สำเนาบทสนทนา 🔗
WEBVTT
00:00.612 --> 00:10.942
[SPEAKER_02]: On this week's show, construction is the proving grounds for complexity with Lubomir Burdev and Cameron Azirbal from Crime Point.
00:21.572 --> 00:24.035
[SPEAKER_02]: Alright, you guys might notice.
00:25.285 --> 00:29.546
[SPEAKER_02]: It's just me because I wanted to bring two people because this is going to be a fun story.
00:29.586 --> 00:30.786
[SPEAKER_02]: So Cameron, how you doing buddy?
00:31.107 --> 00:31.667
[SPEAKER_02]: I'm doing well.
00:31.707 --> 00:32.127
[SPEAKER_02]: How are you?
00:32.447 --> 00:41.109
[SPEAKER_02]: I'm wonderful that I finally tracked you down and got you onto a microphone so we can catch up officially excited because I've been around prime point for a while here.
00:41.129 --> 00:42.669
[SPEAKER_02]: So we're going to have some fun talking about that.
00:42.689 --> 00:45.090
[SPEAKER_02]: But Lubremeer, it's great to have you on the show as well.
00:45.390 --> 00:48.451
[SPEAKER_01]: Thank you for inviting us, looking forward to chatting.
00:48.742 --> 00:54.145
[SPEAKER_02]: Yeah, I mean, we'll get into it to the history here before we do you guys know the deal.
00:54.665 --> 00:59.568
[SPEAKER_02]: Check us out on youtube.com forward slash at the content crew.
01:00.268 --> 01:04.450
[SPEAKER_02]: Follow us on social follow the guests on social share it with your friends.
01:04.590 --> 01:07.572
[SPEAKER_02]: You never know if you tap us on the show or you might even be the next guest.
01:08.015 --> 01:15.758
[SPEAKER_02]: All right, so I'm going to bring this because this was cool because I get to kind of do these things once in a while and say like, hey, I think this story is really interesting.
01:16.358 --> 01:21.340
[SPEAKER_02]: Between the technology side and the construction side so I'm going to start with lube america's
01:22.235 --> 01:23.577
[SPEAKER_02]: that's the technology side.
01:23.637 --> 01:31.448
[SPEAKER_02]: So we kind of talked at the beginning of this, but what's what was sort of the impetus behind your history and how you got here today?
01:31.749 --> 01:32.189
[SPEAKER_01]: Okay.
01:32.430 --> 01:37.637
[SPEAKER_01]: So I grew up in Bulgaria about time, a communist Bulgaria.
01:37.798 --> 01:37.878
[SPEAKER_01]: And
01:38.438 --> 01:45.300
[SPEAKER_01]: soon after the fall of communism, I came to the United States as a student at Brown University.
01:45.340 --> 01:49.261
[SPEAKER_01]: I did my bachelor's and master's there in computer science.
01:49.361 --> 01:53.822
[SPEAKER_01]: Then I came to the Bay Area, started working at Adobe for some time.
01:54.542 --> 01:58.503
[SPEAKER_01]: Adobe Research on Photoshop Illustrator, Acrobat.
01:58.963 --> 02:03.809
[SPEAKER_01]: And that's where I developed the first phase detector.
02:03.849 --> 02:10.757
[SPEAKER_01]: That was my first kind of introduction to computer vision, which we shipped to Photoshop elements.
02:10.817 --> 02:17.926
[SPEAKER_01]: And that was one of the very first, if not the first commercial application of phase detection back in 2005, I believe.
02:19.227 --> 02:34.518
[SPEAKER_01]: And I found computer vision really exciting and so I wanted to pursue a PhD, so I ended up going to Berkeley for my PhD while I was part-time at Adobe and then after that I joined Facebook.
02:35.218 --> 02:46.444
[SPEAKER_01]: And I was excited about Facebook because it was the largest place where you can like scale was really important for computer vision and Facebook had the largest scale.
02:46.744 --> 02:59.690
[SPEAKER_01]: And so I was the first person that Facebook hired to do computer vision and I found that their first computer vision team, we built this technology that understands every photo on every second of every video.
02:59.890 --> 03:06.115
[SPEAKER_01]: And after that, I did this start-up inspired by the HP or show Silicon Valley.
03:06.355 --> 03:09.278
[SPEAKER_01]: It was on video compression with neural nets.
03:09.858 --> 03:11.600
[SPEAKER_01]: We got really great results.
03:12.020 --> 03:13.621
[SPEAKER_01]: I saw the company to Apple.
03:14.182 --> 03:19.726
[SPEAKER_01]: And then after that, I wasn't excited about a large company.
03:19.746 --> 03:21.308
[SPEAKER_01]: I was thinking what to do next.
03:21.628 --> 03:24.050
[SPEAKER_01]: And that's where LLams were really taking off.
03:24.290 --> 03:41.719
[SPEAKER_01]: I ended up focusing on construction because the construction industries, one of the largest industries and the kinds of challenges in construction were just within the range of what AI could actually help solve for the first time.
03:42.467 --> 04:00.097
[SPEAKER_02]: Yeah, I mean, it's a very interesting history of being at the kind of beginning of different technologies and seeing their usefulness and then bringing them all the way of fruition through scale, which I think is really unique because that that leaned into why you picked the construction industry and it's where the beginning sort of came from is.
04:00.357 --> 04:15.003
[SPEAKER_02]: There's a complexity in collaboration and AI has an ability to help in that area if it's done correctly and well as you said it in the beginning, where's the most complex collaboration there is it's in a construction project no matter how little or how big.
04:15.503 --> 04:21.286
[SPEAKER_02]: So you came up with those ideas and co-founder and started working on it and a cool history.
04:21.386 --> 04:34.652
[SPEAKER_02]: If you guys remember the show here, I actually got to meet Luba Mayor before they came out of stealth and then they came out of stealth on a show that I was co-hosting for Nate Fuller, which was super cool because I, you know, I met you guys then, but I also met Cameron then.
04:34.692 --> 04:41.655
[SPEAKER_02]: So Cameron, talk about how meeting these guys still stealth and why did it appeal to you?
04:42.635 --> 04:50.744
[SPEAKER_02]: We get a lot of people, no, and Loubury Rears not one of these, but we get a lot of
Load full transcript
people that know better than the construction industry coming on in from tech.
04:51.064 --> 04:52.566
[SPEAKER_02]: But that's not this world.
