific question , why does so few GCs self-perform anything ?

And they are there to help us manage risk and manage the project . However , there's like a change order , right ? We're paying for all the change orders .

Like we're going to squash them down for you , owner . Or they're passing the change on to the customer , right , the owner's paying for it . So I'm like , are they managing risk ? I guess that's managing risk , but it also feels like it's passing the buck around , right ? It's like just don't charge me for anything , right ? So that was one area of thinking .

And then I've got in my mind like every day I think about this . Right , we all think about certain things . I'm a nerd , I think about things . Right , we all think about certain things .

I'm a nerd , I think about things you know , probably less interesting , but it hit me the other day that if AI is going to design buildings , are they designing buildings with the means and methods of humans in mind , or are they designing with the means and methods of what the robots are capable of ?

In other words , you know , like one of my poor cozokibo that doesn't know . Of course I'm plugging them because I'm an investor . So big surprise . But they do automated drywall finishing and painting .

And if I ask them like , hey , what percentage of projects are designed in a way that the robot can't paint and finish the entire space , they're like well , you know , there's always a percentage that like well , if you get feedback to the designers and said , hey , like , move this wall out a little bit , move this in a little bit so that it's optimized for the

robot to do a hundred percent of the project , wouldn't that be better ? And they're like you know , actually we have a tool where we can upload the drawings and we can decide what areas the robot can do , which areas they can . I'm like but are you giving it back to the designers to say redesign it ?

And they're like no , we're not doing , like we're not there yet . I'm like okay , that's fair .

So then I started thinking , taking from that , so if we have AI designing buildings so that robots can build them , if I'm a GC , if the risk is in managing resources , scalability , all these things that we got out of that post , which I suggest you talk to some of these people that posted , they said look , look , you know it's a different skill set , it's a

this and that . But if they can buy a robot , which GCs do have big balance sheets to manage the risk instead of subbing it out , can they just buy the drywall robot ? Can they just buy the robot that can um structurally frame a building ? Can they buy a robot that lays the flooring ?

And if they can just buy the robot , then the scale of um of capacity goes away . The risk of quality actually maybe goes away . The risk of schedule maybe goes away .

So what kind of ding ding ding in my head was if AI designs buildings so that robots can build them , if the GC is going to start self-performing everything , that's because they're going to buy more

robots .

But how many people realize , like , if you are an architect who's really thinking about the constructability of of your building and I can , I can remember specific examples of this like , hey , we can't do that because the framer can't actually get in there and swing a hammer , or we can't actually fit a nailer in that space to get that nailed .

You know what , you know fastened , something like that , and I mean , that's , that's real world stuff , right ? So what you're talking about is a different version of that .

Um , now , of course , before we hit record , one of my , one of my questions , you know sort of playing devil's advocate was okay , well , why don't we design the robots that are optimized to a certain to a certain size or something like that ? But you know , we've got building material , modules and things like that .

But I think , you know , I think what you're saying is really intriguing and one of the most common responses , I think , at least as I look at my screen here , and I don't know how the comments are sorted here on my screen , but one of the most common comments was well , it's a different business model .

And I read that and I was like , hey , I think that might've been the point of the question .

There's like six hot dogs but eight buns yes , same thing it's like that , right , like we're going to design a building , so there's leftover drywall , so there's leftover carpet . And it's interesting because , as I was like , of course people text me and people DM me about this stuff , right .

And then they're like , well , what you're talking about is very complicated , and literally the next day or that day , microsoft talks about quantum computing and with that , you know , like their their breakthroughs , the 17 year breakthrough that they've had on quantum computing .

And the first thing is , I do a deep dive on that and I'm like , oh my gosh , like so if you're saying like we have to rationalize what the owner wants in terms of a building and the ability to optimize the construction to get to a zero waste environment , oh , that's too complicated . And I'm like , no , no , we just have , we have quantum computing coming .

It's not that complicated , right ? Quantum computing can run these , run these scenarios for us . And so , anyway , I think it's , it's interesting , like this post , especially like we got more comments than we got reactions , which is always fun .

But I think we have to start thinking about this stuff and , and you know , do architects ever design buildings and not think about the means and methods ?

At some point when you're in school , you are taught about the modules of construction , like studs laid out at 16 inches on center wood framing perhaps , or maybe 24 depending , and then , of course , the , the four-way sheet of drywall , etc . All your nominal dimensions , etc . Etc .

And you're you're taught to think in that way and and I think we talked about this on a past episode when I , when I did my internship about 120 years ago , the two there were , there were three partners in this firm . Two of them were the architecture partners and one was interior design partner .

