Subscribe
Share
Share
Embed
🧠Podcast Episode Description:
AI isn’t just generating art or helping you write emails—it’s quietly transforming how architects work, communicate, and design the built world. In this episode of CheckSet, we sit down with Alex Mead, a former construction laborer turned systems PhD and AI engineer, now helping build Redmark, a next-gen tool for architectural communication and collaboration.
AI isn’t just generating art or helping you write emails—it’s quietly transforming how architects work, communicate, and design the built world. In this episode of CheckSet, we sit down with Alex Mead, a former construction laborer turned systems PhD and AI engineer, now helping build Redmark, a next-gen tool for architectural communication and collaboration.
Alex breaks down:
- Why so many AI tools feel like toys—and where the real power lies
- How Redmark is streamlining meeting notes, decision tracking, and client feedback
- What “technical debt” means for architecture firms (and why it’s coming for your workflows)
- Why AI won’t replace architects—but the architects using AI might
- How privacy, trust, and mentorship are being reshaped by intelligent tools
If you're curious (or skeptical) about what AI means for your job, your firm, and your future—this conversation is a must-listen.
🎙️ “AI will not replace architects. But it will change who thrives.”
Creators and Guests
Host
What is Check Set - Architecture and Technology?
Brian Swem, Senior Architect at Pure Architects, discusses the future of Architecture in an AI world.
[Speaker 2] (0:00 - 0:13)
Welcome back to CheckSet. We're here with Alex Mead today, an AI engineer who's also helping us to build Redmark. Thanks for joining us, Alex.
Yeah, it's good to be here, Brian. Can you tell us a little bit more about yourself and kind of how you got here?
[Speaker 1] (0:13 - 0:24)
Yeah, totally. So, broad background, I actually grew up working construction, and I think you and I talked about this very briefly, but I have a degree in structural engineering actually from Michigan.
[Speaker 2] (0:24 - 0:25)
Oh, nice.
[Speaker 1] (0:25 - 1:16)
I know, right? And then decided I wanted to work with energy and energy systems and went down this whole grad school rabbit hole and went to UC Berkeley and got a PhD in what's called systems engineering, which kind of basically means like electrical engineering and kind of the interface of computer code with the real world processing sensor data and then usually actuating some kind of sensors to change the real world based on that computation. I got into AI actually quite a while ago in 2012, I think was the first like course I took on machine learning or AI, if you will.
To get a little technical, this was before what's called like the deep revolution or the deep neural net revolution. So, like before the ImageNet challenges, before the chat GBT kind of things, all the techniques were quite a bit different. So, I hate to say this, but it was it was rather boring.
[Speaker 2] (1:16 - 1:39)
So, just like that's that's one of the things that I think is surprising to me in that when you really look into or when you start to learn more about AI, people have been talking about it for a very long time or when working in that field for a very long time, but it feels like it's really only been popularized or to people like me more recently.
[Speaker 1] (1:39 - 2:05)
Totally. Yeah, I mean a lot of these things and you know, I'm kind of a geek and I'm always like, we've got to start with the definitions. And, you know, technically to like a computer scientist, artificial intelligence would be anything that would traditionally require a human.
You just have the computer do. So, this is a little bit of a moving goalpost type definition, if you will. But but yeah, people have been talking about these these tools since the 50s and even before that.
[Speaker 2] (2:05 - 2:05)
Yeah.
[Speaker 1] (2:06 - 2:06)
So, yeah.
[Speaker 2] (2:07 - 2:18)
Cool. So, can you talk a little bit more about like what you're doing currently or maybe how you're helping out with Redmark or how that started or how you feel about it?
[Speaker 1] (2:18 - 3:05)
Yeah, totally. So, very broad question. But basically, I'm an AI engineer, if you will.
It's kind of an emerging job title, if you will. Literally in the last like year or two, people are trying to say like, OK, what are these engineers who are kind of at a certain layer in like what's called the technology stack? So, you're not building these like foundation models per se, where you're doing all this like crazy work with large language models like LLMs, huge sets of data.