04:52.886 --> 04:56.010
[SPEAKER_01]: How did that appeal to you and what did it appeal to you about what they were working on?
04:56.110 --> 04:56.410
[SPEAKER_00]: Yeah.
04:56.570 --> 04:59.353
[SPEAKER_02]: Well, maybe give a little bit of your background, too, because I just jumped through that.
04:59.393 --> 05:00.795
[SPEAKER_00]: So give a little bit of your background, too.
05:01.035 --> 05:01.635
[SPEAKER_00]: Sure, sure.
05:01.675 --> 05:03.776
[SPEAKER_00]: Yeah, very different path from Lubimer.
05:03.796 --> 05:15.382
[SPEAKER_00]: I think the one thing we have in common is we both went to Berkeley, but besides that, I did civil engineering at Berkeley and I went straight to the GC world after that.
05:15.562 --> 05:28.268
[SPEAKER_00]: I worked for a Turner for a couple of years after college and then I spent most of my career at web core working on large commercial projects in the Bay Area.
05:28.548 --> 05:38.652
[SPEAKER_00]: So, I've come from the general contracting background, when I met Lubimir and Hammid, I was in the middle of that complex collaboration process myself.
05:39.732 --> 05:46.935
[SPEAKER_00]: We met, we chatted a couple of times, I threw some ideas at the team and I was very impressed and amazed.
05:47.095 --> 05:53.037
[SPEAKER_00]: We chatted about an idea and we met like a week later and they already had built something to address it.
05:53.197 --> 05:53.817
[SPEAKER_00]: within a week.
05:54.057 --> 06:06.521
[SPEAKER_00]: And the speed of which they were moving and how they were thinking about tackling the problems that we are seeing, even without having a very extensive background in construction, it was really impressive to me.
06:06.801 --> 06:11.943
[SPEAKER_00]: And it really spoke to a lot of the processes and pain points that I wish I had.
06:11.983 --> 06:18.625
[SPEAKER_00]: And I really thought that this is something huge that we could really impact the industry with.
06:18.945 --> 06:19.845
[SPEAKER_00]: And that was it.
06:20.133 --> 06:20.873
[SPEAKER_02]: That was it, huh?
06:20.894 --> 06:21.834
[SPEAKER_02]: That was you were sold.
06:21.894 --> 06:29.579
[SPEAKER_02]: So Luma Mere, I want you to talk a little bit about actually the problem that you've focused this, you know, prime point on.
06:29.739 --> 06:35.623
[SPEAKER_02]: So people get an understanding, because they're listening in and they're going down the road and they're like, okay, so it's another problem of collaboration.
06:36.143 --> 06:42.647
[SPEAKER_02]: Explain what it meant to you and what you were doing underneath, and then I'll have Cameron kind of show why that was helping him at Web Corps.
06:42.967 --> 06:57.845
[SPEAKER_01]: Well, so collaboration was the inspiration of it, but one we ended up working on was really where we realized that getting into an truly understanding
construction data unlocks a lot of possibilities.
06:58.185 --> 07:13.145
[SPEAKER_01]: And so our platform takes all the project data and does the parsing and builds a knowledge knowledge graph out of this, the features that we provide really following to two categories reducing risk and increasing efficiency.
07:13.805 --> 07:18.127
[SPEAKER_01]: and for different customer's efficiency means different things.
07:18.167 --> 07:20.547
[SPEAKER_01]: For data centers, time is of the essence.
07:20.827 --> 07:23.268
[SPEAKER_01]: For others, cost reducing costs and so on.
07:23.428 --> 07:32.591
[SPEAKER_01]: But that's basically kind of the viral proposition, but at a big picture, I see that, like, construction data.
07:32.651 --> 07:40.674
[SPEAKER_01]: For example, there was a big transformation, led by one grid, where the data was from paper to electronic.
07:43.635 --> 07:48.437
[SPEAKER_01]: you can now pull the data on your device, you can search quickly, find information.
07:48.958 --> 08:00.023
[SPEAKER_01]: So, and I think we're on the verge of a second big transformation, which is that we take this data and we represent it in a semantically coherent way.
08:00.123 --> 08:09.448
[SPEAKER_01]: Because the data inside the PDF is still very difficult to work with, like, information about a room is distributed in many different places.
08:09.968 --> 08:11.429
[SPEAKER_01]: You want to talk about the door.
08:11.509 --> 08:14.170
[SPEAKER_01]: There is a graph that shows where the door is.
08:14.791 --> 08:18.193
[SPEAKER_01]: The property is about the door in some table, somewhere else.
08:18.233 --> 08:19.994
[SPEAKER_01]: The jump detail is somewhere else.
08:20.114 --> 08:23.375
[SPEAKER_01]: And the hardware set is in the packets all over the place.
08:24.036 --> 08:32.220
[SPEAKER_01]: And so we see an opportunity to use the eye to really reorganize the data in a way that is semantically makes sense.
08:32.380 --> 08:33.221
[SPEAKER_01]: And it makes sense
08:37.443 --> 08:52.555
[SPEAKER_02]: It's a unique, I like that you use that analogy because, you know, way back in the beginning of the day of the show, we had the team from playing great on and, you know, Tracy started with a real problem of paper plans and not even being able to physically slip sheet them or carry them and or have all that information just.
08:53.195 --> 08:55.176
[SPEAKER_02]: in a mobile format so that you could see it.
08:55.256 --> 08:58.938
[SPEAKER_02]: And so it really was a problem searching for help.
08:58.998 --> 09:03.401
[SPEAKER_02]: And, you know, luckily she had Ralph to help her, you know, do that.
09:03.601 --> 09:04.441
[SPEAKER_02]: This comes along.
09:04.541 --> 09:05.502
[SPEAKER_02]: What is that process?
09:05.522 --> 09:07.103
[SPEAKER_02]: So I love the knowledge graph.
09:07.123 --> 09:11.065
[SPEAKER_02]: I love what you're talking about and turning it into data and all those things being in different places.
09:11.545 --> 09:14.587
[SPEAKER_02]: But on a day-to-day to somebody listening Cameron, what did that mean?
09:14.707 --> 09:16.788
[SPEAKER_02]: Where does this come into play and what you were doing?
09:16.908 --> 09:27.831
[SPEAKER_00]: The theme that I saw a lot in my career is in the beginning of the project, once you win a job, everyone's so excited about building this thing.
