The AI Agent Arms Race Begins

What happens when everyone's AI agents start talking to each other—and you're stuck without any?

In this episode of KP Unpacked, KP Reddy and Nick process the Zero RFI launch aftermath - from 3,500 resumes in 24 hours to a top-tier VC introducing themselves like KP's never heard of them. But the real conversation pivots to what happens when everyone deploys AI agents: cognitive overload, the spy-vs-spy escalation of automation, and why construction's suicide crisis gets worse when information flows faster but judgment disappears.

KP breaks down why engineering firms are drowning in RFIs that should just say "read the damn drawings" (but legal won't allow it), why text messages with no context create work handoffs disguised as communication, and why the people automating everything on X probably don't have real jobs. Nick counters with diligence innovation—using Claude Code for VC code review, building Slack analysis tools to measure founder leadership styles, and whether term sheets should include MCP server access to accounting systems. The through-line? Defense agents, offense agents, and the realization that humans should only handle judgment and exceptions—but the magnitude of those decisions just went exponential.

Key questions answered:

• What was KP's favorite response to the Zero RFI launch announcement?
• Why did a top-tier VC introduce themselves like KP's never heard of them?
• How many resumes did Zero RFI receive in the first 24 hours?
• Should VCs use Claude Code for startup code review during diligence?
• Can you measure founder leadership style by analyzing their Slack history?
• Should term sheets include information rights to connect MCP servers to bank accounts?
• Why are engineering firms drowning in cognitive overload from RFIs?
• What happens when everyone's AI agents start responding to everyone else's agents?
• How do you separate real AI demos on X from complete fabrications?
• Why is construction robotics funding only $1.78B total—and is that enough?
• What's the right business model for robotics: sell machines, lease them, or become a subcontractor?
• Should robotics companies target OEM distribution partners like Milwaukee Tools?

If you're drowning in notifications wondering when AI actually helps, a VC trying to figure out what diligence looks like in 2025, or a founder posting fake demos on X hoping no one notices, this episode will force you to ask whether your agents are creating leverage—or just more work for someone else's agents to handle.

Listen now.

BuildingWorks & Brookwood Sponsors

Transcript

SPEAKER_01 0:34

Hey Nick, how's it going?

SPEAKER_00 0:37

I've got a question for you. More questions. This is a podcast, you know. Question is what was the single best response that you got from the zero announcement? Um could be funny, could be uh, you know, could could be opportunistic, could be you know insult insulting.

SPEAKER_01 1:03

I think the thing that made me laugh out loud, and I shared it with like one of the most reputable largest venture firms reached out to me and introduced themselves to me, like, hey, I don't know if you know who we are, do to do to do. And I'm like, it just made me laugh out loud. I'm like, if I don't know who you are, you shouldn't do business with like it's like that kind of thing. And that just made me just, I mean, I was sharing with you. I was like, it just made me crack. And anyway, so it's just but like your point, right? Also, like the polite thing to do, like you can be top three in the world and still like kind of you know, a humble slash whatever thing to do. But I was just like, if I don't know who you are, that should that should be you should self-select, I should just self-select out.

SPEAKER_00 1:54

It's like the founder of SpaceX politely introducing himself when he reaches out cold via email or the president of the United States of some someone of that of that caliber. Right. I don't know if you know about SpaceX, but let me tell you. I mean, I don't know. My reaction is it could be either amazingly polite and humble. And yeah, you never want to assume that the person on the other side has studied the history of your industry, or it's a humble brag. Yeah, and they're just like basically just you know, putting their chests out, you know, saying, Hey, you may not know who we are. We're just one of the world's largest asset management venture venture firms, and you know, yeah.

Betting On A Risky Name

SPEAKER_01 2:41

We've only invested in a deal. Yeah, we've only invested in every company you've ever heard of, but I'm sure you've never heard of us, right? Yeah, but yeah, no, and I'll take the broader thing that I thought was just not a specific outreach. I really, you know, being a bit in the industry forever and being a little bit myopic and some might say sometimes uh stubborn in my thinking, right? I did not know how the name would go over. I really didn't, right? I really I wasn't I was unsure, right? I've had so much positive feedback on the name because you and I talked about the name early on. It's kind of I mean, we uh me and you have a lot of history with the name, right? Yeah, and we're but we were unsure. I mean, in fact, like even what months before like launch, we're like, hey, should we rename? Like, we're having those conversations, right? Like, is it too industry? Is it too ambitious? Like just all the things, right? Yeah, but the feedback I've gotten has just been really, I mean, I was talking to Sun Mi, our CMO, and I was just like, I really didn't think it would go over this well. Like, I was like, I thought that actually. What's that? I'm surprised to hear that actually. Did I felt that way or did it went over that well?