The two architecture partners to me , of course , I was what like 22 years old or something like that . To me , these dudes were old . Yeah , these are the old guys , but these guys had Mies van der Rohe as a professor when they were in school .

They were old , they're really old , but what they brought to it , what they brought to their practice , and and they're so they went to iit in chicago and there's this infamous project that you do at iit . They still do it . I saw , saw somebody , uh , the other day on instagram or something , um , a student had posted they still do this project .

I think they refer to it as the brick project , but they design some building . Whatever the program is of the building , they design it as brick and they have to figure out every tiny little module , including the mortar , joints and everything else , and they have to document the whole thing .

And it's part of that is the process , or the , the , the motion of understanding the modules . So the students should be learning this , do they probably ? Do they carry it into the profession ? Maybe are buildings designed with that in mind , maybe not . Yeah , but I think , to get to what you're talking about , we , we have to right we , we have to .

If , if you you're talking about , we have to right , we have to . If you're going to reduce waste , you have to figure out how to reduce waste , and part of that is the modules that you construct with .

It's called design for manufacturing and so we're building these prototype robots like custom prototype robots , right . And then we go to a contract manufacturer that's going to manufacture hundreds of these right to scale us , right , and we paid these guys . They came in , even if they're going to manufacture it .

They come in , you pay them a ton of money and their engineers say look , this railing that you built for this robot to run on , um , it's 10 feet , it's 10 feet long and that's a custom size . I can get you a nine foot one for a third of the price .

off on the cost of that motor . And they went through piece by piece by piece and took this custom Frankenstein machine that we built that worked and said here's all the SKUs that are available , the nuts I mean down to the nuts and bolts , right ?

Because we're seeing some weird sized bolt that it turns out , if we had chosen a different one , it's like half off , because it's a highly produced SKU , right ? So they go through and do all this stuff and , by the way , this is what Apple goes through when they deal with Foxconn to manufacture their phones .

They go through this DFM process and I found it absolutely fascinating because they took a robot that the custom one that we built cost us a million dollars . They got it down to 50 grand .

We keep them in inventory and we just drop ship them right to the right to the site . And it was really fascinating and I thought you know we try to do that a little bit , but it's almost like too late , right ?

It might mean , you know , it's optimizing for the availability of pieces and parts and for means and methods and everything doesn't have to be custom . Right , like it's custom , like oh , we're getting this custom window because the architect , the fancy architect , decided this is the window I want .

Well , there's one that's like three inches shorter , that's like a tenth of the price , and that green that you're so like enamored with , there's an off green of that . That's half price . Yeah , and if I'm the owner , I'm like I don't care about the green is green maybe .

When you're building that prototype , the point is not how much it costs , right , it's , it's how do we build it so that it does what we want it to do , it gets the job done , et cetera , et cetera . And then and then we go through the process then of of optimizing everything cost and operation and everything else .

So we start thinking about this in terms of scale . You know , okay , we built one . You know , we used that nut that was laying there on the or that bolt that was laying there on the workbench , because it was laying right there and it fit in the hole or whatever it was .

And now they figured out that we can get , we can use a different bolt just as effectively for half the price , or something like that . So if we translate that to construction , if , if you're thinking about , if you're an owner that builds lots , of , lots of buildings and I worked for a firm years and years and years ago and we did .

It was a strange firm , we did everything we did . I worked on the Museum of Contemporary Art in Chicago , and then we did distribution centers for , like , albertsons and Safeway Foods and things like that .

So , tilt up concrete versus custom everything right In a museum of contemporary art , but if you're the Alberbertsons or the , you know the whatever that build lots of buildings , this makes total sense . Oh yeah , we want you to optimize everything . How do our trucks get in and out ? What are the bays ? Like ?

Um , you know our sites , the , the land that we buy and we develop and all those things . And and we might think like that , like how do I optimize this ?

Which makes total sense , and what you're saying in terms of construction , we have to think about it in terms of a systems approach , and it's not just this one time that we're going to build , it's every time that we're going to build . And you know , maybe on the other end of the spectrum , a custom home or something like that , that's a very different animal .

But still , in the building materials world there's no translation from tilt-up concrete panel to wood stud . But in the building materials world , there's tons of overlap there where we especially if you're a contractor too it's like how am I going to build these things ? Why am I not thinking about optimizing ?