09:27.851 --> 09:30.691
[SPEAKER_00]: You look at the rendering, you're like, we're going to collaborate.
09:30.791 --> 09:33.332
[SPEAKER_00]: The first kickoff with everybody is amazing.
09:33.552 --> 09:38.953
[SPEAKER_00]: Everyone's going to do their part, collaborate, architect, owner, all the subcontractors.
09:39.013 --> 09:39.633
[SPEAKER_00]: Everyone really
09:43.294 --> 09:47.697
[SPEAKER_00]: But after a while, kind of issues start bubbling up, right?
09:47.737 --> 10:01.885
[SPEAKER_00]: You find issues in the documents, you are delayed on commitments, and that relationship between everybody kind of gets eroded a little bit, and the process gets a little bit less exciting.
10:02.045 --> 10:03.646
[SPEAKER_00]: People are less collaborative.
10:03.826 --> 10:05.007
[SPEAKER_00]: They're managing risks.
10:05.047 --> 10:07.208
[SPEAKER_00]: They're covering your basis and
10:11.170 --> 10:20.116
[SPEAKER_00]: not spending enough time up front on digging into what you're committing to, finding errors that are already in the documents.
10:20.136 --> 10:26.021
[SPEAKER_00]: You just didn't get to look at them or just like you didn't appreciate some of the complications early enough.
10:26.441 --> 10:33.126
[SPEAKER_00]: What that means day to day for us, what the tools we have in the features that we have on our platform is.
10:33.386 --> 10:36.208
[SPEAKER_00]: Let's say you're trying to do a constructive ability review.
10:39.410 --> 10:45.353
[SPEAKER_00]: Someone with a lot of construction experience digging in and saying, hey, is this design constructable?
10:45.654 --> 10:50.496
[SPEAKER_00]: Can we use different methodology material or, you know, does it meet the budget?
10:50.597 --> 10:56.100
[SPEAKER_00]: And what we ended up doing a lot of times is just going through the documents and see if the design is coordinated.
10:56.380 --> 11:01.063
[SPEAKER_00]: Or if there is errors in the documents that could end up being delays or change orders.
11:01.323 --> 11:07.146
[SPEAKER_00]: So what if we used AI and computer vision to tackle some of that low-hanging
11:08.847 --> 11:15.551
[SPEAKER_00]: So you can spend some time focusing on the deeper critical thinking aspect of the projects.
11:15.631 --> 11:26.498
[SPEAKER_00]: And you can apply that to getting revisions, understanding what's in the revision, reviewing a product that a submittal assisting with RFIs, disfinding information.
11:26.518 --> 11:32.622
[SPEAKER_00]: There's so many different processes that I feel like we are finally at a point with AI that we could help either.
11:32.982 --> 11:39.079
[SPEAKER_00]:
Automate a mundane piece of it or facilitate making the project teams faster and solving them.
11:39.405 --> 12:04.104
[SPEAKER_02]: Yeah, I mean, you hit on a couple of really critical pieces there that I think people overlook right and especially in the process because, you know, this is why you guys started it is a complicated owners need to have something an asset of some sort those assets are very different in what they want a data centers very different than a hospitals very different than a power plant a power station than a water treatment plant.
12:04.464 --> 12:11.667
[SPEAKER_02]: Then a road, then a brick, I mean, like these are all construction projects, so they all want very in need, very different things.
12:12.228 --> 12:17.430
[SPEAKER_02]: But they follow some standard processes that can be sped up.
12:17.490 --> 12:26.134
[SPEAKER_02]: I like how you set it that way, because you get in reds, and especially as, you know, in this new modern world, we're like trying to move really, really fast, right?
12:26.154 --> 12:30.776
[SPEAKER_02]: So the owners giving you that basic idea of what they need, the architects giving you the basic,
12:31.216 --> 12:44.965
[SPEAKER_02]: The engineers getting there, but I rev over rev, things are changing, and like you said, we're we're digging over thousands of things to find the needles in the haystack, and that's actually not getting us any closer to that outcome.
12:45.266 --> 12:46.526
[SPEAKER_02]: And that's what we want to do, right?
12:46.547 --> 12:48.348
[SPEAKER_02]: That's the constructability review.
12:48.868 --> 12:51.870
[SPEAKER_02]: That's the ability to provide new means and methods, right?
12:51.890 --> 12:56.633
[SPEAKER_02]: That's the other problem in construction is we have so much time checking boxes.
12:56.934 --> 12:57.234
[SPEAKER_02]: We're not
12:58.123 --> 13:02.126
[SPEAKER_02]: thinking differently and applying all these new things that are coming out, right?
13:02.146 --> 13:12.393
[SPEAKER_02]: We have, you know, computer vision and events, but there have led to so many new things that we can do from a manufacturing perspective and off-site perspective and automated perspective.
13:12.453 --> 13:20.039
[SPEAKER_02]: So it's really cool to see you guys take this and give us time to what I hope is understand the project.
13:20.419 --> 13:24.002
[SPEAKER_00]: Something I would add to your point a lot of these processes.
13:24.967 --> 13:28.589
[SPEAKER_00]: repeat over and over from project to project.
13:28.809 --> 13:42.317
[SPEAKER_00]: What I felt like we tend to start over again, like the same problems with this new team members now, and you can almost predict the movie on what's gonna come up, what's the crescendo and how it's gonna end.
13:42.837 --> 13:44.578
[SPEAKER_00]: And part of it for me was,
13:45.038 --> 14:12.377
[SPEAKER_00]: I can't tell you how many times in the project there was a scope of work that was coming up in the field and I knew that there was like five things I should check from my past experience to make sure all the puzzle pieces go together but that meant like sitting down in quiet space spending couple hours opening up couple of shop drawings open up the drawing open the spec open this put them all together really focus and dive in to see oh
14:13.498 --> 14:19.320
[SPEAKER_00]: Actually, these two things are not coordinated, or we're going to have a problem here with something that is coming up later.
14:19.840 --> 14:29.283
[SPEAKER_00]: However, I had 20 things calling my attention every day in the form of an email, in the form of a deliverable, that where
14:29.963 --> 14:32.644
[SPEAKER_00]: quick, fast, easy task, but they had to get done.
14:33.005 --> 14:37.667
[SPEAKER_00]: And I couldn't, like, you wouldn't get to that deep critical thinking early enough.
14:38.027 --> 14:43.510
[SPEAKER_00]: What we can do now is maybe take care of that low hanging fruit or the deep critical thinking.