SPEAKER_00 4:02

That that it was that well received. I thought I thought, I mean, you know, I just look around at the traditional owners' rep firms and how and the you know, the nomenclature of you know how you would go about naming that sort of business and what's accepted in that world. And to me, so zero zero RFI, which is obviously the name of your company, it just has like a you know, it has a technology bent, extremely ambitious. Some might say, you know, impossibly ambitious, like there never will be a world of zero RFIs, right? But that's but in in a beautiful way, that's the intent of the name is like that's the entire mission of the company is to create a world where there, you know, we we don't we don't have to make adjustments in the field, we don't have to process change, everything is perfectly predicted, and we we we get there through technology, right?

SPEAKER_01 4:52

Right. I mean, I would say like even the day after launch, Devin was like, you know, should I just introduce myself as being with zero? Like that somehow like people won't get it, right? Like it'll get lost, right? And so it's it's been I've been really surprised by that. And to your point, right, I had a a large hospital owner reach out, and they were just like, Love it. Is it reality that you'll hit it? Probably not, however, the name has ambition, it describes kind of what the mission is, and it's ambitious. Yeah, I didn't know. I mean, you don't know.

SPEAKER_00 5:31

It also it also signals differentiation in a way, in that the mission is held to the highest standard, so much so that it's in the name of the company, right? Yeah, and no one ever will compete with you for that name because they're not striving for that sort of goal, right? Yeah, like you're in a category of one that's trying to that's trying to truly achieve that in that space.

SPEAKER_01 5:55

Yeah. It it's been it's been surprising, and I don't know, like just in so many ways, like you do things, you try, you know, to me, it's always like shots on gold, you just keep trying, and eventually it pops in. This is one of the few times I feel like it just worked, right? Feedback has been fantastic, it's felt good to like not. I mean, up until launch, I was like, you know, still not can still change the name.

SPEAKER_00 6:22

So did you have any other other good candidates for names?

SPEAKER_01 6:27

Not really. That might be why we stuck with it. Is like I couldn't.

SPEAKER_00 6:33

It's not nothing close.

Owner Demand Versus Industry Suspicion

SPEAKER_01 6:34

Yeah. God, naming companies is so. Um, but no, it's been it's been fantastic in terms of just the feedback and all that. So it's been good.

SPEAKER_00 6:51

What and yeah, in terms of some of the some of the feedback, any patterns of response from from owners and then and then from the industry, like anything that yeah, you think would you would call out as a as a trend line in terms of how people were receiving the business and you know what they were excited about, what they were skeptical, what they were skeptical about in the near term, some of those things.

