บทถอดความฉบับเต็มของ Buildots at a Billion, Autodesk's Fragmentation Problem, and a Fine Nobody Will Feel - September 2026 Reviewed as Noted

coming on. I think, you know, September was an interesting month. 00:02:15.326 --> 00:02:18.916 We had some stories come out. I think just looking at the stories we've got 00:02:18.916 --> 00:02:22.316 here, so we've got BuildDot's kind of large valuation spike. 00:02:22.316 --> 00:02:25.909 We've got some new agentic AI stuff coming out from Autodesk. 00:02:26.437 --> 00:02:30.796 You know, potentially a little bit like Groundbreak, seeing kind of what's actually 00:02:30.796 --> 00:02:33.622 real in the strategy will be interesting over the coming months. 00:02:34.174 --> 00:02:40.061 New Jersey, letting unpermitted generators keep running on a data center campus. 00:02:40.455 --> 00:02:45.116 And there's a fine associated. We got Dodge's wild kind of mega project driven 00:02:45.116 --> 00:02:48.691 construction starts and record low construction unemployment, 00:02:49.243 --> 00:02:53.672 along with Caterpillar and Field AI announcing a partnership, which is interesting. 00:02:54.171 --> 00:02:56.586 So I wanted to start with BuildDots. 00:02:57.091 --> 00:03:02.416 I guess, Aaron, to maybe kick us off here, everyone's tried to solve for progress 00:03:02.416 --> 00:03:07.755 tracking in some capacity. The reality capture field is pretty crowded. 00:03:09.407 --> 00:03:13.987 What makes BuildDots kind of differentiated to allow them to get this $1 billion 00:03:14.347 --> 00:03:16.446 valuation, in your opinion? 00:03:17.083 --> 00:03:22.827 In my opinion, so for clarity, we're not investors with BuildDots. Okay, yes. 00:03:23.447 --> 00:03:28.237 And I'm not like that close with the team. So anything, my opinion here is just 00:03:28.237 --> 00:03:31.425 like me as an observer, not me for our fun. 00:03:32.127 --> 00:03:37.257 So here's what I have observed from the sidelines as not being a direct user 00:03:37.257 --> 00:03:39.367 of GoldDots, but, you know, kind of, 00:03:40.047 --> 00:03:46.087 through the channels, is they built a more comprehensive link between how do 00:03:46.087 --> 00:03:53.099 you tie out the video walkthrough into a reporting function with less hands-on keys, right? 00:03:53.337 --> 00:03:58.207 Are those hands-on keys worth the billion-dollar valuation or not, 00:03:58.747 --> 00:04:02.669 I think is an interesting question from a pure tech side. 00:04:03.069 --> 00:04:05.861 And I'm not close enough exactly to articulate that. 00:04:06.503 --> 00:04:11.602 I do think that it sounds like there's some wind in their sails from the data center push. 00:04:12.561 --> 00:04:17.955 But, you know, just and maybe Stephen, why you like my approach to venture, 00:04:18.268 --> 00:04:22.860 I'm always similar to a contractor worried about diversity of your client base 00:04:23.185 --> 00:04:24.539 and when the other shoe falls. 00:04:24.539 --> 00:04:29.899 Right. Most great GCs are thinking hard about what's my what's the next capital 00:04:29.899 --> 00:04:35.381 plan that I want in my backlog so that as the centers eventually dry up. 00:04:35.798 --> 00:04:38.306 I don't know if that's been three years, five years, or eight years. 00:04:38.818 --> 00:04:47.460 They have enough work to continue a cash flow generating, reasonably profitable, 00:04:47.571 --> 00:04:49.010 high-employment machine. 00:04:49.475 --> 00:04:54.569 And so I always look at these big valuations saying, after data centers, 00:04:54.569 --> 00:04:55.780 what's next for progress? 00:04:56.123 --> 00:05:01.209 And so that's kind of like my big quandary as I look through some of the valuations 00:05:01.209 --> 00:05:02.269 that we see in the market today. 00:05:02.594 --> 00:05:07.132 Noting there's a lot of billion-dollar companies and, you know, 00:05:07.341 --> 00:05:12.165 Tahuwa has been around 17 years or something like that, just crossed, 00:05:12.351 --> 00:05:14.563 you know, rumored over the billion-dollar valuation. 00:05:14.957 --> 00:05:18.028 Very different business, you know, very different timeline to get there. 00:05:18.464 --> 00:05:21.313 But, you know, it's certainly not unheard of. 00:05:22.042 --> 00:05:24.751 I mean, it's also interesting to maybe, you know, argue like how much of this 00:05:24.751 --> 00:05:29.191 is just inflationary of like the monetary supply is expanding. 00:05:29.191 --> 00:05:34.281 And so valuations are kind of going up accordingly. But I do think there's there's 00:05:34.601 --> 00:05:35.763 real tech behind that as well. 00:05:37.161 --> 00:05:41.381 Hanu, from the GC side, obviously, you guys are potentially obviously a user 00:05:41.701 --> 00:05:44.036 of many of these reality capture systems. 00:05:44.356 --> 00:05:49.221 Do you see anything kind of intrinsically true about build dots from a product 00:05:49.221 --> 00:05:51.369 set that's differentiated from what you guys have used? 00:05:52.083 --> 00:05:55.781 Yeah. I mean, I think you mentioned earlier, like it's a pretty saturated market. 00:05:55.781 --> 00:05:58.671 I mean, I think the one thing that the funding tells is just like, 00:05:58.671 --> 00:06:01.021 look, investors believe in progress intelligence. 00:06:01.021 --> 00:06:04.424 It's becoming an important category. And I actually do think that, 00:06:05.090 --> 00:06:06.541 I mean, this is kind of the future, right? 00:06:06.541 --> 00:06:10.191 I mean, the tech is real. It's not perfect, right? 00:06:10.191 --> 00:06:13.291 And the funding is real too. I mean, they're making real money on these jobs 00:06:13.291 --> 00:06:14.891 and people are willing to pay for them, right? 00:06:14.891 --> 00:06:19.721 We've actually just gone through year-long evaluation of a couple of these different 00:06:19.721 --> 00:06:24.241 solutions like BuildDots and Doxel and Cupix. And there's, I could probably 00:06:24.241 --> 00:06:25.664 name a dozen others, right? 00:06:26.250 --> 00:06:30.441 And I think we are trying to solve a real problem, right? I mean, 00:06:30.441 --> 00:06:33.651 progress tracking on construction projects is

pretty inconsistent. 00:06:33.942 --> 00:06:39.151 And it's, you know, the reality capture isn't the real hard problem really to 00:06:39.151 --> 00:06:43.071 solve anymore, right? The problem is turning that observation into something 00:06:43.071 --> 00:06:47.621 useful to be able to make operational decisions quickly enough that we can actually 00:06:47.621 --> 00:06:48.906 still change an outcome. 00:06:49.758 --> 00:06:57.801 But I think we're still trying to really see what the ROI actually is. I mean, based on our. 00:07:00.421 --> 00:07:03.639 Early evaluations, obviously it's value in the documentation itself. 00:07:04.515 --> 00:07:08.565 Progress reporting, schedule management, like, but those are three different 00:07:08.565 --> 00:07:12.131 things and you shouldn't automatically treat them as the same ROI. 00:07:12.270 --> 00:07:17.375 But, but just getting everything in one place is like a single source of truth. 00:07:17.375 --> 00:07:21.875 It's like a good start, but you know, it's, you know, these automated, 00:07:21.875 --> 00:07:27.555 it's not an easy button, right? Like there are no automations that are magical human in the loop. 00:07:27.915 --> 00:07:32.158 It still takes, you know, anywhere from like 24, 00:07:32.835 --> 00:07:39.155 to 48 hours to get information so there's a challenge of this lag and it's not 00:07:39.155 --> 00:07:42.607 really the primary differentiations between these solutions is just, 00:07:43.214 --> 00:07:47.625 how quickly can they turn it around how cheaply can they do it, 00:07:48.211 --> 00:07:52.205 and you know how accurate is the information because the reality is is if you 00:07:52.205 --> 00:07:55.781 still can't measure something you can't capture with a camera right so 00:07:56.216 --> 00:08:00.385 it's not a perfect solution with 100% accuracy by any means but it's getting 00:08:00.385 --> 00:08:03.380 a lot of traction and I think you know we're pretty bullish that, 00:08:03.978 --> 00:08:08.515 this technology is here to stay, and we're continuing to work with all of them 00:08:08.515 --> 00:08:09.835 to try to make them better. 00:08:09.935 --> 00:08:13.215 All right, spots to break. Lars, real quick, when was the last time you got 00:08:13.215 --> 00:08:17.699 a new set of drawings from an architect and they didn't cloud anything? 00:08:18.070 --> 00:08:23.021 Well, the one that cost me a couple hundred grand was back, I think, in 2017, so... 00:08:24.183 --> 00:08:27.772 Lesson learned there. Yeah. So, you know, what I like about what Trunk Tools 00:08:27.772 --> 00:08:34.462 is building is they're looking at tools to help kind of remove the mundane from the job site. 00:08:34.462 --> 00:08:39.132 And, you know, same as you, every time there was a bulletin or an addendum, 00:08:39.452 --> 00:08:42.982 I would go back to designers and ask them to at least highlight or give me a 00:08:42.982 --> 00:08:44.032 cover letter on the changes. 00:08:44.352 --> 00:08:47.772 And that was always like pulling like teeth, but Trunk Tools is now automating 00:08:47.772 --> 00:08:51.784 that. So you get a bulletin come in, it automatically highlights the changes. 00:08:52.173 --> 00:08:52.992 You don't have to do a thing. 00:08:53.503 --> 00:08:56.732 Well, it's also hitting RFIs, right? It's also hitting submittals. 00:08:57.232 --> 00:09:01.328 And when we think about standardizing ops, right, and providing that, 00:09:01.833 --> 00:09:07.912 you know, sort of resource to those field operators, what I see it as, you know, 00:09:08.586 --> 00:09:14.049 it's sort of like this, like, mentor within the company that structure standardizes, 00:09:14.211 --> 00:09:17.132 keeps it moving in the right direction, looks out for risk. 00:09:18.172 --> 00:09:22.442 And just this last week, you know, I had a QC manager just pass along the submittal 00:09:22.442 --> 00:09:26.366 to an owner without doing a review because he was too busy and it was wrong. 00:09:27.512 --> 00:09:31.777 And, you know, the classic owner statement is, did you even review this? Right. 00:09:32.394 --> 00:09:36.545 So this to me is carte blanche. We absolutely did. 00:09:36.916 --> 00:09:43.092 And think about getting that first pass before your field manager ever even 00:09:43.092 --> 00:09:45.372 actually has to do the first pass. Someone's doing it for them. 00:09:45.372 --> 00:09:50.935 So to me, it's about speed. It increases accuracy and helps standardize ops out in the field, right? 00:09:51.388 --> 00:09:54.732 Yeah. And this is the kind of AI that we've talked about on the podcast, 00:09:54.732 --> 00:09:58.672 really appreciating, right? Which is looking at ways to elevate the craftsman 00:09:59.012 --> 00:10:00.212 at the heart of project delivery. 00:10:00.532 --> 00:10:05.872 Give me time back so I can be out in the field managing risk, right? Yeah, 100%. 00:10:06.284 --> 00:10:11.502 So, all right. If you're a little curious, check it out, trunktools.com forward slash podcast. 00:10:12.048 --> 00:10:14.757 And if you mentioned construction conversations, you get a discount. 00:10:15.228 --> 00:10:19.518 Discount demos are free and we get, and you guys get a discount. So let's go. 00:10:20.872 --> 00:10:26.292 I want, can I push on the valuation? I want to push. So like the valuation side is one side of it, but. 00:10:27.625 --> 00:10:31.075 I think we're seeing a really big trend here with a lot of the investments that 00:10:31.075 --> 00:10:36.496 are being made, because like software that only lives on a screen is really easy to copy right now. 00:10:36.995 --> 00:10:40.365 So project management softwares and things. So I think investors are looking, 00:10:40.845 --> 00:10:43.805 really, really hard at what are those things that reach down into the field 00:10:43.805 --> 00:10:46.615 level where I actually need people to go out and capture something and bring 00:10:46.615 --> 00:10:48.657 it back to me for processing. 00:10:48.953 --> 00:10:52.195 And so