So it's 2000 degrees up in the attic and I'm trying to like change this out and can't see and everything else I'm like this is insane , right . And so think about that . Every six months I was getting tortured by this air conditioning system , to the point I just told these guys like I'll pay you a hundred dollars , you just come do it for me , right ?

And so think about how many times in a utility closet in a boiler room there's some handle that's just like awkwardly put there , and so there's all these intricacies about not just how do you build it , but how do you maintain it and the lack of clearance we give .

You know , if you really want it optimized , you know they had , like in a warehouse they had those machines that automatically drive back and forth to clean the concrete , right , that wash the concrete a little suction thing . You see them in airports and stuff . Should it be optimized for two widths ?

So you're driving up once and back and you're done , because if it was two and a half , then all of a sudden I gotta go like that's three , right . So you start thinking about even operating this thing and the auto . I mean , if we think , if we have these roombas in our houses , are we not going to have industrialized cleaning bots ?

right right right , I mean , I saw this great demo this is 10 years ago a caterpillar , a fully automated grading equipment , greater it dropped . Dropped the grading plan and it runs . But , guess what ? It can't grade everything , because if the civil engineer didn't design it a certain way , they were like nooks and crannies . It couldn't get into Right .

How important were those nooks and crannies ?

But now I became a little conflicted there for a minute , because you're talking about hey , we're going to design this or we're going to optimize it , so the , the cleaning machine , makes two passes , once up , once down , and obviously you know that there's , there's an oversimplification in the example , right , there's okay , well , that's , we're going to design this

for the cleaning , but what about the forklift that's stocking the , the upper things , and what about the , the exiting requirements for , for fire code and all those things ? But but your other point about the quantum computing which isn't necessary for this example , but all this stuff is , is figureoutable . It's not that hard to figure that out .

Once we understand all the things that need to go in there the cleaning , the stocking , the exiting , whatever the things are Then we can come to some number , and one of them is going to be the control number .

I think I posted another post that said hey , we're moving from means to methods , to orchestration and agents , right , this is like a massive symphony of sorts that we're trying to put together , right ? And I think if you look at quantum and I don't even know if you need quantum I'm just really hyped about it because Satya is hyped about it .

I get hyped about anything Satya Nadella is hyped about . But you start thinking about all this stuff , right , and then you go back to the building product manufacturers and so that's like the second wave . I'm building right , so I've been talking to all the owners .

I've been deliberately not been talking to AEC because I just don't think they're mission aligned to solve the bigger problem . They're too busy building time and delivering change orders . They're just too much in it for me to pull them out of that mindset .

So now I'm talking to building product manufacturers and I ask the simple question how often do you get feedback on how to improve your product ? And they're like hardly ever to none . Nobody ever comes to us and says hey , you know that light picture you designed . If you did it this way , it would be easier for me to install .

If you did it that way , it would provide enough lighting that I could reduce the number of canned lights when I'm designing it , right ? So you start going down that that that way , about every you know , of course , we all know like the number of skews in a building are wild , right , it's doors and hinges and doorknobs and hvac .

I mean there's so many skews , right , and

I think it's fascinating . If you take this dfm approach um , which , ethan , you know , producer ethan , you might want to put some notes in there about dfm , since you know , maybe our , our listeners don't know much about it . You can put something in the show notes about design for manufacturing .

But if you take a DFM approach , there's a constant feedback loop to the conveyor system , people to the automated welding to the CN . There's a constant feedback loop and I feel like we have no feedback loop to the manufacturers . Feedback loop and I feel like we have no feedback loop to the manufacturers .

In fact , I was in a situation with an owner , a large grocery store chain , and every one of the stores was having a problem with a specific piece of equipment and it was funny , they didn't even complain to corporate .

They were selling the equipment on eBay and buying a different manufacturer that they liked better and corporate didn't even know this was going on . But it was like this underlying thing . It's like well , guys , why didn't you let corporate know that this piece of equipment is not working for you ? I don't know , I'm just , I'm trying to run a store man .

I'm trying to get stuff done and one of the other stores . I was talking to him and he had the same problem . He was like well , how'd you solve it ? He's like this . I was like so there was no like feedback loop . To corporate I was the feedback loop , which is wild .

And also to the manufacturer to say , hey , we're this big grocery store chain , we've standardized on this skew . We're getting some feedback from our store managers for x , y and z . Have you guys heard like are you doing anything about it ?

zero feedback so I think , when we start thinking about that , the feedback that can be provided by the design community and the construction community and the operations people , the building product manufacturers . I feel like we can solve so many like optimization problems .