SPEAKER_01 7:14

Yeah, you know, I I think what's interesting. So owner response has been tremendous, like just lots and lots of inbound. You know, I mean, a lot of it's not like, hey, I want to hire you. A lot of it's like, hey, can you come to a lunch and learn from my team? Can you come speak at our conference? Because I think other people will want to hear what you're working on, like stuff like that, right? Not a lot of, hey, send us a contract stuff, obviously, but so but the lots and lots of positive signals. I think the recruitment, we got like 3,500 resumes in the first day. That is with our AI filter running, right? So, not a lot of nonsense there. So we just had coincidentally, like we thought it was gonna be kind of a problem. We had two talent acquisition people start this past Monday, and they were like, Okay, we're in the in the ground running, are we? Yep. So that's been fantastic. The the interesting thing on the kind of AEC side of things, there's been two types of responses, right? One, like, hey KP, it looks like you're trying to push down our fees, our costs. Looks like you're gonna be the bad guy in the room these days, so to speak, which I hate hearing that from my friends, right? These are people I've known forever, like that, you know, I'm gonna use AI to beat the crap out of them or something like that. And I was like, I'm like, if you're overcharging, yes, but like in in the world of fairness, no, like, you know, nobody's ever done well by going with the lowest bid. It's never happened, right? We we know this to be true. So I think I've had to defend that a little bit. But the more interesting thing is some of the AEC firms that we know and love that have very strong innovation ambition, they've been very, very positive. And what they've said is like, you know, our problem is when when there's an owner's rep involved and we show up with some level of innovation, it gets discounted. Like, who cares? Apples to apples. So one of our investors is a structural engineering firm, and they said, look, as a standard practice, we provide reinforced concrete design, struck steel design, and wood design. That's just what we do. It's part of our process. We've got some tools that make it fast for us. And they said they go to that with clients saying, hey, here's how we show value. We preemptively have created optionality so that later on, if it prices out too much as steel, you don't have to, we don't have to go back to the drawing board, right? And they said, like nine times out of 10, an owner's rep will say, Yeah, that's cool. Don't do that. I need to compare apples to apples. And the other three engineering firms are only designing steel. So please go redo your proposal. So what they said is like, I feel like as an owner's rep, you guys are going to be much more open to the long game of innovation and not putting everything into apples and apples and oranges and oranges, but actually seeing like what the actual value that the innovation could bring. So it's like that, you know, the best client's an educated client. That's how they're feeling about it, right? Which I I I thought that was great. That was positive news, right? Because a lot of times they feel like they have to eat the innovation. All these firms have spent so much money investing in innovation. And CFO says, What's the ROI? Tell us, tell us what project you got because of innovation. And they they can't point to anything. Yeah.

unknown 10:48

Yeah.

SPEAKER_01 10:48

So that's been that's been different.

SPEAKER_02 10:50

Yeah.

SPEAKER_00 10:52

Any any near-term goals or you know, basically, I'd say maybe previews that you want to share with the audience what they should be on the lookout for, just uh in terms of everything zero.

Cloud Code For VC Code Review

SPEAKER_01 11:08

I mean, I think you'll you'll start to see more acquisitions and then you know announcements, a lot of team announcements coming up, which I think is fantastic. Near term goal, fill fill in the roster, fill in, fill in the team out. Yep. Yeah, yeah, yeah. I mean, I think like if you look at our roster right now, I mean, I think we're making 70 hires between now and the end of the year is our hiring plan. Not trivial. So um yeah. So yeah, no, I think that I think between acquisitions, we're gonna start you know, you can't buy our tech, so to speak, right? It's part of our service, but I think we're gonna start sharing like our tool sets and our toolbox and what that looks like. So yeah, it's pretty, you know. So obviously there'll be a lot of a lot of news, so to speak. I was gonna tell you switching topics. Um something super interesting. You know, what you and I have talked about cloud code, cloud code, cloud code, right? What I discovered today is cloud code, every VC should be using cloud code for code review. It's so good. For code review? Yeah, I mean, if you're investing in a startup, right, a lot of times it's like, well, you know, they have the right person. You know, Matt on our team used to do some cursory stuff, right? But using curse using cloud code for code review, it's very, very good. So the reason I bring that up is if you're a tech startup, right? Not that a lot of VCs listen to this show, but maybe they do. I don't know. But if every VC at pre seed, seed stage, one person, two-person firm that doesn't have a techno, you know, we've always loved having Matt around because hey, we have a CTO. We're we're I mean, how many venture firms have a CTO at our size? Very few, right? And we always thought that that was something kind of special at Shadow Ventures. But now think about it: if every VC has code clawed code and can do code review, especially in a day where maybe people are slepping some vibe-coded thing or whatnot, founders better have their shit together.