14:43.550 --> 14:48.692
[SPEAKER_00]: Maybe we can facilitate getting the fast set because everything's now connected in this knowledge graph.
14:49.192 --> 14:50.313
[SPEAKER_00]: And you can just dive right in.
14:50.571 --> 14:53.633
[SPEAKER_02]: I want to bring back now to Liebermeer, like you did.
14:53.653 --> 14:57.956
[SPEAKER_02]: You mentioned the knowledge graph and how you're talking about different revisions.
14:57.976 --> 14:58.937
[SPEAKER_02]: You're talking about some Middle East.
14:58.957 --> 15:00.238
[SPEAKER_02]: You're talking about RFIs.
15:00.558 --> 15:04.160
[SPEAKER_02]: How is this knowledge graph and what you're capable of doing?
15:04.801 --> 15:06.722
[SPEAKER_02]: How's it facilitate all that happening?
15:06.762 --> 15:11.125
[SPEAKER_02]: Like there's a lot of people think you just chat with your drawing and that's not where you're talking about.
15:11.545 --> 15:12.906
[SPEAKER_01]: How is facilitates?
15:13.326 --> 15:16.429
[SPEAKER_01]: So first of all, for my Joseph or Spectif,
15:17.269 --> 15:25.676
[SPEAKER_01]: Imagine, I guess a good example is you want to click on a one detail and you want only information about this detail.
15:25.777 --> 15:31.722
[SPEAKER_01]: Where is it being used to reference that this back referencing doesn't exist in as far as I know in other products?
15:32.242 --> 15:34.244
[SPEAKER_01]: Who is responsible for doing this?
15:34.684 --> 15:36.506
[SPEAKER_01]: When is it going to be on the schedule?
15:36.826 --> 15:39.348
[SPEAKER_01]: And how does it look from a previous division?
15:39.529 --> 15:42.752
[SPEAKER_01]: Are there any other files and shop drawings associated with this?
15:43.292 --> 15:46.995
[SPEAKER_01]: All the information, like, just associated with this?
15:47.456 --> 15:49.638
[SPEAKER_01]: So this is what the Knowledge Graph provides.
15:50.178 --> 16:06.332
[SPEAKER_01]: and it provides to the user in our intelligent navigation platform, connected documents platform, but it also
provides to our agents that are navigating the Nordic Graph that are looking for answers to questions or automating tasks.
16:06.734 --> 16:18.585
[SPEAKER_02]: Okay, so explain that a little deeper for me on Give me an example of where I might be because we hear about this, but how your platform might be grabbing more for me and surfacing it, so I can see it.
16:18.725 --> 16:20.968
[SPEAKER_02]: What would an agent be doing in that?
16:20.988 --> 16:25.432
[SPEAKER_02]: Because I kind of get how the two go together, but I'm wondering where the rubber meets the road.
16:25.832 --> 16:51.165
[SPEAKER_01]: Okay, so concrete example there is like if you want to ask a question and we have this sample project and one question is like on which floor is camera 35 for example and we have a schedule for cameras that has camera 35 and that drawing references of floor so it's like you have to do multiple hopes in order to answer this question.
16:51.365 --> 16:56.389
[SPEAKER_01]: And it's like for many questions, the answer is not immediately obvious in one place.
16:56.449 --> 16:59.731
[SPEAKER_01]: You just have to follow these connections to get the answer.
17:00.492 --> 17:12.000
[SPEAKER_01]: And at a glance, it's not like they will never be able to do this without our knowledge craft, but they're getting more and more powerful, but it is quite a big of an indirect.
17:12.461 --> 17:16.764
[SPEAKER_01]: You're basically hoping that they're going to recognize this figure out what it is.
17:21.168 --> 17:23.902
[SPEAKER_01]: We're making it much easier using this Norwich Graph.
17:24.238 --> 17:51.488
[SPEAKER_02]: Okay, yeah, that makes a lot more sense because that sure the algorithms are getting better and we're getting better at finding the needle in the haystack faster than, you know, Cameron could used to be able to find it in a quiet room with all the details and everything up, but your knowledge craft is actually guiding it so it's acting as sort of a table of contents for, hey, this is where you're going to go and where you're going to grab things and bring things together to speed it up a create more accuracy and less probably false positives along the way or false negatives along the way.
17:51.548 --> 17:51.728
[SPEAKER_02]: So,
17:52.328 --> 17:59.762
[SPEAKER_02]: When you hear that Cameron and you're thinking about submittals or RFIs, where is this the most impactful for somebody using it?
17:59.822 --> 18:02.927
[SPEAKER_02]: Like that where you see in that this makes a ton of sense.
18:03.406 --> 18:26.004
[SPEAKER_00]: There's so many examples in operation, so part of it is navigating the knowledge graph, navigating a drawing like a person with or part of it is like just knowing everything that's in there a lot of times the issue that I had you know generally what should be there, but there was always things that surprised you whether it was a note that was hidden somewhere that you didn't know about.
18:26.304 --> 18:34.326
[SPEAKER_00]: or there was an RFI written about a topic that you didn't know about because a team is so big and there's so many documents going past and forth.
18:34.486 --> 18:36.587
[SPEAKER_00]: Or you didn't know a status of a submittal.
18:36.647 --> 18:45.790
[SPEAKER_00]: So imagine today, if you're a superintendent and you are preparing for a slap for of the third deck.
18:46.490 --> 18:47.711
[SPEAKER_00]: and that work is coming up.
18:47.891 --> 19:02.605
[SPEAKER_00]: You would probably have to go bug a couple of engineers ask them what the status of the rebars submittal is and are we all ready to go with are the RFIs closed and you know generally what the drawings are telling you and what the spec is telling you.
19:02.665 --> 19:02.805
[SPEAKER_00]: So
19:03.406 --> 19:06.068
[SPEAKER_00]: There is one piece of, okay, where is everything?
19:06.168 --> 19:08.209
[SPEAKER_00]: Who do I need to talk to to understand?
19:08.729 --> 19:18.255
[SPEAKER_00]: Now imagine if you went on your schedule and you click on the third level slap or activity and all of this information was presented to you from the Knowledge Graph.
19:18.595 --> 19:26.400
[SPEAKER_00]: So you know where the status of the submittals are, you know what RFIs are related, you know all the notes, all the details are the specs that are related to that thing.