Think about , like you know , you're an HVAC company or Lennox or Oda Carrier , whoever hey , we design in these three tonnages or these four tonnages . Well then it turns out like I'm designing and it's like , hey , it turns out I'm always having to over over buy the tonnage for my cooling systems because , that's all that's available .

Everything else is too undersized .

Why does the design of this particular elevation of this building or something like that , why is it being driven by the size of this one material , doesn't make any sense at all . Right , it doesn't have anything to do with with any of these other things , except that that's the size that that material comes in . That's just , you know , really superfluous example .

But I think your point of of the feedback loop being constant , being continual and through all the players , I think that's what's going to be required for this business model , if you will , this idea that you're talking about . Otherwise , it starts to fall apart pretty quickly .

That's just the nature of physics . Right Things break . Oh , it turns out we bought this product and spares aren't available . You know what I mean ? It's just like this thinking around the long-term viability and waste and all that around buildings . I think we can learn a lot from that DFM process . Yeah , dfm process , yeah , and I think owners are open to it .

You know , I think the when I talk to owners , they kind of get bullied by their design teams . It's very , it's a very strange dynamic . I'm paying you and I'm asking for x and you're like well , x won't work .

Sorry , like uh , but I'm the customer right , like this , like the customer , the customer right , like this , like the customer , the customer is right doesn't seem to apply .

Certainly , behind behind the scenes , we're getting ready to , uh , to host a sustainability um what ? do we call it sustainability forum um , on the way to launching a new mastermind group which is really going to be focused on resilience . These are the types of things we're going to have to to think about . You know , how do we actually ?

What's the actual life cycle cost ? What's the what's the process of designing , building and operating ? Because it is , I think .

I'm really looking forward to , like you know , starting to talk about these topics . You know , I think , argue about these topics , you know , I have this thing that I really believe that , like someone says well , it's the way we've always done it . Yeah my five year old , my inner five year old comes out , but why ?

So I think we've , really , if I think , if we're all committed to doing better which is where it starts right , commitment and passion is where it starts we're like the smartest industry in the world .

It doesn't enter into that , into that realm , but yet we evoke that term all the time . You know that we use that product , or we use , we make it that size or we you know whatever it is , um yeah , imagine if you went to your doctor and you said , hey , I have a fever .

But I think it really strikes at the heart of the evolution that we're seeing in the built environment , the AEC world , and the way that we design and we build and the materials go into it , the way that we develop them and and own and operate in the built environment .

So let me , as we get ready here to wrap it up , let me read this post one more time and then again , if you're not following KP on LinkedIn , you should be . Just look for KP , the letters K and P and then ready , r , e , d , d , y , you can follow him .

You won't probably won't be able to connect with him because he's usually capped out at LinkedIn's maximum number there . But follow KP or more more posts about what's happening in the built environment , and you'll find gems like this it says . I hear lots of answers to this question and I have my own ideas . Why do so few GCs self-perform anything .

If your answer is risk , please be specific . So go find that post when you listen to this at some point in the future at least from when we're now , when we're recording go find this post and type your answer into the comments and let's see you know . Join . Join the conversation .

Join dozens of other people that are reacting and commenting on this particular post . This is really interesting . I'm sure we'll revisit some version of this conversation again before too long because , like I said , I think it's one that's going to be on our radar a lot in the near future .

, because this is the way we've always done it .

And also the founder of Shadow Ventures , as KP said , I guess by the time you hear this I didn't think about this as I said it , but producer Ethan is not going to have this one published before we head to Phoenix for our next one day mastermind event , but we will be in Atlanta in Q2 for our one-day mastermind event in May , so keep an eye out for info

on that . I know tickets are already going for that event and we have other offerings out there , obviously .

I think may is going to be one of those ones where we actually have to tell people like , hey , we don't have room right , which is which is usually just a marketing ploy by the events team , um , but this time it might actually because I'm just .

I think there were so many people that couldn't make it to phoenix and are automatically signing up for atlanta , right , and of course , I've spent so much of my life in atlanta , there's also like a good following there as well .

It's our one day mastermind event , our Q2 one day mastermind event , coming up in Atlanta , may 20th 2025 . So , looking at show notes or registration link for that event , kp , as always , thanks for joining me today and for all of you that are out there listening . Thank you for listening . Thank you for joining us again . Follow KP on LinkedIn .

Go over and find this post . I hear lots of answers to this question . Have my own ideas . Why do so few gc's self-perform anything ? Go find that post .

in and talk about my posts . So you might be one of the special ones that get picked , that's right we're uh , we're starting a troll k club .