Measuring Teams Through Work Tools

SPEAKER_00 13:35

Yeah. Do you I've actually thought a decent amount about some of these, like how diligence might change and how it maybe should change given the tools that we have access to. And I thought about it less from the code review standpoint up until this point. I actually think it's a great idea. And I was thinking about it mostly from a team, a team organizational, basically figuring out what is happening behind the scenes operating-wise. What's their leadership, what's the founder's leadership style? How hard are people working internally? And the reason I think it's it's a it's an interesting time to try to solve that problem is a lot of teams really over the last decade, like we've obviously seen a major shift to remote work, right? And you have these these platforms like Slack or Microsoft Teams, GChat, where there's like this repository of data on everything that's happening organizationally at the company. Like, you know, like basically I think there's like a way to measure the performance, like organizational performance and understand exactly what the leadership style is of the founder and their their capabilities, their strengths and weaknesses, communication style, just by like looking at looking at those, looking at the text. And like there's no better tool that that basically that was just unleashed to the world than like you know, LLM, which is fantastic at digesting lot large amounts of raw text that you could extract from these databases. So I've actually like worked on I've been working on building that product. And I think the biggest thing I'm running into, as I've talked about it with our founders, and as I think about, as I think about, you know, any sort of scale that this product might have diligence-wise and venture or PE, is permission from the founder. Like, are they like nearly every founder I talk to about this, they know I they they agree, they validate that the idea is like is incredibly good, and then they're like they start blushing. They're like, holy cow, I can't. But not with me, but not with me. Like if you got access to that, like I'm I'm really concerned.

SPEAKER_01 15:53

Yeah, and it's nothing about that, yeah. And it's nothing about that. That's different. That that's like a different level, right? But but think about the spreadsheet we send out, which a lot of VCs send out, like, hey, tell us what's your revenue run rate, like all of our reporting functions. And there are some firms that require, I forgot what that there's some product where founders have to put their info in. Forgot what it's called. We've never deployed it. But what if we said, like, hey, our MCP server needs to be connected to your accounting system and your CRM so that we can get real-time reporting about how you're doing. I don't think that's technically that intrusive because it's facts and data, and it maybe drives a discipline. Like, don't throw BS in your sales pipeline. In fact, if I was a founder, I don't know that I would mind that. Because I'm gonna go to my head of sales and say, hey, look, you know, our VCs have access to our data. Don't put any BS in there, man. Like you can try to BS me, but these guys are gonna ask me about it. At least that's more analytical. I understand like Slack, like, hey, you're gonna monitor my email. Thanks. Like, you know.

Permission And Term Sheet Implications

SPEAKER_00 17:04

Yeah, yeah. I mean, yeah, there's there are permissions to be, you know, to be concerned about. Like, we, you know, the intent would not be to read DMs, like we're not gonna read private messages. But any public channel that's like a work-driven channel, I think it should be fair game, you know, in terms of basically, you know, almost like for personality, leadership style, strengths, weaknesses. Like to me, it actually should replace the personality test, like the questionnaire that a lot of VCs will send out to founders. Um, just show me like, like, let me just connect the software product and like I'll just I'll I'll do the assessment through what has actually occurred in the business. But the intent is like not to not you know, not to you know, infringe on where all this stuff manifests itself, right?

SPEAKER_01 17:50

Is in information rights. So do you start rewriting term sheets? We should talk to our lawyers about it, right? Like if you're on the other side of it and then our information rights said, we have the right to connect to your Slack, we're gonna connect to your accounting system and your CRM. Yeah, blah, blah, blah, right? Like, I mean, I have a friend who in information rights gets a view only log into the bank account. That's part of his deal. Yeah, he's like, look, I mean, this idea that founders manage cash well, like this I getting the call, like, hey, we're about to run out of cash in three weeks, can you help us out? Like, he was like, it's happened to him way too many times that he gets like gets it from the, you know, gets it right from where he wants to see, which is like the bank account. Yeah, and you know, I mean, if if you include it in information rights, you know, if they agree to it, now they might try to negotiate it out or whatever. Yeah, but you know, it is interesting. Like, how does this maybe how does some of this like access to data? Yeah, exactly.

SPEAKER_00 18:52

Yeah, the the the ability to to connect to data data repositories is you know now abundant, right? Like we have that ability to connect any system, extract the data in any way that we want. And so yeah, there's like going to be, I think, a reckoning on what who has permissions to what. And I think I I could see V, I could see V there's gonna be groups of investors that build a business around not infringing on you know, on data. Like there's, I mean, this this has already occurred to some extent, the founder-friendly movement of like, hey, I'm not gonna put any egregious info right terms into the term sheet. I'm gonna, it's gonna be standard vanilla. I'm not gonna, I'm not gonna require a board structure. I just get out of the way. I write you a check and I get out of the way. People have built a fundraising or you know, a entire fund strategy around that idea. And then you have like the, you know, you the other side of the coin, which is no, we're gonna like, we're gonna run this business properly, we're gonna uh install proper governance, we're gonna have a board, we're gonna do checks and balances on bank accounts, we're gonna like, we're gonna get in the weeds with you and try to help you build the business.