I think that's the reason why we're seeing a lot of these moves right 00:10:52.195 --> 00:10:57.408 now thinking at some of the other acquisitions like drone deploy and things along those lines. 00:10:58.505 --> 00:11:01.775 I feel like I'm just seeing more and more of a trend investors know like hey 00:11:01.775 --> 00:11:06.189 software only isn't as super safe as it used to be 00:11:06.630 --> 00:11:11.445 whereas right now I want things out in the field like I want assets out in the 00:11:11.445 --> 00:11:13.748 field capturing things and driving value, 00:11:14.329 --> 00:11:18.021 so I mean yeah we could argue about the valuation but, 00:11:18.675 --> 00:11:21.467 You know, I can see why the money is being raised here. 00:11:22.285 --> 00:11:23.605 So I guess to that point, though. Well, they need the money, 00:11:23.605 --> 00:11:26.395 too, right? I mean, just a quick note. They need the money. I mean, 00:11:26.395 --> 00:11:31.425 they're spending quite a bit of either tokens to process this data, 00:11:31.425 --> 00:11:33.465 and then obviously they're still this human in the loop. 00:11:34.105 --> 00:11:38.145 And they need a lot of this money to stay afloat, but it's also about training 00:11:38.145 --> 00:11:41.765 these agents, right? That's going to cost them money, and I think they're getting 00:11:41.765 --> 00:11:45.661 smarter, getting better, right? But the vision AI is getting smart better. 00:11:46.385 --> 00:11:52.581 And obviously, the more they train these machines on the construction data, 00:11:53.074 --> 00:11:54.456 hopefully, the better they'll get. 00:11:54.601 --> 00:11:57.065 And hopefully, we'll be able to bring the cost down as well. 00:11:58.626 --> 00:12:03.936 Well, I don't know build lots well enough, but just from the brief research 00:12:03.936 --> 00:12:05.236 I've done, it seems like, 00:12:05.896 --> 00:12:11.656 you know, the kind of the frontier for a long time was, can you accurately kind 00:12:11.656 --> 00:12:15.935 of translate work in progress into digital progress tracking? 00:12:16.440 --> 00:12:21.613 And it seems like they've kind of taken it beyond that to now, 00:12:22.796 --> 00:12:26.926 starting to scratch the surface of kind of predictive forecasting or being able 00:12:26.926 --> 00:12:28.500 to point back at your schedule 00:12:28.923 --> 00:12:33.996 based on trends inside its kind of progress tracking brain what that's going 00:12:33.996 --> 00:12:37.356 to do to your schedule have you seen any results from that or you heard any 00:12:37.356 --> 00:12:41.178 kind of positive stories hanu from the field of like hey We're not seeing, 00:12:41.800 --> 00:12:44.331 some kind of beginnings of predictive. 00:12:45.526 --> 00:12:50.346 I think we're not there yet. But I mean, if we can get to a place where we can 00:12:50.346 --> 00:12:55.316 get perfect information about what was installed yesterday, like on a consistent 00:12:55.316 --> 00:12:59.536 basis, because right now a lot of the

capture is done on a weekly basis, right? 00:13:00.326 --> 00:13:05.206 There's some constraints around doing it daily, whether it's people or whether 00:13:05.206 --> 00:13:09.556 it's just processing time, right? They can't, you know, none of these solutions 00:13:09.556 --> 00:13:11.776 can turn it around really fast enough today. 00:13:12.083 --> 00:13:14.716 That's why it was my comment earlier was like, who's going to get there first 00:13:14.716 --> 00:13:18.336 where they can give you that information the next day is really going to be 00:13:18.336 --> 00:13:19.937 a big differentiator. So. 00:13:20.841 --> 00:13:23.734 And tying that to a schedule, because, I mean, obviously the capture doesn't 00:13:23.734 --> 00:13:27.084 know when did the work actually start unless you're out there doing it daily. 00:13:27.084 --> 00:13:32.094 So it's really getting to that information that's consistent and aligning. 00:13:32.094 --> 00:13:37.324 There's obviously a lot of alignment between like just schedule and model and 00:13:37.324 --> 00:13:39.784 what's actually being captured, where it's on capture. 00:13:39.784 --> 00:13:44.734 Right. So once we get there, I think it'll be able to help drive those decisions. 00:13:44.734 --> 00:13:48.524 We'll be able to start getting much more predictive trends. Right now, 00:13:48.524 --> 00:13:51.453 it's just telling me was something installed or not installed, right? 00:13:52.103 --> 00:13:56.604 But when we are getting that consistent capture, we can start getting some trends 00:13:56.604 --> 00:14:00.514 and metrics around production and performance, which could then also help us 00:14:00.514 --> 00:14:05.424 sort of do some analysis on, hey, are you likely to meet that end date or not, right? 00:14:05.424 --> 00:14:07.916 So it's getting there, but it requires a lot of setup. 00:14:08.503 --> 00:14:14.559 I waited to minute 15 to spice it up. So taking that, like, you know, 00:14:14.932 --> 00:14:18.404 let's talk about the schedule piece. So we're investors in an amazing company 00:14:18.404 --> 00:14:21.666 called SmartPM, right? And I think a lot of you guys have probably bumped into them. 00:14:22.095 --> 00:14:25.784 So a big question is, is your schedule built correctly, right? 00:14:25.784 --> 00:14:32.271 And like you guys have seen, like schedules are just a disaster in plenty of projects, right? 00:14:33.041 --> 00:14:38.379 And so also there's only like three folks in the general conditions budget, 00:14:38.779 --> 00:14:42.959 you know, for a whole bunch of the work out there in the world, you know? 00:14:43.602 --> 00:14:47.034 And they got to update the schedule and like make sure the supers are paying 00:14:47.034 --> 00:14:50.654 attention and like arguing to get the electrician on the site and you know, 00:14:50.654 --> 00:14:53.006 all the other stuff that comes with just running a job. 00:14:53.534 --> 00:14:57.784 And so this idea of spending a big chunk of your general conditions budget to 00:14:57.784 --> 00:15:01.624 have this like automation for the vast majority of projects, 00:15:01.624 --> 00:15:03.751 which are not mega site data centers, 00:15:04.262 --> 00:15:08.064 is like, you know, I think pretty disconnected from reality, right? 00:15:08.721 --> 00:15:13.714 And that's why I'm like super long on like, let's nail the big chunk of the 00:15:13.714 --> 00:15:15.554 market exceedingly well. 00:15:16.415 --> 00:15:20.944 In the cost constraints that is general construction today, which is relatively 00:15:20.944 --> 00:15:22.365 lean general conditions budgets. 00:15:22.911 --> 00:15:28.019 Really good tech for really big projects that can do stuff in 24-hour turnaround 00:15:28.251 --> 00:15:31.322 might make sense, but I would guess that there's only, let's say, 00:15:31.873 --> 00:15:37.863 200 to 500 of those projects, you know, that hit that at Ballywood, right? 00:15:38.113 --> 00:15:40.098 And there's 10,000 that are like. 00:15:41.219 --> 00:15:46.777 Do I even have project controls? You know, do I understand the trade allocated 00:15:46.777 --> 00:15:51.057 to each one of the line items in my schedule so that when I see it slipping, 00:15:51.057 --> 00:15:53.032 I call the guy that actually makes the difference? 00:15:53.624 --> 00:15:57.517 That's kind of our philosophy, at least. And like, that's kind of where I see 00:15:57.517 --> 00:16:02.229 some of these disconnects between what can tech do and like, where is the market? 00:16:03.023 --> 00:16:06.147 Sponsor break. If you spend any time running a project, you know, 00:16:06.147 --> 00:16:09.304 the answer is almost always sitting in the contract document somewhere. 00:16:09.763 --> 00:16:15.447 The catch is that it's almost never in one document. It's split across the contract, 00:16:15.447 --> 00:16:19.067 the specs, and some flow-down clause three layers deep that nobody connects 00:16:19.067 --> 00:16:20.839 until it's already cost money. 00:16:21.048 --> 00:16:24.927 That's the whole premise behind Document Crunch. That risk doesn't live inside 00:16:24.927 --> 00:16:27.660 any single document. It lives in the gaps between them. 00:16:27.869 --> 00:16:32.366 And most teams never look at all of them together until they're standing in a dispute. 00:16:33.063 --> 00:16:37.617 The average construction dispute in the U.S. now runs about $60 million, 00:16:38.311 --> 00:16:40.685 according to Arcadis. It's a lot of money. 00:16:41.086 --> 00:16:45.057 And the root cause they point to isn't fraud or bad intent. It's people not 00:16:45.057 --> 00:16:46.669 fully understanding what they agreed to. 00:16:46.895 --> 00:16:49.769 So Document Crunch builds a platform around closing that gap. 00:16:49.972 --> 00:16:55.580 Crunch AI reads the documents, surfaces the risks, and it cites the exact page and clause that's big. 00:16:55.928 --> 00:16:59.757 So you're not just taking its word for it. Project Assist lets your team act 00:16:59.757 --> 00:17:03.007