19:26.420 --> 19:37.907
[SPEAKER_00]: So you don't have to deal with surprises by an inspector or ride before you're getting ready for the poor by your quality control manager that you didn't do a piece of the work necessary to do it.
19:37.927 --> 19:40.949
[SPEAKER_00]: Now you have to delay it or rework if you end up to
19:41.409 --> 19:42.271
[SPEAKER_00]: making a mistake.
19:42.732 --> 19:51.225
[SPEAKER_00]: That is really for me the power that it unlocks in making someone do their job so much faster, so much better with better quality as well.
19:51.409 --> 19:52.650
[SPEAKER_02]: Yeah, I mean, you nailed it.
19:52.870 --> 19:56.751
[SPEAKER_02]: It's the ability to have all that information in one place and know what's missing.
19:56.771 --> 20:04.195
[SPEAKER_02]: Instead of all the time being spent finding, it's learning because that's one of the things I hear as a negative.
20:04.315 --> 20:09.997
[SPEAKER_02]: I've got a lot of videos out there and stuff and all here, you know, I don't want people doing too many repetitive tasks.
20:10.017 --> 20:14.399
[SPEAKER_02]: And I had somebody say, well, but repetitive task is how you learn, how you get intimate with it.
20:14.419 --> 20:16.380
[SPEAKER_02]: And I'm like, okay, in school, yeah.
20:17.000 --> 20:19.501
[SPEAKER_02]: But you're talking about a complex project
20:21.382 --> 20:25.905
[SPEAKER_02]: You know, floor three, you're going to do it on floor four, floor five, floor six.
20:26.125 --> 20:28.307
[SPEAKER_02]: You don't have to start over at the beginning every time.
20:28.707 --> 20:29.908
[SPEAKER_02]: It's about effectiveness.
20:29.948 -->
20:36.332
[SPEAKER_02]: So it's not really, you're not losing any of your familiarity with the project.
20:36.472 --> 20:37.573
[SPEAKER_02]: That's what they're worried about.
20:37.913 --> 20:43.257
[SPEAKER_02]: It sounds to me like maybe you're even getting more familiarity in an easier fashion with the project.
20:43.711 --> 20:43.991
[SPEAKER_01]: Yes.
20:44.392 --> 20:45.092
[SPEAKER_00]: Go ahead of me.
20:45.292 --> 20:47.133
[SPEAKER_01]: Louvre is like, yeah, I gotta say that with you ahead.
20:47.474 --> 20:58.161
[SPEAKER_01]: We are getting feedback from customers that they're getting up to speed with the project faster because the concern is AI will just make your like answer the questions then you don't learn anything.
20:58.181 --> 21:00.523
[SPEAKER_01]: You just have to go and ask AI and you can't.
21:00.623 --> 21:08.024
[SPEAKER_01]: So, but using our product, it actually makes you go much faster to, you don't spend time searching for information.
21:08.144 --> 21:15.566
[SPEAKER_01]: Like a great analogy that I like to use is imagine if you have the internet, but you never invented the notion of hyperlink.
21:15.726 --> 21:21.947
[SPEAKER_01]: And every time you want to find something, you have to go to this web page and search for this thing to go to the next item.
21:22.267 --> 21:24.987
[SPEAKER_01]: And we're introducing these hyperlinks right there.
21:25.567 --> 21:28.408
[SPEAKER_01]: And once you kind of experience that there's no going back.
21:28.688 --> 21:31.690
[SPEAKER_02]: It's huge, too, because I mean, we're seeing this across the board with AI.
21:31.710 --> 21:32.770
[SPEAKER_02]: Like, there's two ways to use it.
21:32.790 --> 21:35.972
[SPEAKER_02]: You can use it, B-dummer, can answer your questions for you, and you don't learn anything.
21:36.072 --> 21:37.613
[SPEAKER_02]: You're just sort of the button in the middle.
21:37.693 --> 21:40.995
[SPEAKER_02]: And yeah, you might get automated, but that's probably not a good idea.
21:41.135 --> 21:46.058
[SPEAKER_02]: I have a son, you saw him walk through the camera before, and so I'm interested in how teachers are approaching it.
21:46.078 --> 21:48.719
[SPEAKER_02]: And it's like, I had a at the AGC national.
21:48.759 --> 21:52.001
[SPEAKER_02]: We were on with a professor who said, listen, I love him use it.
21:52.321 --> 21:55.823
[SPEAKER_02]: I just spend now time, I don't, I talk to them about it.
21:56.083 --> 22:08.769
[SPEAKER_02]: If they went in and used an LLM to help them articulate what they were thinking and answer this question in a thoughtful way and learn it and present it and present their findings, I can tell in a second when I talk to them.
22:09.129 --> 22:18.254
[SPEAKER_02]: If they're knowledgeable, if they can speak to it, if they're really in green news like and when it's done right, they're farther in detail than they would have been.
22:18.754 --> 22:21.376
[SPEAKER_02]: had I not allowed them to use those mechanisms.
22:21.396 --> 22:32.186
[SPEAKER_02]: Like you said, Lumimere, to just collect all the hyperlinks, just collect the pieces of the information that you're really need to dig into and learn and then digest those and be able to to use them.
22:32.266 --> 22:33.146
[SPEAKER_02]: It's exactly
22:33.847 --> 22:36.448
[SPEAKER_02]: you're talking about in doing in the project.
22:36.768 --> 22:39.509
[SPEAKER_02]: And you're not talking about removing the human from the loop.
22:39.529 --> 22:45.050
[SPEAKER_02]: And I think Cameron, that's really important from your perspective for the industry is like, this isn't about replacing people.
22:45.350 --> 22:55.533
[SPEAKER_02]: It's really about augmenting something that I mean, can many times you think you took a risk without this kind of stuff on something and roll the dice on that slab and that poor.
22:56.082 --> 22:58.423
[SPEAKER_00]: Yeah, I was actually talking to someone about this.
22:58.823 --> 23:08.507
[SPEAKER_00]: There was a lot of times where, you know, because on large projects, you tend to handle a piece of the scope, and there's another team that handles another piece of the scope.
23:08.767 --> 23:23.733
[SPEAKER_00]: And, you know, the hope is you've done enough coordination in the design and the shop drawing process, where when you, let's say you build a curb, a concrete curb for the drywall to sit on or the current wall to sit on,
23:24.844 --> 23:29.646
[SPEAKER_00]: You kind of hope it's at the right elevation, and it's a right thickness and all of that.