SPEAKER_01 20:05

Yeah, I think what's what starts to get I have a friend that runs a Series A plus firm. He has a CFO and he has an entire accounting team, and they spend a lot of their time collecting quarterly data from all of their portfolio companies because their LPs require it. So he's got like 10 accountants, right? They just sit there and collect data from all his port code. And he was like, Yeah, like unfortunately, we have some really big institutional investors, and you know, we have to comply with things for them. So we, you know, in a slater stage, but the you know, the real question would be what's the value, you know, or pre season, like what's the value in like tracking anything? Like, I'm gonna track your financial screen. Yeah, yeah, you know, it's it and to your point, like the founder-friendly movement, so to speak, I think applies in super early stage where you're just betting on a founder and their judgment and all those things, you know, or people that have been there done that, but you know, at some point you get to series A. And you know, unless the LP start telling people, like, oh, don't worry about it. You know, if there's a GP friendly movement, then maybe that trickles down. But you know, at A, B, and plus, I mean, there is no GP friendly movement. It's like a bunch of lawyers and accountants telling everyone what to do.

Construction Robotics Funding Reality

SPEAKER_00 21:25

Yeah. Yeah. It's gonna be interesting to watch that play out. Yeah. I have a few other Yeah, go ahead. I was just gonna say, I have a few other topics to hit on today if you're if you're open to progressing the conversation. First is construction robotics and funding trends. By the way, I think I've cracked the X algorithm. I've been spending more time on you've been hosting a lot. I've been spending more time on on X, formerly Twitter.

SPEAKER_01 21:55

So dropping some Frank Lloyd Wright stuff. I was like, what is this? What is his AI agent doing for him out here?

SPEAKER_00 22:07

Feed feeding me interesting data on everything, our industry at all times. So yeah, but anyway, one of the things I posted yesterday was just a a general observation that we've covered a lot on this podcast about how basically the right business model choice for robotics companies. I feel like we talked about this like on many episodes, right? Which is like the to be a vertically integrated subcontractor and perform work, the choice to sell machines, lease machines, the choice more or less bifurcating into like, do you want to invent products, engineer, you know, in engineer machines, or do you want to run services in the field? And most companies that that I see that we see, it's very easy to to put someone in this in a specific camp based on what they've done historically, like looking back. So when I when I posted this, anyway, like my it was basically my this post was my attempt at at addressing the investor concern on like what is the right business model? And what and and frankly, like what is this, what, what should it be if you're a founder? Like how should you think about this, knowing that there the biggest risk to your business is both not getting enough traction, which is like potentially a business model issue if the, if the, if the, if you have a great product, but but the the the the market doesn't want it because you haven't structured the the company the right way, the product the right way, or capital. The other big risk is capital, which is how do you keep investors excited knowing that the industry we're operating in is conservative. We do buy it, you know, there there is a slower nature to how deals get done and how they you know how contracts land and expand. And we've seen just a lot of a lot of companies that have built good product not build a great business. And I don't think it's because they could they they they they couldn't I think it's like I I think it's I I I look at the flaws and and maybe some of this is investor driven like what the investors pushed them to do, the choices they made early business model wise that maybe were not were not great choices. And it left them in a position where they didn't have enough traction and the story the narrative about the company and what they were and what they've done historically didn't meet the standard that the next investor the next buyer needed to see. And so that was the intent of the post but I was going to share my screen and just show you the the funding data that I pulled when I was looking at this because uh it's quite it's quite interesting and I just wanted to get your take on it.

SPEAKER_01 24:56

And for the listeners just in case you have been switched over to Nick unpacked yeah you're welcome it's tired from the time it's pirate radio here. It's now Nick unpacked let me tell you about what I posted on that one back.