on it, ask questions across the entire project document set, 00:17:03.007 --> 00:17:07.387 hand-draft your submitals, notices, and RFIs right from the chat window, 00:17:07.602 --> 00:17:10.637 and the whole thing runs on your own company standards. 00:17:10.637 --> 00:17:15.386 The office and the field are working off the same playbook for bid through closeout. 00:17:15.676 --> 00:17:20.677 No disconnects. It's live now across the U.S. and Canada. Go to documentcrunch.com 00:17:20.677 --> 00:17:24.030 forward slash construction conversations and book a demo. 00:17:24.947 --> 00:17:27.901 Yes, your point is this is potentially... 00:17:28.882 --> 00:17:36.075 Like a mal direction for where the tech is not necessarily needed it, 00:17:39.037 --> 00:17:43.345 the good thing is to take so take so take what someone said and then like twist 00:17:43.345 --> 00:17:45.387 it into something that they don't want to be said, 00:17:46.155 --> 00:17:48.815 i guess i'm just trying to understand do you think the tech shouldn't be pointed 00:17:48.815 --> 00:17:53.265 in this direction or that the tech doesn't work no no i'm sure the tech works 00:17:53.265 --> 00:17:55.303 right Right. But like, how big is the company? 00:17:55.901 --> 00:18:00.272 Right. So if there's two to 500 projects out there that need this daily reporting, 00:18:00.609 --> 00:18:05.090 how, you know, like, how big is that company? Is that a $20 billion, $30 billion company? 00:18:06.026 --> 00:18:12.358 Well, is your point also that the cost hurdle is such that it can't be pointed 00:18:12.358 --> 00:18:15.082 at every project over a certain size? 00:18:16.178 --> 00:18:19.726 Totally. Well, and the reality of how most projects operate, which are like. 00:18:21.818 --> 00:18:25.703 Just not doing daily walks. Like they're just getting the blocking and tackling done. 00:18:25.854 --> 00:18:29.008 You know what I mean? Right. And like the tech that solves the blocking, 00:18:29.008 --> 00:18:33.278 tackling at a price point that is applicable to everyone in the market and then 00:18:33.278 --> 00:18:35.102 built up with killer analytics. 00:18:36.198 --> 00:18:40.838 Wait do we do we build anything that isn't a large-scale data center at this 00:18:40.838 --> 00:18:49.058 point i thought that's that's that's all we do that's all we do now man that's the entire u.s gdp right. 00:18:50.729 --> 00:18:55.348 Well are there other buildings i'm confused yeah i know i know i uh i changed 00:18:55.348 --> 00:18:57.271 my toilet seal so yeah yeah, 00:18:58.258 --> 00:19:01.228 look steven to answer your question maybe a little bit and then i think it's 00:19:01.228 --> 00:19:05.015 an and like i think you need sort of enhancements I mean, this is why we're, 00:19:05.729 --> 00:19:10.008 kind of living a little bit of this tech boom, construction tech boom. 00:19:10.008 --> 00:19:14.194 I guess the trend has been up since 2020. 00:19:15.278 --> 00:19:19.748 But yes, you need a good plan, right? But

what this is solving is, 00:19:19.748 --> 00:19:22.008 are you even building according to the plan? 00:19:22.008 --> 00:19:26.308 Can we just get the field and the office in alignment with like, 00:19:26.308 --> 00:19:28.498 hey, where are we building and how fast are we building? 00:19:28.498 --> 00:19:31.358 Are we meeting the schedule? I mean, you got to have a good schedule if you're 00:19:31.358 --> 00:19:32.608 garbage in, garbage out, right? 00:19:32.608 --> 00:19:35.608 You've got to start with a good schedule. And I think to, Aaron, 00:19:35.608 --> 00:19:41.665 your point, yes, you need some checks to validate that the schedule is sound, 00:19:42.187 --> 00:19:44.158 makes sense, has the right flow. 00:19:44.178 --> 00:19:48.858 Because what ends up happening, if it doesn't, then the field basically goes 00:19:48.858 --> 00:19:51.478 and builds what they think is best versus following the plan. 00:19:52.444 --> 00:19:56.548 But this is solving a different problem, right? And I think both are valid problems 00:19:56.548 --> 00:20:03.209 to solve for. And obviously there's technology here that we can use for both use cases, right? 00:20:03.592 --> 00:20:06.998 Yeah. How do you sync those up though is just really, I think, 00:20:06.998 --> 00:20:10.233 what you need to set up a good process around regardless. 00:20:11.102 --> 00:20:14.624 I mean, you can't just plug chat GPT into it and just have it like push the 00:20:14.624 --> 00:20:16.987 data back. What are you talking about, honey? 00:20:19.904 --> 00:20:24.904 It's easy. It's easy. It's easy. You just do it. All right. So from build dots 00:20:24.904 --> 00:20:31.254 to Autodesk, I was actually at AU this year. Honey, were you at AU? I didn't see you. Yeah. 00:20:31.604 --> 00:20:34.264 I popped in there for a few days. Yeah. I didn't stay the whole week. 00:20:34.284 --> 00:20:36.044 There were a couple thousand people there. 00:20:36.444 --> 00:20:40.468 Yeah, that's right. I did run into a run of people. I'm surprised I didn't run into you, man. 00:20:40.961 --> 00:20:45.478 I do, I like that you helicopter into conferences. You're like in and out, pretty surgical. 00:20:46.504 --> 00:20:49.484 You're like, all right, sorry, I just landed. I've got a 2 p.m. 00:20:49.484 --> 00:20:53.924 And then I'm going to fly out my private jet later. Yeah, two days max in Vegas 00:20:53.924 --> 00:20:56.432 for me, man. I've been there. Yeah, no, that's a good rule. 00:20:57.244 --> 00:21:02.295 So I guess, you know, one of the takes I had as I walked through the floor was, 00:21:02.834 --> 00:21:09.006 seems like more space on the exhibit floor was taken up by Autodesk than maybe in previous years. 00:21:09.314 --> 00:21:14.729 And there was a big focus on kind of autodesk agentic presence i think, 00:21:15.493 --> 00:21:19.744 the the big question especially after groundbreak last year which then kind 00:21:19.744 --> 00:21:25.740 of resulted in data grids acquisition is how much of what we