23:30.227 --> 23:42.353
[SPEAKER_00]: But to an extent, there are so many times we did the work and just hope that we, Clayter, we didn't get a call back from the next subcontractor saying, sorry guys, this is not gonna work for me.
23:42.753 --> 23:43.013
[SPEAKER_00]: Right?
23:43.173 --> 23:45.474
[SPEAKER_00]: You guys did something wrong and we need to redo it.
23:45.534 --> 23:55.039
[SPEAKER_00]: That happens a lot because there is a lot of parties involved from consultants to architects and subcontractors and there's so many opportunity for things to go wrong.
23:55.526 --> 23:57.187
[SPEAKER_02]: It's the nature of complexity, right?
23:57.307 --> 24:07.913
[SPEAKER_02]: From the beginning of your career, when I started in this till now and Lumereg, you've seen this, it's like a $250 million project was a big project, a half a billion dollar project was like a unicorn.
24:08.033 --> 24:14.696
[SPEAKER_02]: There, we're talking three, six, 10 billion dollar projects that are multiple JVs.
24:14.716 --> 24:19.999
[SPEAKER_02]: They're JV and at the GC level, you're even JVing in the trade level, you're subcontracting it.
24:20.339 --> 24:24.382
[SPEAKER_02]: So the complexity of current jobs, I never even thought about it.
24:25.202 --> 24:41.764
[SPEAKER_02]: Even if you could sit down Cameron, now you're in a JV project, you're not even you only have your portion you don't have it all you can't be everything some of
these projects there is no way that you could wrap your head around everything
24:42.325 --> 24:54.313
[SPEAKER_00]: So you hope that everyone does their job really well and recognizes where they can impact the next scope and they communicate that very clearly.
24:54.813 --> 24:57.695
[SPEAKER_00]: And you can imagine how that could go wrong, right?
24:57.715 --> 25:00.556
[SPEAKER_00]: There's just so much that people need to handle.
25:00.636 --> 25:05.339
[SPEAKER_00]: And into your point of replacing people, no, we don't see that.
25:05.560 --> 25:11.003
[SPEAKER_00]: What we are hearing and seeing is there isn't enough people to do all of this work.
25:11.683 --> 25:14.586
[SPEAKER_00]: with the Dana Center work and all of the larger projects.
25:14.806 --> 25:18.969
[SPEAKER_00]: It's not a matter of replacing people is increasing their capacity.
25:19.450 --> 25:28.037
[SPEAKER_00]: So that if they're not doing some of these mundane, yes, you do it the first or second time you learn it, but you don't need to do it 60 more times yourself.
25:28.157 --> 25:31.580
[SPEAKER_00]: Maybe you can leverage AI to automate that process.
25:31.840 --> 25:33.962
[SPEAKER_00]: Now instead of you just managing,
25:34.582 --> 25:44.328
[SPEAKER_00]: They drywall scope, maybe you can now manage the drywall and the current wall scope and increase the capacity for that organization and that team to do more projects.
25:44.710 --> 25:57.736
[SPEAKER_02]: Yeah, and I hope eventually it opens up like you started on the constructability review and means and methods and really bringing in the newer means and methods that we're going to need to build at scale because we're never going to have enough people in the field.
25:57.856 --> 26:00.378
[SPEAKER_02]: And you're we're not going to have enough with experience in the office.
26:00.758 --> 26:04.660
[SPEAKER_02]: By the way, I have a theory here at a feeling that and I'll bring you the Luma here and see that
26:05.280 --> 26:07.941
[SPEAKER_02]: In the future because we got to know where's this thing going to go.
26:07.961 --> 26:12.463
[SPEAKER_02]: I think there's an opportunity to sort of democratize that experience.
26:12.503 --> 26:16.125
[SPEAKER_02]: So as Cameron's working with the tool and the knowledge craft and it's learning over time.
26:16.605 --> 26:26.670
[SPEAKER_02]: There's this ability to democratize what he knows because that's our biggest problem right we have a gap between what a new person coming in the industry knows and can do.
26:27.450 --> 26:30.653
[SPEAKER_02]: and what we're unfortunately, we're kind of handing over to them.
26:30.673 --> 26:35.197
[SPEAKER_02]: We're handing over a lot of risk at the current moment, and we're not giving them near enough guardrails.
26:35.698 --> 26:43.265
[SPEAKER_02]: So I'm wondering if there's this idea and where it's going to go in the future on, not only just doing that, but what else do you have planned, Luba, America?
26:43.285 --> 26:46.127
[SPEAKER_02]: Because I can't imagine you're thinking, oh, yeah, we're good.
26:46.468 --> 26:50.571
[SPEAKER_02]: Well, this sell this thing, grow the team, sell this thing, you know, a lot of people in the
26:53.013 --> 26:56.575
[SPEAKER_01]: So, you mean AI, you mean prime point?
26:56.715 --> 26:58.976
[SPEAKER_02]: Yeah, prime point, where do you see it going?
26:59.016 --> 27:04.238
[SPEAKER_02]: But also, if you want to get outside the lines and just tell some other folks what they might want to be working on, that's fine too.
27:04.658 --> 27:12.021
[SPEAKER_01]: So, I guess I shared the vision and inspiration that Hammett, my co-founder, has
27:20.185 --> 27:26.568
[SPEAKER_01]: In the vision of Cloud Cloud is you're just throw any project at it and it will just do the work for you.
27:26.848 --> 27:31.311
[SPEAKER_01]: So, or at least it can help you get started on the project.
27:31.351 --> 27:36.593
[SPEAKER_01]: So, we see a lot of, so, and by Cloud Cloud, I don't mean Cloud specifically.
27:36.633 --> 27:38.354
[SPEAKER_01]: We use Codex and cursor.
27:38.394 --> 27:39.915
[SPEAKER_01]: There are many solutions like that.
27:40.195 --> 27:50.868
[SPEAKER_01]: But software has completely been revolutionized by AI and we see this transformation happening in will happen in construction as well.
27:51.449 --> 27:55.534
[SPEAKER_01]: It has taken longer because construction is a lot more difficult.
27:55.614 --> 27:58.417
[SPEAKER_01]: There are a lot more like drawings or challenge.
27:59.058 --> 28:02.199
[SPEAKER_01]: their unique things in each of the projects.