Reading The Capital By Category

SPEAKER_00 25:14

Stealing stealing the mic here. All right so I have a question for you Nick Yeah can you see this? I cannot oh I think I need to add you there you go there we go how about now yep okay so historical funding data and construction robotics probably interesting for the audience to know 1.78 billion in in capital raised the categories here 3D printing has raised 650 million ish excavation and earth moving notably bedrock robotics who's the who's somewhat of a new entrant in the last two years they've raised 350 million you get on the list you have the layout and surveying robots dusty robotics has raised 70 million you have some prefab and and offsite companies autoval is one of those drywall and finishing you have canvas who raised 43 okos a uh a player in that in that market kawazo material handling so we have by the way super excited about kawazo like those guys a lot we could just never we could never get in like just never we could never I don't know there's always some reason why we couldn't do that deal but love the founder super when I saw that press release come by that he raised 25 million I was like cheering them on super excited about that. Yeah good call out so yeah as you look at as you look at these categories I'm curious to just to hear your your reaction are you are you surprised by that when you look at the trades do you think that like it's somewhat of a small pool of money when you take out the 3D printing companies you take icon out you take icon out and you take bedrock out like it's just it's not that much capital that's been much there's too much right yeah but yeah any other any other reactions no I think I think all that's about right you know so I think we've spent a lot of time on robotics and just like you put up the screen right the the capital stack and how we think about funding them.

Business Models That Actually Fit

SPEAKER_01 27:18

But I but I think there's a different way to look at some of this right I think we should look at it I I think we beat a dead horse around like hey capital comes in different forms and if you're building robotics you should probably be thinking you should need to expand your thinking beyond venture capital right we've we've we've we've covered that pretty well in the past but I think there's there's a couple things to start from which is how is it done today right who are the players that do it today and then the cost of your product right and then so I'll say is the existing models the cost of your product your build the amount of customization required to your robot the additional labor requirements to operate it and then what percentage of the work does it actually complete so if you take the CSI nomenclature and you start picking off CSI codes you start saying okay for this CSI code here's what it looks like today. Now the reason I bring that up is in the world of specialties right there are companies that all they do is core concrete it's all they do they core concrete that's what they do for a living they've always done it they run around these vans ABC coring XYZ coring and they do coring right so there's a company I met I think it's called August Robotics they drill holes for data centers right so they drill the little holes for data centers to put up their server racks right that's that would normally be someone that has a coring business right these mom and pop coring businesses. That's a very narrow use case so if I was if you could sell that product to a coring and concrete coring business it doesn't require a lot of configuration you could probably if it was priced right you can by the way priced right either you can sell it to them or you could probably lease it to them or rent it to them right and they can operate it they don't need an operator to stand behind it. It does a hundred percent of the job right it's drilling holes does all the job right it's not like oh I have to come back through with my labor crew to finish the things it couldn't finish and it fits into an existing model right so I think that's how we've got to look at some of this stuff. And I think the problem with you know so icon you know one of ours challenge with icon is it 3D prints a house it's very custom you have to have someone there to monitor it's you're dependent on like concrete delivery and a materials play that's complicated and it's not 3D printing cabinets right so it can't even do it can't do the whole building yeah right and there's not an and so like if you put out an RFP to find concrete masonry for residential you're not going to get a lot right there's not a lot of concrete construction done in residential right there's not a lot of in place concrete done in residential and then there's like you know truck pulls up and pours a slab right beyond that there's not a whole lot of talent specific to that so which also kind of shows you why icon had to raise so much money right there is no ecosystem that understands this stuff so I think that's part of it. I mean I think the with the technology today and what you can do you're seeing a lot of good core robotics systems right you can buy a rover system you can buy a gantry system there's all there's there's a lot of out of the box systems that are coming to market. Yeah um you can build proprietary end effectors not that hard to AI design and get fabricated you see that all over X like hey look what I 3D printed you know kind of thing right so I think those are trends and then just the programmability of motion control with robots is getting easier and easier right you you know you don't have to learn ROS you can get claud code or you know look at things and build stuff right I mean I'm simplifying a bit but it it's definitely an unlock so that unlock means that you can now build kind of niche playing like can you start a construction robotics company for a million bucks maybe right maybe if all it does is score if if it's a rover that just goes around the job site with a little magnet underneath it to pick up all the nails and screws that people drop around that are a safety hazard probably probably yourself and run it right people would probably buy it there's nothing to do it's like buying a Roomba right unless free business idea for everybody. So I think that's that's where I think things are changing.