saw at adu was was marketing, 00:21:26.238 --> 00:21:30.604 versus like real products did do you get you probably had some conversations 00:21:30.604 --> 00:21:32.520 with autodesk while you're there what was your sense, 00:21:33.535 --> 00:21:37.409 Well, I mean, obviously, we're a longtime customer with Autodesk, 00:21:37.409 --> 00:21:40.849 and we've been in these beta programs for, you know, probably well over six 00:21:40.849 --> 00:21:43.139 months before the announcements, right? 00:21:43.139 --> 00:21:47.129 So they are real. Some of them are obviously ready. Some are still in beta, 00:21:47.251 --> 00:21:49.853 and some of them are still in sort of the vision. 00:21:50.560 --> 00:21:54.629 But, you know, we're pretty excited about them. It doesn't obviously magically 00:21:54.629 --> 00:21:57.689 fix the fragmentation of our construction information. 00:21:57.689 --> 00:22:02.198 But you know i think everybody's sort of building these agents and their platforms, 00:22:02.754 --> 00:22:08.359 you know but you know like before like you know before they're going to be solving 00:22:08.359 --> 00:22:11.329 all of our world problems obviously we got a lot of work to do and just maybe. 00:22:12.389 --> 00:22:15.689 Entering the right information in there making sure that you know the fragmentation 00:22:15.689 --> 00:22:21.234 doesn't create more problems because now we're consuming the wrong information faster right so, 00:22:22.053 --> 00:22:26.319 but i think the fragmentation is the story Like, I think when I think of Autodesk 00:22:26.319 --> 00:22:30.819 and their kind of app environment, I think of like 10 different environments 00:22:30.819 --> 00:22:32.959 where duplicate data potentially lives. 00:22:32.959 --> 00:22:38.139 And so I'd be interested to understand within the master prompt of any agents 00:22:38.139 --> 00:22:43.179 that they're building, how are they handling elements that contain different 00:22:43.179 --> 00:22:47.297 information depending on the forum, whether it's Forma or, you know. 00:22:47.778 --> 00:22:51.039 How are they handling their own internal fragmentation? Yeah, 00:22:51.039 --> 00:22:53.206 exactly. From all the acquisitions they've done over the years. 00:22:54.089 --> 00:22:57.449 Yeah. I mean, the fact that they are sort of building, I guess, 00:22:57.449 --> 00:23:01.949 this sort of an ecosystem of information for the project construction lifecycle, 00:23:01.949 --> 00:23:03.999 I mean, it's a challenge to solve for. 00:23:03.999 --> 00:23:09.253 I mean, I think, but the fact that, you know, they are still, 00:23:09.415 --> 00:23:12.581 for lack of a better term, a system of record, right? 00:23:12.744 --> 00:23:16.809 And, you know, obviously it's how well are you managing that system has to do 00:23:16.809 --> 00:23:19.628 with how well is that information to be available and used for you. 00:23:20.400 --> 00:23:24.835 But you know I don't know that we're anywhere near really being able To get

to like, 00:23:25.691 --> 00:23:31.058 You know, what's the right baseline schedule versus the current schedule? 00:23:31.058 --> 00:23:34.288 Are we building from the contract document from the working model? 00:23:34.288 --> 00:23:38.264 Like, does it really know how to analyze, like, what is the most current information, 00:23:38.827 --> 00:23:41.348 and, like, what actually has authority over what? 00:23:41.348 --> 00:23:44.528 I think that's something that I think we're going to have to continue to train 00:23:44.528 --> 00:23:47.808 these things to get smarter. But it's going to be really good at, 00:23:47.808 --> 00:23:52.568 like, just retrieving information back to you. Think of like a chat TPT built 00:23:52.568 --> 00:23:55.600 into, you know, your construction management platform. 00:23:56.207 --> 00:23:59.448 And, you know, it'll retrieve information for you faster. 00:23:59.448 --> 00:24:02.168 Maybe that's a good problem to solve, too, right? It'll just, 00:24:02.528 --> 00:24:04.798 hey, here's all the information that's available to you. 00:24:04.798 --> 00:24:08.808 And we don't have to go try to find information buried in multiple different 00:24:08.808 --> 00:24:13.558 folders and across the different modules, right? So there are going to be some 00:24:13.558 --> 00:24:17.260 workflow enhancements that are going to come out of that regardless. 00:24:18.368 --> 00:24:23.708 I just imagine this like virtual trailer where all the different agents are 00:24:23.708 --> 00:24:30.168 going to be hanging out complaining about the humans and and they're just disparate data sources like tyler, 00:24:30.708 --> 00:24:33.665 that's already happened that's moltbook dude go check that out right, 00:24:34.488 --> 00:24:38.608 but like i want moltbook for construction i would do that all day long yeah 00:24:38.608 --> 00:24:42.042 that actually would be pretty fun tyler so when When you look at, 00:24:42.928 --> 00:24:46.906 literally everyone has an agent kind of being integrated into their systems, 00:24:47.648 --> 00:24:51.469 and you've got personal agents, you've got probably project level agents. 00:24:52.508 --> 00:24:58.088 What is the future of, I guess, as a part of the architecture of project startup, 00:24:58.088 --> 00:25:03.088 is it going to be this like orchestration layer of who owns hierarchy of decision 00:25:03.088 --> 00:25:07.294 making between the agents? Like how do we have the agents all make sense? 00:25:08.054 --> 00:25:11.343 I've been noodling on this a little bit myself. I was thinking about that last 00:25:11.343 --> 00:25:16.013 night is like, we almost need a manager of all of the agents in a way, right? 00:25:16.333 --> 00:25:20.273 It's almost like a whole other project management practice that you would need 00:25:20.273 --> 00:25:25.723 to do of, you know, how do I get this AI agent to go out and do a specific task 00:25:25.723 --> 00:25:28.749 every single time and then like slap it on the hand when it doesn't work right 00:25:29.126 --> 00:25:31.173 and solve that for everybody across the board. 00:25:31.193 --> 00:25:34.113 I mean, you don't need violence, Tyler. It's not necessary. Yeah, 00:25:34.113 --> 00:25:37.533 actually that, yeah, let's not do that because when it takes over, that's going to be bad. 00:25:37.553 --> 00:25:42.543 Grokbots they'll they'll know yeah yeah hey man physically i know i'm first 00:25:42.543 --> 00:25:47.823 on the hit list for this anyway so i just accepted my fate no i was saying about 00:25:47.823 --> 00:25:52.110 this a lot so have you guys played with grokbot at all you mess with that one yet go. 00:25:53.153 --> 00:25:56.983 Like go check that out because i think it's a really interesting sort of flow 00:25:56.983 --> 00:26:03.064 and it it's similar to an open claw it's simpler or similar to a hermes type agent that you can set up, 00:26:03.611 --> 00:26:06.753 open claw and hermes are really difficult to manage though and they're not great 00:26:06.753 --> 00:26:08.603 for like an enterprise level company. 00:26:08.603 --> 00:26:11.836 It's just, it's flaky. It's got too many things going on with it. 00:26:12.418 --> 00:26:16.365 Whereas GrokBot has been like the only thing that has come out recently that's like, 00:26:16.952 --> 00:26:20.139 it does that, but it can do that at scale, 00:26:20.751 --> 00:26:25.533 for a company to where you can have your chief of staff, you could have your 00:26:25.933 --> 00:26:29.423 marketing agents all in different threads, but then you can also put them in 00:26:29.423 --> 00:26:33.029 a group chat together and they can all communicate with each other. 00:26:33.627 --> 00:26:38.468 So it's like you have all of your employees in a Slack channel and you say, 00:26:38.828 --> 00:26:42.642 hey, how are we going to handle this? And your employees can work together to figure it out. 00:26:43.195 --> 00:26:48.333 It has been one of the, I'll say it's been more eye-opening for me to step into 00:26:48.333 --> 00:26:50.058 that and go, oh, that's where we're going. 00:26:50.360 --> 00:26:52.798 It feels like that's where we're moving. 00:26:53.560 --> 00:26:57.163 And also, if you look at ChatGPT, which they are opening eye, 00:26:57.163 --> 00:27:00.840 they just announced dots, which is a very similar concept to that. 00:27:01.223 --> 00:27:03.516 So we're going to start seeing, I think. 00:27:04.279 --> 00:27:08.139 Like anthropic coming out with their own version of this, Google coming out 00:27:08.139 --> 00:27:12.509 with their version of this to where it's, it's like Slack and you, 00:27:12.509 --> 00:27:17.189 all four of us can be in that Slack channel and we can tag in agents and agents 00:27:17.189 --> 00:27:18.859 are communicating with this as we go. 00:27:19.283 --> 00:27:24.329 Right. So it's going to be like, Hey, Tom, the RFI writing agent, 00:27:24.329 --> 00:27:26.759 you know, go out and do this for me and knock it out. 00:27:26.759 --> 00:27:30.839