28:02.519 --> 28:10.222
[SPEAKER_01]: And also, unlike software, it's a good analogy, but software also is a static thing that you're building over time.
28:10.802 --> 28:12.323
[SPEAKER_01]: The project is evolving.
28:12.603 --> 28:15.584
[SPEAKER_01]: In construction, things happen beyond your control.
28:16.224 --> 28:38.536
[SPEAKER_01]: Oh, like this material is delayed or we discovered this problem underground and like there is things are constantly changing and what you want is a system that really can anticipate the change can be more proactive and say, hey, this is what I propose like this will happen here and this is where I can shine by the way beyond the elements.
28:38.816 --> 29:03.909
[SPEAKER_01]: you need to have a very good planning, taking into account everything that can happen like or imagine your building a skyscraper and there is a crack developing and now you need to take photos and you need to figure out from this multi-model experience, the photos and the plans and material science and projecting future love and all of these factors together what is the plan?
29:04.069 --> 29:05.290
[SPEAKER_01]: How should we go forward?
29:05.790 --> 29:13.078
[SPEAKER_01]: This is an incredibly complex problem that we have one way prime point and AI in general will be able to tackle.
29:13.856 --> 29:35.702
[SPEAKER_02]: Yeah, I mean, we've seen real
world examples of that going wrong, and it would be amazing if not only they were able to do it actively but passively, there's so many photos, there's so much documentation going on, it's not even getting cracked at the surface of what computer vision can do with it and then enable everybody else to react so a bridge is cracking.
29:36.162 --> 29:43.586
[SPEAKER_02]: You know, how many of those that we're seeing we're seeing drones do that now and fly them and video them and try to process that.
29:43.626 --> 29:45.347
[SPEAKER_02]: But why are we having to do it from there?
29:45.427 --> 29:49.969
[SPEAKER_02]: Why can't there all these passive things that we're using be collected over time?
29:50.009 --> 29:54.611
[SPEAKER_02]: So I love to hear that opportunity to move it into forward action, right?
29:54.631 --> 29:58.213
[SPEAKER_02]: Because that you're building something that can live down the road, too.
29:58.253 --> 30:01.635
[SPEAKER_02]: I'm wondering Cameron in your world, like, where do you see it going?
30:01.695 --> 30:03.776
[SPEAKER_02]: And maybe if it was the, where would you see it go?
30:04.156 --> 30:05.478
[SPEAKER_00]: Yeah, that great question.
30:05.578 --> 30:14.748
[SPEAKER_00]: So one thing we've learned so far is whatever we're building today needs to kind of fit in with the workflows that people are used to.
30:14.968 --> 30:22.397
[SPEAKER_00]: It can be drastically new thing that people need to learn because you have to change your process and you have to add
30:23.117 --> 30:26.218
[SPEAKER_00]: and immense amount of value for people to just change their process.
30:26.238 --> 30:30.399
[SPEAKER_00]: So we're trying to show the value and fit within the process.
30:31.179 --> 30:37.181
[SPEAKER_00]: And we're not asking people to build complicated agents to accomplish these tasks.
30:37.261 --> 30:41.242
[SPEAKER_00]: So what we're doing today is we know past that needs to be done.
30:41.542 --> 30:51.345
[SPEAKER_00]: We either use prime prime platform to automate a piece of it or facilitate the project manager and engineer superintendent to do it faster.
30:51.685 --> 31:11.594
[SPEAKER_00]: Our view for the future is what you just mentioned, we have this idea of a lessons learned, sort of feature that's more of an enterprise level, where I remember seeing so many 30, 40 years carpenter to general superintendent type of characters leading
31:12.454 --> 31:13.155
[SPEAKER_00]: the industry.
31:13.235 --> 31:17.338
[SPEAKER_00]: And I remember how big of a void that retirement was.
31:17.658 --> 31:21.340
[SPEAKER_00]: That was a super power within the company.
31:21.480 --> 31:22.541
[SPEAKER_00]: It's no longer there.
31:22.621 --> 31:24.743
[SPEAKER_00]: And those people are just irreparable.
31:25.023 --> 31:32.728
[SPEAKER_00]: And we're seeing less and less people like that coming through the field into the ranks of, you know, a general contractor hire up.
31:32.888 --> 31:57.197
[SPEAKER_00]: Our hope is we can tap into some of that organizational knowledge from previous projects whether learning things that became problems in terms of change orders rework, things that became RFIs, how the schedules evolve on different types of projects and what if we could apply all of that to new projects that are coming up and automatically give you a report of
31:59.858 --> 32:07.347
[SPEAKER_00]: given the database of history that you've had and pass on some of that knowledge onto the upcoming projects.
32:07.407 --> 32:11.252
[SPEAKER_00]: So that's something we're pretty excited about tackling in the near future.
32:11.651 --> 32:13.853
[SPEAKER_02]: Yeah, I mean, you're hitting, it's wild.
32:13.873 --> 32:16.615
[SPEAKER_02]: We've been doing this now at the show for 10 years.
32:17.275 --> 32:29.965
[SPEAKER_02]: And even in the beginning of this, we said, the problem with construction is most of what we're doing, contractually during the project, during anything is we're looking over our shoulder for the problem and running into the next one.
32:30.385 --> 32:38.270
[SPEAKER_02]: We're never getting to turn our heads forward because so much of it has to be reactionary so you're really creating an environment where we can get there.
32:38.490 --> 32:41.852
[SPEAKER_02]: I want to give you guys a little credit there because like a lot of.
32:42.805 --> 32:44.166
[SPEAKER_02]: you could run real fast right now.
32:44.326 --> 32:48.849
[SPEAKER_02]: And I would say that I know you guys are moving quickly, but I don't even know how to capture it.
32:48.869 --> 32:51.631
[SPEAKER_02]: I mean, what's the, Louvre Mere, what's the culture like?
32:51.651 --> 33:00.177
[SPEAKER_02]: Because you are, you're moving fast, but you're not breakneck through the baby out with the bath water speed company.
33:00.577 --> 33:05.323
[SPEAKER_01]: Well, I mean, when I was at Facebook, there was this slogan, move fast and break things.
33:05.803 --> 33:08.727
[SPEAKER_01]: Now we have to move fast, but we cannot break things.
33:08.947 --> 33:11.310
[SPEAKER_01]: Move fast and don't break anything.
33:11.470 --> 33:14.493
[SPEAKER_02]: Which is, it bites sound easy, but it really hits it.