Distribution Partners And OEM Exits

SPEAKER_00 32:06

So I think we'll see an increase in funding in robotics but the funding sizes may go down right so higher frequency lower amount might be the thing do you think that because it's because many of those categories in companies are tied to services they're either selling product selling product to subs that are performing the work or they're maybe they've decided to do it themselves. Do you think there's a different allocation strategy that a venture firm should have when backing that company versus kind of more of a standard maybe maybe like a good example here would be Icon who's actually building brand new brand new machinery inventing a category selling selling basically the best 3D printing equipment on the market versus just performing the work.

Cognitive Overload From RFIs

The AI Agent Arms Race

SPEAKER_01 33:01

Yeah I look I think it's all going to come in shapes and sizes but I think ultimately you have to understand the markets really well you have to understand the technical requirements really well and you have to build distribution very quickly. I mean if you can build something that's highly repeatable that Milwaukee Tools can put their logo on think about the distribution. Right? But you have to understand what distribution at scale looks like you have to understand like it can't be complicated right if there's a battery system you should probably comply with whatever batteries so I think you're gonna see some of these interesting OEM relationships start to evolve in some of these markets where I mean who bought was it canvas that got bought yeah by JLG? Yes they make scissor lifts and stuff right I don't know what they paid for it hey if anybody actually knows send us a note I don't think we'll need we know what they actually paid but man the distribution JLG has massive massive distribution right so in their mind like they're not just gonna pay for the IP they probably don't care about the revenue the revenue is just proof points that the product works but they mostly care about the IP so they already have the ability to produce at scale the product like to manufacture the product and then they have massive distribution yeah right so I think I look at someone like Milwaukee tool right absolutely right they had they they know how to manufacture at scale they're a strong brand in the construction market. People love their Milwaukee tools and then they can actually drive distribution but it has to fit in with their in the reform factor. I think Hilti has tried some of that right with mixed results. So I think but I think founders have had to raise a lot of money to come up with a big idea maybe there's like some really interesting niche plays if you sign up with the right distribution partner it gets to be pretty big big yeah yeah super insightful cool okay I'll uh transition transition the show back to KP unpacked thank you for that thank you for indulging me in that brief intermission what other what other what other what other posts did you have this week that you thought had an interesting response that you wanted to talk through you know and for not following on Substack I'm posting a lot more on Substack these days on LinkedIn mostly because I'm just getting spammed like crazy on LinkedIn it's just insane so it's gonna be tough but I I wrote this article about cognitive overload right just that you know if if you had to provide a submittal typing it out on your IBM selectric typewriter and folding it and putting into an envelope to mail it into the client you'd probably do a lot less submittals right and when I talked to I was talking to some engineering firms and we're talking about you know because once again like the zero RFI thing people are like oh I love it love it love it. Okay. So some of our engineering friends called me and we're talking about it they're like they're like KP you know how many like RFIs we get that I what I wish my response could be is like read the damn drawings. Right. And what they're saying is basically we are getting overload the overload of submittals and everything from everybody just like we get overloaded with emails right however the information they're getting has a shelf life if you don't respond you could be subjected to legal action you could be there's just lots of things right so you have to respond and apparently you're not allowed to respond he said he checked with the general counsel you're not allowed to respond like listen dumbass read the drawings apparently it's not an allow you're not allowed to respond that way but you actually have to like look at it and study it and give it back so this this cognitive overload which has been well studied right with fighter pilots and a lot of other airline pilots in general with all the blinking lights and switches and everything else right you just you can't handle it right it creates the level of cortisol and stress it creates is just absolutely massive right so I think with the advent of some of these system of record platforms that we like to beat up on the alerts here's another alert in your inbox here's another alert clear out these alerts right we we and it's so easy to submit anything that people just keep doing it right so I think it's super fascinating in a world where it's a high you know we we we deal with you know construction highest suicide rate is in construction already look already very stressful generally not a safe business right it's not I'm sure some of our engineering friends that are sitting in cubes are probably fine in their air conditioning but in general right the amount of data we're generating and pushing around is is kind of unsafe right what it's doing to people and then if you think about like the the lack of staffing the lack of you know talent you throw that like oh we have a shortage of talent and now we're gonna auto-generate a thousand emails right and so I kind of wrote about that and I got a lot of really good feedback from people saying like hey you know you're right all this technology has actually made it worse it's just too easy there's not enough friction it's made what specifically worse? Just the amount of info the flow of information right people don't have to think through something because there's no friction right they don't have to like look at a situation process it use some level of their own professional judgment all they do is like swivel chair yeah information like oh this person in the field said this great let me forward that over to the to the architect right yeah you're push you're pushing information you're not actually thinking as much yeah that's totally that's totally a real thing i mean in the in in the world that we're living in using agents pretty consistently that's kind of the job like it's like the the the I fe it feels like the game is like how how how little can you think but how fast can you manage with with with real clarity right the job is like to compress your thoughts and be extremely extremely clear when you're communicating with your agents and pushing them information.