That's a simple workflow, but I think that's where it's moving. 00:27:30.839 --> 00:27:33.859 And that's what I was kind of curious about with autodesk is are they doing 00:27:33.859 --> 00:27:36.697 that where they're putting agents inside of a thread, 00:27:37.179 --> 00:27:42.019 and you're able to tag them in and they go out they have no idea it's mine it's 00:27:42.019 --> 00:27:46.989 my thesis or or is it just a chat based thing where it's like where did i put 00:27:46.989 --> 00:27:48.129 this and it gives you a link 00:27:48.449 --> 00:27:52.189 and it's not really an agent workflow it's it's more of like a chat gpt type 00:27:52.189 --> 00:27:53.976 of experience what did they show. 00:27:54.841 --> 00:27:58.709 Yeah i mean i think initially it is that but obviously there's different agents 00:27:58.709 --> 00:28:03.269 being built for the RFIs, for the submittals, for the schedule, right? 00:28:03.269 --> 00:28:06.799 I think it's just a matter of time until you start actually connecting that, 00:28:06.799 --> 00:28:11.279 right? Because I think where some of this agentic AI becomes interesting is 00:28:11.279 --> 00:28:14.353 when it sits on top of like disciplined information. 00:28:14.742 --> 00:28:18.079 And to your point, like the agents can actually start communicating and creating 00:28:18.079 --> 00:28:22.770 like some actual meaningful references. And so you have a powerful reasoning layer on top of those, 00:28:23.388 --> 00:28:26.800 what are some inconsistent inputs but it's maybe able to like rationalize and 00:28:26.800 --> 00:28:30.250 be still be able to give you hey at least options and then there's i think there'll 00:28:30.250 --> 00:28:34.429 still be a human in the loop to make a decision i think that's just uh yeah, 00:28:34.911 --> 00:28:37.810 gonna have to be part of the workflow i think so, 00:28:38.390 --> 00:28:42.050 a workflow and i'll use this just in my personal life because i i play with 00:28:42.050 --> 00:28:45.546 these things on the personal side before i start deploying them into you know work 00:28:45.969 --> 00:28:52.170 and so like for grok i set it up and i have my assistant i say every two weeks 00:28:52.170 --> 00:28:54.994 i want you to look and see if i've got a haircut on the calendar, 00:28:55.540 --> 00:28:58.670 and if i don't have a haircut on the calendar i want you to go to this website 00:28:58.670 --> 00:29:05.230 for my barber book a time haircut every two weeks tyler every three you're barbed so but it it looks, 00:29:06.627 --> 00:29:09.570 it looks on a routine this is more revealing. 00:29:12.705 --> 00:29:17.410 Yeah so it it looks on a routine though and it will it will go to my calendar 00:29:17.410 --> 00:29:21.770 see if i've got two haircuts minimum on the calendar if the second one is not 00:29:21.770 --> 00:29:26.578 there it will go out of its way to book that haircut for me at a certain time that i've specified 00:29:26.927 --> 00:29:32.290 and it does that consistently it's always there it's always working so i have not had to 00:29:32.690 --> 00:29:38.010 physically go and book a hair appointment with Alejandro in weeks. Alejandro. 00:29:38.530 --> 00:29:41.910 Yeah. There we go. A little shout out. We'll put his link in the show notes. 00:29:42.650 --> 00:29:47.700 Alejandro, dark parlor, baby. Let's go. I go to him because he just like gives 00:29:47.700 --> 00:29:50.592 me all sorts of crap every time I go. And it feels like, I don't know. 00:29:50.749 --> 00:29:52.398 It feels like I get to hang out with Eddie. 00:29:53.350 --> 00:29:56.456 Yeah. Right. Aaron, how many haircuts a month do you get? 00:29:57.071 --> 00:30:00.250 I don't know. But when I forget, Lloyd texts me and says, yo, 00:30:00.250 --> 00:30:04.043 man, I haven't seen you in a bit. You want to swing by? You're like, your guy's Lloyd? 00:30:04.647 --> 00:30:10.231 That sounds like a barber. Lloyd's great. Lloyd's got a barber tattoo here. Rides a motorcycle. 00:30:10.836 --> 00:30:16.890 Plays in Blue Man. He does not have a website. And Lloyd is chatting. 00:30:17.430 --> 00:30:18.505 Our barbers are very different. 00:30:19.197 --> 00:30:24.215 Yeah yeah and i don't have enough hair left to be able to justify it yeah yeah yeah yeah. 00:30:27.535 --> 00:30:30.545 I'm showing off my fade over here you know you gotta do that was a good day 00:30:30.545 --> 00:30:32.897 i'm not gonna lie no he did great he did it's, 00:30:34.633 --> 00:30:37.395 clean yeah did you find him on instagram no 00:30:37.835 --> 00:30:41.815 no no those clips and you were like so no i drove i drove past i drove past 00:30:41.815 --> 00:30:44.825 it was like on my way to church or something i was like i stopped by there then 00:30:44.825 --> 00:30:47.505 i've been going there ever since it's great so i love it but Yeah, 00:30:47.505 --> 00:30:50.935 they have a website with a little like booking page. It knows. 00:30:51.295 --> 00:30:53.355 So it has computer use inside too. 00:30:53.802 --> 00:30:59.415 So GrokBot uses a computer that is built into it to go to the page, 00:30:59.415 --> 00:31:03.015 navigate, click through all the buttons, book it. Do you arm it with all your 00:31:03.015 --> 00:31:05.452 sign-in information already? It doesn't have. 00:31:05.818 --> 00:31:08.555 That one doesn't have sign-in information because like you don't need to sign 00:31:08.555 --> 00:31:11.565 in to book a time. Yeah, you just sign up. It's like a calendar sort of link 00:31:11.565 --> 00:31:13.684 where you just. I would arm it with a name. 00:31:14.320 --> 00:31:16.955 Huh? That's all I need. You armed it with your name. Oh. 00:31:17.455 --> 00:31:22.735 And even more fun too is that I had an issue, like a clash with when it was 00:31:22.735 --> 00:31:25.291 scheduled and I said, hey, go change the time. 00:31:25.854 --> 00:31:29.245 And it went and rescheduled it for me. I didn't have to tell it to do anything 00:31:29.245 --> 00:31:32.396 specific. It

just, it knew where to go and how to get that done. 00:31:32.739 --> 00:31:37.075 So I say all that to illustrate, I feel like those are the types of things that 00:31:37.075 --> 00:31:40.915 will be, like they're already possible, but we'll start seeing more of that 00:31:40.915 --> 00:31:45.255 type of workflow within construction over the coming months. Those are low-hungry. 00:31:45.375 --> 00:31:50.124 For task automation, workflow automation. I think we're still a little bit maybe, 00:31:50.727 --> 00:31:54.565 like predictive analysis is still a little bit further out. But I mean, 00:31:54.565 --> 00:31:56.317 these kind of assistants, I think we're... 00:31:57.136 --> 00:32:00.130 You know getting a bit proposal agents i mean you're going to see a bunch of 00:32:00.130 --> 00:32:04.960 agents getting built into the different work that's where i think like aaron 00:32:04.960 --> 00:32:10.440 from a vc standpoint i think like cash why are you laughing i'm just having 00:32:10.440 --> 00:32:12.030 nightmares of role level permissioning. 00:32:13.390 --> 00:32:16.710 Right we're role level permissioning agents and i'm like well remember that's 00:32:16.710 --> 00:32:21.380 where i think the money needs to go which is like some kind of like no no like 00:32:21.380 --> 00:32:24.898 like an orchestration layer like Someone needs to build the orchestration layer 00:32:25.212 --> 00:32:29.032 that can be customizable and modular to 00:32:29.432 --> 00:32:32.190 programs, to domains, to all this stuff. 00:32:32.190 --> 00:32:35.360 To be fair, though, I think a lot of the big guys are already starting to build 00:32:35.360 --> 00:32:37.850 those orchestration layers. So I don't think this needs to be some third part. 00:32:37.990 --> 00:32:40.190 You're saying I don't have a billion-dollar idea, Tyler? 00:32:40.590 --> 00:32:43.797 No, I'm saying somebody's already working on it, Steve. 00:32:44.697 --> 00:32:46.717 Another idea killed by Tyler. I'll shut up now. 00:32:48.990 --> 00:32:53.880 Are you seeing are you seeing ideas around that come your way and get pitched to you aaron, 00:32:55.050 --> 00:32:58.551 for what how to run role level permission across multiple agents, 00:32:59.373 --> 00:33:04.370 maybe you make it sound so basic but that does sound like a cool idea no yeah 00:33:04.370 --> 00:33:08.089 i mean we see it all like yeah totally but you know, 00:33:08.681 --> 00:33:12.150 ultimately you also have to know like what is this guy supposed to do within 00:33:12.150 --> 00:33:16.510 my company so like it's not it's not like a easy button at all ever yeah, 00:33:17.110 --> 00:33:19.681 anywhere so that's fair. 00:33:20.809 --> 00:33:23.996 All right two questions on this then we got to go to the next one um let's do it, 00:33:24.685 --> 00:33:34.636 hanu and aaron what is like is there any examples of the use of agents within your companies, 00:33:35.292 --> 00:33:42.867 that has been interesting that you guys are either a part of or

just maybe seeing hanu go. 00:33:43.587 --> 00:33:48.178 Yeah no i think one of their i mean obviously we we're starting to get better 00:33:48.178 --> 00:33:52.608 at organizing our information and We've got an incredible amount of like historical 00:33:52.608 --> 00:33:56.247 information around schedules, around cost, 00:33:56.765 --> 00:33:58.648 subcontractors that we work with. 00:33:58.648 --> 00:34:01.238 I think probably one of the coolest things we're doing internally, 00:34:01.238 --> 00:34:06.108 aside from like predicting like our resources and allocation of resources on jobs. 00:34:06.108 --> 00:34:09.815 But the other one is probably just being able to look at the different, 00:34:10.511 --> 00:34:16.598 subs and propose subcontractors for our projects based on their past performance, 00:34:16.598 --> 00:34:20.110 their availability, how many projects they're on across our different, 00:34:20.755 --> 00:34:22.048 you know, national network. 00:34:22.048 --> 00:34:25.918 So I think those are like really like practical, simple examples. 00:34:25.918 --> 00:34:30.448 But, you know, I think those are leveraging our own data essentially for to 00:34:30.448 --> 00:34:35.157 be able to, you know, either hopefully create more predictable results, 00:34:35.633 --> 00:34:38.538 understand what you're signing up for, understanding where you have risk. 00:34:38.538 --> 00:34:40.672 I think those are some good use cases. 00:34:41.391 --> 00:34:47.320 I like that. Aaron? Do you want it within my venture practice or do you want it within a startup? 00:34:47.667 --> 00:34:50.332 No, on the venture side, let's see what you guys are doing. 00:34:50.978 --> 00:34:56.118 Cool, right? So the Java venture is very much like sales and decision-making, 00:34:56.118 --> 00:34:58.808 I would say, is the two anchors. 00:34:59.449 --> 00:35:06.458 And agents are killer at sales stuff. So of our new opportunity pipeline... 00:35:07.486 --> 00:35:13.601 Half of it comes from personal relationships and referrals, meaning like founders, friends, et cetera. 00:35:14.333 --> 00:35:19.772 10% of it comes from outbound and the remaining 40% comes from other investors, 00:35:19.772 --> 00:35:21.067 particularly lead investors. 00:35:21.669 --> 00:35:25.442 And so we thought really hard about how do you maintain your relationships with 00:35:25.442 --> 00:35:26.632 your co-investor network? 00:35:27.277 --> 00:35:31.202 And we use Affinity as our CRM. 00:35:31.591 --> 00:35:39.112 So how do you help build a simple cloud project in order to link to affinity 00:35:39.432 --> 00:35:40.772 and help me create an email? 00:35:40.772 --> 00:35:45.882 So whenever I go talk with one of our co-investors, I know what they like to 00:35:45.882 --> 00:35:50.272 invest in and I can go through my affinity and say, here are the things I think align best. 00:35:50.690 --> 00:35:55.514 And it like automates a lot of that relationship management perspectives. 00:35:55.665 --> 00:36:00.651