33:14.674 --> 33:16.155
[SPEAKER_02]: It's really hard, especially given.
33:16.916 --> 33:46.603
[SPEAKER_02]: That new toolset you're talking about in building code does give you an opportunity and I'll be the first to say everybody on the shows heard it some people are taking that opportunity to move too fast and are breaking a lot of things I don't mean construction I'm not pointing in my construction people everybody knows my apps on the field right now for my travel and whatnot I'm like can you guys stop slow down and test to say and so it is hard because that you're capable of doing that right now so how are you managing that balance.
33:46.803 --> 33:46.870
[UNKNOWN]: You
33:47.142 --> 33:56.684
[SPEAKER_01]: It is uncharted territory because what we're seeing is things that will
take weeks or months, features can be done in a day or less.
33:57.524 --> 34:10.447
[SPEAKER_01]: At the same time, we are in a position and like our two in construction, we are kind of, unless we actively try to understand the code we can lose track of what's going on on the code.
34:10.788 --> 34:12.888
[SPEAKER_01]: Because it's not visual like construction.
34:13.008 --> 34:15.589
[SPEAKER_01]: Our product will bring you all the relevant information
34:17.289 --> 34:27.600
[SPEAKER_01]: So there is this thing where if we move too fast then we get like we lose track of what's happening at the same time The flip side is if there are bugs fixing them is very easy.
34:27.740 --> 34:29.422
[SPEAKER_01]: You just say go fix that back.
34:29.582 --> 34:33.927
[SPEAKER_00]: I would add from the user experience to just coming from the industry
34:34.207 --> 34:55.473
[SPEAKER_00]: you know how trust is so important in everything we do, and especially when it comes to technology, if you are adventuring with a new technology and you lose that trust in form of giving you wrong information or giving you too many false positives now that's creating more work for you to like sift through what's the signal through the noise.
34:55.733 --> 34:57.674
[SPEAKER_00]: After a while you're just going to go back to
35:01.335 --> 35:04.819
[SPEAKER_00]: that's with or without AI or any kind of technology.
35:05.479 --> 35:10.465
[SPEAKER_00]: So part of it for us is yes, there is a lot of movement, there's a lot of demand.
35:11.045 --> 35:19.034
[SPEAKER_00]: We do need to be addressing those pretty quickly, but we also need to be very careful with the quality of it.
35:19.634 --> 35:27.559
[SPEAKER_00]: and staying close to our doing our own testing, staying close to the clients giving it to trusted design partners and caveat.
35:27.939 --> 35:37.845
[SPEAKER_00]: Hey, we just release this, test it, break it, give us feedback, improve on it before we go on a wider release on Amy Tool or Feature.
35:37.925 --> 35:40.126
[SPEAKER_00]: And that's just so important for our industry.
35:40.361 --> 36:00.381
[SPEAKER_02]: Yeah, it's move fast and don't break things with the asterisk like we'll tweak them a little bit and see how it breaks to our user group that we trust all it is is the same loop it's just being compressed a little bit more and but that's what people you know I hear it all the time all a vibe coded into this and it'll be done and it's like no you can't do that you got there's a reason.
36:01.002 --> 36:02.943
[SPEAKER_02]: that the cycle is what it is.
36:03.003 --> 36:04.524
[SPEAKER_02]: But we can amplify it a little bit.
36:04.684 --> 36:06.465
[SPEAKER_02]: We're doing that in construction right now, too.
36:06.825 --> 36:08.026
[SPEAKER_02]: We're accelerating it all.
36:08.066 --> 36:13.829
[SPEAKER_02]: So it's really important to know your blocking and tackling, but then also be able to deliver.
36:13.930 --> 36:16.411
[SPEAKER_02]: So I'm excited to see where you guys go.
36:16.671 --> 36:19.953
[SPEAKER_02]: I'm just thankful you guys came and sat down now, because I think this is a key time.
36:20.013 --> 36:22.935
[SPEAKER_02]: For people to get involved and kind of understand more about
36:23.395 --> 36:24.435
[SPEAKER_02]: where you're going.
36:24.535 --> 36:27.416
[SPEAKER_02]: So, you know, Lou Meer, Cameron, it was awesome to have you.
36:27.536 --> 36:31.077
[SPEAKER_02]: Where can people go find out a little bit more about you, Cameron and the company?
36:31.377 --> 36:34.998
[SPEAKER_00]: You can't either go to primepoint.au or website.
36:35.098 --> 36:47.822
[SPEAKER_00]: There is a form that they could reach out to us, and we'll get back to them with some time to meet, or they can't email us at contacts at primepoint.au and we will get back to them.
36:48.620 --> 36:50.241
[SPEAKER_02]: So awesome, you should be reaching out.
36:50.521 --> 36:54.562
[SPEAKER_02]: Louvre, where can people go follow you and learn a little bit about more of what you're up to?
36:54.582 --> 36:57.324
[SPEAKER_02]: I'm sure we're gonna have a few of my folks that wanna go follow you.
36:57.824 --> 37:02.165
[SPEAKER_02]: I guess on my LinkedIn page, we'll get you out there and any show is coming up.
37:02.306 --> 37:03.346
[SPEAKER_02]: Any where you guys are gonna be?
37:03.526 --> 37:13.010
[SPEAKER_00]: You know, the second round of the conference season coming up so I think we'll be in Chicago for a built world event in July and then we're doing
37:13.530 --> 37:19.836
[SPEAKER_00]: program break, Autodesk University, later on in September and October as well.
37:20.316 --> 37:20.636
[SPEAKER_02]: Awesome.
37:20.777 --> 37:25.541
[SPEAKER_02]: Well, that is where people can go find you and I appreciate finally getting both the all the sit down.
37:25.601 --> 37:26.722
[SPEAKER_02]: This has been fun to watch.
37:27.102 --> 37:38.112
[SPEAKER_02]: Loubomir, so I hope in another 18 months we get you on here and you can tell us all about the new things you as well, Cameron, and thank you for tuning in today to geek out for our interview with Loubomir and Cameron.
37:38.532 --> 37:43.907
[SPEAKER_02]: To read all our new stories, learn more about our guests and listen to the show of visit thecontecru.com.
37:44.047 --> 37:45.992
[SPEAKER_02]: This is the Contecru signing out.
37:46.433 --> 37:50.043
[SPEAKER_02]: Until next time, enjoy the ride and geek out!