SPEAKER_00 40:10

But the more like some yeah like what what would call what would cause the bottleneck is if you're thinking too much. So don't overthink it is like basically the the feedback loop I I that I that I feel don't overthink it. Be clear be clear but don't overthink it move fast. And when you're on that treadmill it does yeah it does it does really it catch it catches up with you when you're moving that fast.

SPEAKER_01 40:34

Right. So you start thinking about it. So if Nick's gonna start sending me deals right and I you're gonna start forwarding me stuff hey KP look at this deck right and your agents are doing it and you have 50 agents like processing your email and shooting me stuff right then I have to have 51 agents to keep up with you shooting me stuff. Yeah right it's like spy versus spy like how many agents do you have I need just as many to kind of respond to your agents right because you know the internet and email and all this stuff has created a lot of a lot of information not a lot of knowledge right I was on I was on an interview this morning and I said look you know if we if AI can help us sift through all the information overload overload that we all deal with and then we actually get to use human judgment with a set of facts that's massive versus just responding because you know it's like text messaging that's what sometimes drives me crazy with text messaging some of my friends just like no context no decision I'm like are you just sharing dude like what like what what is my action item here you know why are you make why are you putting the work on me which I think my wife actually thinks I do that to her because I will text her style message her stuff right just like moving all the stuff to her desk. Just move all the stuff right forward my entire inbox to her and it's like it's your problem now she's like thanks man like so anyway I think I think it's with AI I mean I think that's gonna be the other thing right like there's a lot of people not using agents now what happens right against the people that are using agents.

SPEAKER_00 42:21

Yeah defense we need defense agents we need offense agents and crazy world out there. Yeah because ultimately as humans we should really be judgment and exception handling right where the agent can't handle it and it like a agent does rejected it I don't know what to do with this right yeah that's really that's really interesting that top that topic like yeah it sounds my guess is that becomes more of a more of a a a hot hot topic hotly discussed item as we move into the future which is just like how do we all manage the yeah manage the the massive the massive amounts of work that are being pushed behind the scenes and we're you know we're we're getting in the mix only when it matters but the the importance of the work didn't change in fact you're doing more important work in most cases and you're taking on more of it. And so like the magnitude of that decision when you are called and when it does show up on your desk and how you sense how you synthesize that information which is basically just managing the cognitive load right like that's a that's a really interesting idea.

Separating Real Builds From Hype

SPEAKER_01 43:33

No and I think the other thing is as we start thinking about that you know I I love watching X right the flow of like I just did this with open claw it's insane. It's like yeah whatever right most of these people posting don't have that much work to do yet they're trying to automate all their work with AI right but but it brings up a thing where like if you're very busy how do you find time and space to do all these automations and literally the people with that aren't doing that much have plenty of time to build these things out and then they're just going to keep creating your workload right you got to step out of the flow for a minute to you know to to build your weapon so to speak to respond to all these people but I try to explain to people like actually with my tech team I'll see stuff on X and it's like oh this is really cool. And Barry said like hey I love the way you post things I post it to the engineering team I'm like real or fake everything I post is like real or fake Fake. Because there's so much fake stuff. I mean, my god, like, you know, open I used Open Cloud Options trade and made a million dollars today. Oh, right, right. Yeah. It's insane. It's like, no, you're insane. Like, what are you doing? Right? Like it's you didn't do any of those things.

SPEAKER_00 44:51

Join my telegram where I'll provide trading advice.

SPEAKER_01 44:54

Yeah. Click this Venmo link and I will. Yeah. All right. Next. It's been a good one of