And that sounds a lot like an SDR function at a lot of startups or a lot of people just generally. 00:36:01.150 --> 00:36:06.915 And so it's been really useful to make sure we have depth in all of our connections 00:36:07.246 --> 00:36:10.659 that is current against our kind of currency of deal flow. 00:36:11.403 --> 00:36:15.090 Yeah, that's cool. We're good. Agents in the wild. 00:36:15.462 --> 00:36:20.742 Tyler, I share a couple of them. So we actually just got done working and deploying 00:36:20.742 --> 00:36:25.223 these out into one of my clients, level five cost estimation. 00:36:25.501 --> 00:36:30.272 So we can run a level five cost estimate through their company off of their 00:36:30.272 --> 00:36:33.948 historical benchmarking for all of the data center projects that they have worked on. 00:36:34.645 --> 00:36:36.467 Are the levels based on complexity? 00:36:37.412 --> 00:36:40.172 Yes, based off of complexity. So level five is super high level. 00:36:40.172 --> 00:36:42.366 So we're moving, building towards class four. 00:36:43.652 --> 00:36:47.469 So four, three, two, one, et cetera. One is super, super dialed. 00:36:48.331 --> 00:36:53.092 The other one is bid ingestion. So get bids from vendors. 00:36:53.392 --> 00:36:57.381 We can ingest that information and then we can do the leveling of said bids 00:36:57.770 --> 00:36:59.232 all within that process. 00:36:59.232 --> 00:37:03.982 So like, but that's all, it's all a workflow and we're injecting the agents 00:37:03.982 --> 00:37:08.992 in where it makes sense for them to do the work versus having a deterministic type of work. 00:37:10.125 --> 00:37:13.879 Architecture where it makes sense there too so i mean you sometimes it doesn't 00:37:13.879 --> 00:37:17.777 make sense to burn the tokens you know there's there's there could be too much variance, 00:37:18.626 --> 00:37:23.026 but yeah those are other examples i mean things we just token strategy is increasingly, 00:37:23.499 --> 00:37:27.374 a relevant conversation so and most strategic conversations go ahead i just 00:37:27.554 --> 00:37:31.649 i just came back from yada out in vegas so i came out from i didn't go to au 00:37:31.649 --> 00:37:35.339 but i did go to yada and i was standing in line talking to a guy at a, 00:37:36.199 --> 00:37:42.029 super large company he's cio over there i was like so like tell me about your 00:37:42.029 --> 00:37:43.986 usage of ai and he was like, 00:37:44.648 --> 00:37:48.269 he was kind of venting for a moment and then he told me about this bill that 00:37:48.269 --> 00:37:53.079 they got he was like yeah so one of our people burned eleven thousand dollars 00:37:53.079 --> 00:37:54.747 worth of tokens in a month, 00:37:55.341 --> 00:37:59.979 dude legend i'm like okay well what did they do and he's like they built a bunch 00:37:59.979 --> 00:38:02.144 of agents that didn't actually work. 00:38:03.530 --> 00:38:06.239 So I think that the return on investment 00:38:06.239 --> 00:38:10.089 conversation is going to start happening up the chain of command. 00:38:10.089 --> 00:38:15.039 It has to, right? So I mean, if we're burning $11,000 worth of tokens and I 00:38:15.039 --> 00:38:19.849 make $100,000, great. But if I'm burning $11,000 and I have nothing to show 00:38:19.849 --> 00:38:24.199 for it and everybody's doing it, we've got a really big problem on our hands. 00:38:24.559 --> 00:38:29.559 So I think there's, yes, token strategy and how we let certain people do certain 00:38:29.559 --> 00:38:31.540 things. We've got to have those conversations. 00:38:33.716 --> 00:38:38.052 All right. So let's move from the digital to the physical. 00:38:38.759 --> 00:38:43.169 We're building a few data centers in the U.S. these days. And there's one in 00:38:43.169 --> 00:38:49.129 particular in New Jersey that had some issue with using some natural gas-based 00:38:49.129 --> 00:38:51.115 generators that were unpermitted. 00:38:51.759 --> 00:38:57.399 They were fined a million dollars and allowed what looks like six weeks to get 00:38:57.399 --> 00:39:00.903 permits in place while they kind of continued operation. 00:39:01.466 --> 00:39:06.247 So, you know, I was on a panel this week at the Rice University kind of talking through. 00:39:07.513 --> 00:39:11.958 Power constraints as it relates to these, just the demand for data centers. 00:39:12.298 --> 00:39:14.748 And I think the average time right 00:39:14.748 --> 00:39:19.835 now from kind of idea to completion of Interconnect is like five years. 00:39:20.387 --> 00:39:25.628 And so increasingly, everyone's kind of thinking through their bridging strategy that relates to power. 00:39:26.146 --> 00:39:29.222 And generators are a big component of that. 00:39:30.041 --> 00:39:33.098 Do we feel like this is an. 00:39:33.948 --> 00:39:41.158 Instance and a trend towards maybe municipalities feeling empowered to kind 00:39:41.158 --> 00:39:45.634 of maybe insure and protect their citizenry? 00:39:45.761 --> 00:39:52.108 Or is a million dollars on a 15 plus billion dollar contract maybe more evidence 00:39:52.108 --> 00:39:54.212 that data centers still have the power? 00:39:54.862 --> 00:39:57.548 I guess I don't really know where there's a question there, but I think it's 00:39:57.548 --> 00:40:03.298 an interesting trend where, hey, the industry and the owner side is starting 00:40:03.298 --> 00:40:06.338 to feel the effects of the constraints of power. 00:40:06.855 --> 00:40:16.212 And are we going to see more of this kind of fining for what is, 00:40:16.514 --> 00:40:19.218 I mean, I'm sure this happens in more places than one. 00:40:19.218 --> 00:40:23.138 I think reading into the story, it sounds like they only really found out of 00:40:23.138 --> 00:40:29.401 this because local residents were complaining of the noise and filed some complaints. 00:40:29.819 --> 00:40:32.850 So, I don't know. There's a question there somewhere. 00:40:33.553 --> 00:40:37.088 Aaron, there's a lot of data

centers being built. Do you feel, 00:40:37.088 --> 00:40:43.178 do you see some pressures at a municipal level to kind of get more active in 00:40:43.178 --> 00:40:45.344 pushing you back against development? 00:40:45.907 --> 00:40:49.338 Or is there an understanding that, hey, we all need data centers like China's 00:40:49.338 --> 00:40:53.335 a variable. What's, what are you, what are you seeing in the corners of, of your world? 00:40:53.959 --> 00:40:55.803 Yeah, look, like. 00:40:56.976 --> 00:41:02.976 The international competitiveness question is great irisorily, 00:41:02.976 --> 00:41:04.742 but sucks when it's in your backyard. 00:41:05.079 --> 00:41:08.556 So I don't think local municipalities are like, oh, this is cool because we 00:41:08.556 --> 00:41:09.769 got to be competitive, right? 00:41:10.994 --> 00:41:15.486 But all developments require a public and private partnership, 00:41:15.486 --> 00:41:17.284 right? It doesn't matter what you're building. 00:41:18.002 --> 00:41:24.086 Again, the city has to see a fair exchange of value for the ultimate resource 00:41:24.086 --> 00:41:30.918 impact, whether that's air quality or resource use or noise or whatever, to their citizens. 00:41:31.720 --> 00:41:35.076 And right now, I think that those asks are going in one direction, 00:41:35.076 --> 00:41:38.696 which is higher, right? You know, data centers make a ton of money from the 00:41:38.696 --> 00:41:40.107 ultimate sale of compute. 00:41:40.695 --> 00:41:45.336 And so there's like, you know, a more comprehensive thoughtfulness from a lot 00:41:45.336 --> 00:41:48.796 of local municipalities to say, like, what's our what's our condos left that 00:41:48.796 --> 00:41:51.578 makes this an equal trade to be located here? 00:41:52.090 --> 00:41:55.136 I don't know how that's going to differentiate between geography. 00:41:55.136 --> 00:41:58.576 The United States is this collection of people that like have really different 00:41:58.576 --> 00:41:59.664 opinions about these things. 00:42:00.136 --> 00:42:04.726 And like those opinions may start to have like stronger competitiveness in rural 00:42:04.726 --> 00:42:08.666 areas and certain parts of the country and not in other parts of countries, 00:42:08.666 --> 00:42:10.793 even though they both might look the same on a map. 00:42:11.069 --> 00:42:15.336 And so my intuition is this is going to have like a really widespread of opinion 00:42:15.336 --> 00:42:20.856 and a really widespread of action as a voter and as a citizen, 00:42:20.856 --> 00:42:22.863 I'm more concerned about like illegally. 00:42:23.674 --> 00:42:28.918 Like driven wells, right? A groundwater use is not as big of a deal today as 00:42:28.918 --> 00:42:35.778 it was historically, but like, you know, I'm sure that the rules have been bent around groundwater. 00:42:35.778 --> 00:42:36.978 And like, that's the one that is 00:42:36.978 --> 00:42:40.388 like, you know, really, really limited resources, especially in the West. 00:42:41.188 --> 00:42:44.918 And so like, that's going to be, I think that's going to be a hotter button 00:42:44.918 --> 00:42:48.533 issue here as, as a policy group reviews it. 00:42:49.668 --> 00:42:53.057 Yeah. Yeah. I mean, obviously there's a lot of, 00:42:53.803 --> 00:42:57.538 A lot of controversy around these things, around data centers at the moment, 00:42:57.538 --> 00:43:00.458 but I think you can't solve, like in this particular example, 00:43:00.458 --> 00:43:04.698 you can't solve like a utility or permitting problem by just making a construction 00:43:04.698 --> 00:43:06.138 schedule more aggressive, right? 00:43:06.138 --> 00:43:09.268 So we have to plan for these things well ahead of time. 00:43:09.268 --> 00:43:12.998 Like the data centers, they're not just construction projects anymore. 00:43:12.998 --> 00:43:18.368 They're energy infrastructure projects that happen to have really large buildings around them, right? 00:43:18.368 --> 00:43:22.838 So conversations kind of used to be as can we build, how fast can we build a 00:43:22.838 --> 00:43:26.528 facility? It's increasingly becoming about can we get the power, 00:43:26.528 --> 00:43:30.753 equipment, supply chain challenges, labor, obviously permitting. 00:43:31.223 --> 00:43:34.828 And like, how do we align and plan for all of that in order for that to be, 00:43:35.409 --> 00:43:37.255 you know, a smooth process? 00:43:37.859 --> 00:43:41.660 And yes, to Aaron, your point, it's a partnership. 00:43:42.066 --> 00:43:46.478 It's a communication planning problem that hopefully we can all kind of figure 00:43:46.478 --> 00:43:49.368 out how to solve. But I mean, the demand isn't going anywhere, 00:43:49.368 --> 00:43:51.508 right? I mean, I think that's just the reality. 00:43:51.854 --> 00:43:57.288 I think we're kind of in this weird midterms cycle right now where there's just 00:43:57.288 --> 00:44:00.868 a lot of information and misinformation out there. 00:44:00.868 --> 00:44:04.188 But I think it's for our industry to solve some of these problems. 00:44:04.188 --> 00:44:07.888 I mean, we have to figure out how we build these things smarter and more sustainably. 00:44:08.648 --> 00:44:13.378 And figure out how we can at least get the facts out there around this stuff. 00:44:13.378 --> 00:44:19.478 Because, yes, they consume a lot of power, but right now, I think the projection 00:44:19.478 --> 00:44:24.408 is that even by 2028, it's going to be equivalent to anywhere like 6% to 12% 00:44:24.408 --> 00:44:27.554 of total U.S. electricity consumption. So it's not like. 00:44:28.445 --> 00:44:33.473 You know, the data centers are evil and bad. It's just making sure the communication 00:44:33.473 --> 00:44:36.853 is clear and, you know, we follow the right rules. And obviously, 00:44:36.853 --> 00:44:38.800 permitting process is part of that. 00:44:40.233 --> 00:44:44.413 First of all, I would like to say I'm shocked because everything is legal in 00:44:44.413 -->

00:44:46.693 New Jersey. According to Hamilton. 00:44:47.813 --> 00:44:50.213 Okay. That was a low blow. Low blow. 00:44:52.053 --> 00:44:55.179 Hey, hey, hey. This is a Hamilton reference. Just back off. 00:44:55.452 --> 00:45:00.883 All right. So, I mean, a million dollars is kind of like a little gnat flying 00:45:00.883 --> 00:45:05.228 around your head if you're a data center, like a $17 billion project. 00:45:05.512 --> 00:45:12.333 Like, who gives a crap at that point? And I think what's funny is that a lot 00:45:12.333 --> 00:45:16.173 of these companies would like you to believe that this is a partnership between 00:45:16.173 --> 00:45:17.703 the community and the data center. 00:45:17.703 --> 00:45:22.553 No, it's not. This is, we are going to bribe the ever-living crap out of you 00:45:22.553 --> 00:45:26.254 to get this thing built because we're going to make a crap load of money off of this. 00:45:26.760 --> 00:45:31.600 And that was one of the eye-opening things of conversations that I was even seeing in the last week. 00:45:31.860 --> 00:45:37.873 There are people that are dedicated to community engagement for data centers. 00:45:37.873 --> 00:45:43.533 They go in and they talk to the community and they try to share information 00:45:43.533 --> 00:45:45.577 about what's going to be happening at data centers. 00:45:45.810 --> 00:45:50.693 But also, if there is a protest, they will organize a counter protest against 00:45:50.693 --> 00:45:55.005 that. So it starts to get pretty freaking interesting when you start getting into this world. 00:45:55.417 --> 00:45:59.833 Like that was kind of mind blowing for me when I when I sit down at lunch and 00:45:59.833 --> 00:46:03.323 I'm like, what do you do? And that comes up like, wait a minute. 00:46:03.913 --> 00:46:09.013 Stop. Really? That's what you do for a living. So I like. 00:46:10.071 --> 00:46:13.301 I tend to think it's like, yeah, OK, congratulations. You got a million dollars 00:46:13.301 --> 00:46:17.571 out of it, New Jersey. But I mean, they're going to pay you whatever it takes 00:46:17.571 --> 00:46:19.385 to get you to go away so they can build this. 00:46:19.872 --> 00:46:26.117 And I mean, but again, I do agree that it's like, what's going to happen to the midterms? 00:46:26.552 --> 00:46:30.581 Because I think that at a certain point, we will start seeing some people get 00:46:30.581 --> 00:46:33.275 more and more emboldened to go out of their way to say, nope, 00:46:33.611 --> 00:46:35.231 it's going to be more than a million dollars. 00:46:35.678 --> 00:46:37.491 And that's kind of got me worried right 00:46:37.491 --> 00:46:40.911 now. Well, again, this is outside perspective. That's got me worried. 00:46:41.351 --> 00:46:46.766 What's going to happen if a lot of these projects start getting shut down because 00:46:47.131 --> 00:46:53.203 there are people that are newly elected into office want to start making a showing of, you know. 00:46:53.871 --> 00:46:59.161 I mean, I've made the point before, and Aaron, you know, interested in your perspective here. 00:46:59.161 --> 00:47:04.411 But I think from a macro standpoint, so much of our, you know, 00:47:04.411 --> 00:47:10.419 economic output and activity is based on kind of the large category of AI CapEx. And so. 00:47:12.631 --> 00:47:17.401 Between that and federal spending, that is the economy. And so I think when 00:47:17.401 --> 00:47:21.881 you talk about demand, Hanu, I think what's interesting is we are all using 00:47:21.881 --> 00:47:24.424 the tools, but we also, none of us, 00:47:25.034 --> 00:47:29.057 according to studies, 70% of us don't want data centers in our, in our counties. 00:47:29.475 --> 00:47:32.127 And so we all, we all want the tools. We don't want the tools, 00:47:32.533 --> 00:47:35.697 like the infrastructure behind the tools by us, which I think is fair. 00:47:36.016 --> 00:47:39.923 So I think it does come down to like community alignment from a incentive standpoint. 00:47:40.851 --> 00:47:45.331 You know, Andrew Yang, the former political or presidential candidate, 00:47:45.331 --> 00:47:54.095 went to Harvard and gave a speech and essentially armed the recent grads at Harvard with the... 00:47:54.851 --> 00:48:01.324 Like ethos that the battle of their generation is to defeat AI, 00:48:01.664 --> 00:48:06.124 using his words, and talked about how AI is not their friend. 00:48:06.124 --> 00:48:08.624 It's not going to make their life better. 00:48:09.089 --> 00:48:15.214 And it got a huge round of applause. And so there's a lot of interesting kind 00:48:15.214 --> 00:48:16.991 of inputs, obviously interesting time. 00:48:17.340 --> 00:48:21.628 But on the macro side of this, Aaron, if data centers go away, what happens? 00:48:22.804 --> 00:48:27.246 Well, first of all, the bet is too big for them to go away. We're like... 00:48:27.826 --> 00:48:32.685 You know what I mean? That ship is sailing. So... 00:48:34.070 --> 00:48:37.705 I, I, I'll bring this back to like a first principle of American culture, right? 00:48:38.078 --> 00:48:44.350 Like a big part of American culture is like, we're not super happy when someone 00:48:44.350 --> 00:48:46.854 does something without asking our opinion, right? 00:48:47.086 --> 00:48:51.330 We love to vote. We, you know, we became a country because we didn't love people 00:48:51.330 --> 00:48:53.551 just arbitrarily taxing the shit out of us. 00:48:54.132 --> 00:48:58.775 And so I think there's this piece that I tried to make a point earlier, 00:48:59.211 --> 00:49:05.760 which is each community is going to have an opinion very locally that matters 00:49:05.760 --> 00:49:07.380 to its citizens most locally. 00:49:07.380 --> 00:49:13.734 And this is the wonder of the American governmental system, is most power is your exact community. 00:49:14.152 --> 00:49:17.230 And the layers of government above that, right, county, state, 00:49:17.230 --> 00:49:22.349 federal, have decreasing amount of

influence on what is your day life on any given Tuesday. 00:49:22.964 --> 00:49:26.750 But your local community really impacts what your life is like on any given 00:49:26.750 --> 00:49:30.391 Tuesday. And so I think there's going to be a spread of opinion for that. 00:49:30.989 --> 00:49:37.210 And I think that the folks that advocate for it to happen in their community 00:49:37.210 --> 00:49:40.570 to have a data center, because it fits their. 00:49:41.403 --> 00:49:45.360 Approach to zoning, their approach to economic development, their approach to 00:49:45.360 --> 00:49:49.160 employment, are going to benefit from whatever the, you know, 00:49:49.160 --> 00:49:50.900 as Tyler so eloquently put it, 00:49:51.417 --> 00:49:54.944 incentives that they come with having it locally. 00:49:55.531 --> 00:50:00.558 And I think that mutual agreement between a local community and a developer, 00:50:01.098 --> 00:50:06.977 is the ethos to this but just blanketing our country with these without having that back and forth, 00:50:07.569 --> 00:50:10.535 is counterculture to a lot of what 00:50:10.965 --> 00:50:15.450 regardless if you're left or right kind of view of like the american approach 00:50:15.450 --> 00:50:19.860 to development which is like very much we have to agree to let you be in our 00:50:19.860 --> 00:50:24.646 backyard and so that's my intuition is you're going to see a very big spread on who says yes. 00:50:25.268 --> 00:50:28.650 Yeah, but it starts with people being able to actually get in the same room 00:50:28.650 --> 00:50:33.680 and talk about facts, right? And like really agree on what's real and what's not real. And, 00:50:34.475 --> 00:50:39.551 Hopefully, I come to some kind of an understanding that there are obviously 00:50:39.551 --> 00:50:43.901 benefits to the community, but then there are the right ways and wrong ways 00:50:43.901 --> 00:50:45.421 of building these things too, right? 00:50:45.421 --> 00:50:50.921 So we are lucky to work with a lot of clients that are taking a much more long-term 00:50:50.921 --> 00:50:53.965 view and sustainable approach to these buildings. 00:50:54.377 --> 00:50:57.929 And they are working with the communities and doing a lot of good things for the community. 00:50:58.469 --> 00:51:02.451 And genuinely doing that because they actually do care. 00:51:02.451 --> 00:51:07.281 Obviously they have incentives to build these things but you know aaron i think 00:51:07.281 --> 00:51:08.661 to your point earlier i mean these. 00:51:09.781 --> 00:51:12.852 Demand isn't going away ai is not going away and, 00:51:13.457 --> 00:51:17.941 you know i think the reality is is that this is the new new revolution of our 00:51:17.941 --> 00:51:23.441 age our time and yeah the the way we do work now from where it's going to be 00:51:23.441 --> 00:51:25.856 five ten years from now it's going to be quite different. 00:51:28.371 --> 00:51:33.841 I think that the interesting thing, if we were to zoom out, is I think

that, 00:51:34.541 --> 00:51:39.661 if we were to bifurcate and categorize AI as like the narrow use case, 00:51:40.461 --> 00:51:44.939 the narrow model trained versus like AGI in the pursuit of AI. 00:51:45.948 --> 00:51:52.179 Like the most utility, most of the utility that you and I and a majority of, 00:51:52.179 --> 00:52:00.491 I would say, the population can be had from open source, non-frontier models, AI. 00:52:00.990 --> 00:52:05.409 And yet, and so I think there is a question that I haven't heard a good answer 00:52:05.409 --> 00:52:14.478 on of other than like, we can't lose to China, of why the pursuit of large AGI? 00:52:15.099 --> 00:52:20.689 Which in theory, if they're telling us that it's going to take most of the jobs, 00:52:20.689 --> 00:52:22.402 doesn't serve the citizenry. 00:52:23.244 --> 00:52:32.594 And yet narrowly trained models do, what is the benefit of the pursuit of AGI to the population? 00:52:33.035 --> 00:52:36.299 You know, taking that question like head on, which is like the dumbest shit 00:52:36.299 --> 00:52:39.791 I've done in a while, is, you know, I... 00:52:40.572 --> 00:52:45.216 I think there's three reasons. One is frontier models, you know, 00:52:45.216 --> 00:52:49.896 have evidenced the best results from deep academic research, right? 00:52:49.896 --> 00:52:57.516 So, like, how do you, like, create a new GMO-modified corn or wheat or whatever 00:52:57.516 --> 00:53:03.355 that could potentially be planted in arid environments and solve starvation? 00:53:03.848 --> 00:53:09.563 You know, like, that is found on a frontier model, not on a desktop model, you know? 00:53:10.068 --> 00:53:15.226 And so you get like a new corn but potentially maybe like all the banks are 00:53:15.226 --> 00:53:17.236 hacked right like there's also risk on, 00:53:17.716 --> 00:53:22.414 the oh oh my god like totally like we're building in my fair city like you know 00:53:22.722 --> 00:53:27.496 the biggest installation of quantum computing right and like yeah this is like 00:53:27.496 --> 00:53:32.262 a ball of string that i do not want to unravel and you know it's it's, 00:53:32.916 --> 00:53:35.768 i'm trying to argue the upside here versus It's just the downside. 00:53:36.796 --> 00:53:40.786 But on the flip side of that, though, Aaron, it's like if we took all of the 00:53:40.786 --> 00:53:43.856 money we're investing in data centers and actually went out and bought food 00:53:43.856 --> 00:53:46.796 for everybody, then, oh, my God, we just solved world hunger. 00:53:47.616 --> 00:53:50.643 So it's like we're building AI so we can solve world hunger. 00:53:51.269 --> 00:53:55.376 What the frick, man? We just go solve world hunger with the money that we're spending on this crap. 00:53:55.776 --> 00:54:00.606 So it's a little bit of a balance. So could Elon personal, you know? Right. Exactly. 00:54:01.296 --> 00:54:06.226 Exactly. So this isn't this isn't for for any sort of generosity sake here. 00:54:06.226 --> 00:54:07.320 Let's not kid ourselves. 00:54:07.575 --> 00:54:14.346 This is for money, money. That's it. This is the key, the key to a good conversation 00:54:14.346 --> 00:54:16.876 is to open the can of worms right before the end of it, you know. 00:54:18.076 --> 00:54:19.469 And have to have to cap it. 00:54:20.576 --> 00:54:23.956 I think I would love to have a conversation just about this because I don't, 00:54:23.956 --> 00:54:25.681 I don't think we're like, we could stop. 00:54:25.814 --> 00:54:30.458 Like we could not build the thing we think is going to take all the jobs from our kids. 00:54:30.592 --> 00:54:35.036 And, and yet, like, there is a nuanced way of having that conversation yet. 00:54:35.036 --> 00:54:40.006 And I don't think it's, you know, I'm not saying we need to stop building everything. 00:54:40.006 --> 00:54:41.679 But there is a there's a nuance there. 00:54:42.004 --> 00:54:44.116 The one thing I want to get a take on before we go is, 00:54:45.037 --> 00:54:48.119 Caterpillar, one of the largest kind of equipment manufacturers, 00:54:48.578 --> 00:54:53.979 is partnering with Field AI, which is essentially building kind of the brain behind robotics. 00:54:55.061 --> 00:54:58.791 And if you look at some of the other raises that have happened in the last few 00:54:58.791 --> 00:55:04.101 months, there's definitely money coming into, and it seems like the frontier 00:55:04.441 --> 00:55:08.169 of equipment getting smarter and autonomous is maybe here. 00:55:09.019 --> 00:55:15.301 Hanu, what are you seeing on the GC side? Are you seeing heavy civil contractors 00:55:15.301 --> 00:55:18.504 come with an autonomous plan for projects yet? 00:55:19.260 --> 00:55:23.962 We haven't seen that yet. I mean, the physical AI is pretty interesting, but... 00:55:25.205 --> 00:55:29.085 When we're looking about it in the context of robots and gadgets. 00:55:31.758 --> 00:55:37.655 It's starting to supplement some workflows, but I think we're still pretty far 00:55:37.655 --> 00:55:42.845 from fully autonomous workflows. It's a workflow standardization problem. 00:55:43.905 --> 00:55:48.155 We've had that for ages. How do you make a construction site like a manufacturing 00:55:48.155 --> 00:55:50.065 process? Yeah, these sites are semi-variable. 00:55:50.785 --> 00:55:54.485 Yeah, and there's a lot of... Well, they're not the same. 00:55:54.485 --> 00:56:00.049 And all the same and it's not a there's a lot of volatility in all sorts of things so, 00:56:00.676 --> 00:56:05.565 you know i think construction autonomy i guess just isn't it won't be just an 00:56:05.565 --> 00:56:09.445 ai problem i think we need to be able to figure out how we can get more consistent 00:56:09.445 --> 00:56:11.516 with construction make it more predictable, 00:56:12.213 --> 00:56:16.285 maybe more modular i mean i think we're trending in that direction and that's 00:56:16.285 --> 00:56:20.205 where we can hopefully maybe introduce some of these solutions that, 00:56:21.273 --> 00:56:26.425 are become part of our production system yeah good stuff well we're out of time 00:56:26.765 --> 00:56:33.001 guys appreciate you hopping on this has been fun yeah it's been a good conversation alright guys, 00:56:33.770 --> 00:56:38.005 we're in the age of AI man computers we thought computers were going to take 00:56:38.005 --> 00:56:42.565 all of our jobs right but here we still are on the next episode we're going to talk about, 00:56:43.348 --> 00:56:45.002 256 encryption, 00:56:46.345 --> 00:56:49.737 yeah just get into quantum alright you guys.