But you're not like just using chat GPT. You're kind of in this like middle area where you're using a lot of these tools. Another tool would be like a graph database, which we can get into or an embeddings database.
And you're kind of using all these tools to plug together and build novel applications that bring value to people.
[Speaker 1] (3:06 - 3:08)
Yeah, that's kind of the way I think about myself.
[Speaker 2] (3:08 - 4:00)
That that begs another question, just because when I'm experiencing, you know, when I'm just on the Internet or watching YouTube or basically anywhere, the things that I'm seeing people use AI for seem almost like a novelty, right? Like, yes, they have value, but it's it seems like a really narrow use case. So I can find an AI tool that will help me get rid of emails I don't want to see or find openings in my calendar for me.
Right. But I I never run across tools that exist across platforms or would have seemingly like large impact on my day to day life. Yeah.
So do those systems exist or how do you find them?
[Speaker 1] (4:00 - 4:17)
Yeah, well, I think, again, it kind of gets to like, well, what do you want to call AI? I'll give you like a very silly example. Probably 10 years ago, facial recognition, like literally just identifying a person's face in an image would have been like bleeding edge AI.
Now everyone has that in their phone.
[Speaker 2] (4:17 - 4:17)
Yeah.
[Speaker 1] (4:17 - 5:10)
So it's kind of like, OK, well, here was this like really hard problem not that long ago for computers to figure out how to actually put a box around the face of someone. Because it's kind of a complex, like, you know, object detection problem. Right.
And you could be multiple of those people. And what's a person? What's a picture of a person?
You know, there's, you know, in a picture. Right. So that would I think would be a kind of a silly example.
Maybe maybe you would still think that's silly, but that could be one example. Another example that was that is still considered AI, ML, machine learning is like recommendations of products or content. So when you're, you know, going to YouTube and you're looking for a video or you're going to Amazon and you're looking for a product, there is a huge amount of AI computation going on behind the scenes to figure out what to recommend to you.
Sure.
[Speaker 2] (5:10 - 5:47)
So which that's all fascinating to me and it's probably for a different conversation. But sure. Yeah.
Yeah. Short version. I decided I didn't want to use Google for a minute.
That must have been hard. And then very quickly realized that it had learned like what I wanted to see. So at the time, I was getting most of my information from other people through forums.
Finding those forums without having that backup was much more difficult. So I very quickly went back and was like, OK, I'm just, you know, yeah.
[Speaker 1] (5:48 - 6:55)
But yeah, I think there's also a lot of work being done in like genetics, for example, that I think we could all agree would be like a non-toy example. Yeah. So, you know, I have a friend and I don't fully, I'm way over my skis on this.
I don't understand a lot of the genetics stuff, but they're using AI to process like human genome data, which basically since the Human Genome Project, it cost a hundred million dollars to sequence one human. Now, I think it's on the order of like a thousand dollars, something like that. So just like ridiculous drop in cost.
And then now suddenly you're like, OK, sequencing someone's DNA becomes a lot lower of a lift. Now, can we use kind of AI or machine learning to explore these growing data sets of human genomic data to learn about things like, you know, how does heart disease actually form? Is there some kind of genetic markers?
How, you know, are there trends in the DNA that maybe influence, you know, other diseases like autism or something like that that we didn't fully understand before and something we're doing now is mutating our DNA? And, you know, so there's like real medical genomics research going on, too, as well.
[Speaker 2] (6:55 - 7:33)
Yeah, that's great. And I think I, you know, I know that that stuff exists, but it seems like, which I guess is appropriate, right? Like everything starts when it has like money or real purpose behind it.
So these large private systems that are doing really interesting and amazing things with data make sense to me. But it's just it's a little surprising that it seems like there's a very large separation between like what's happening very privately and very nearly behind the scenes versus what's happening out in the public.
[Speaker 1] (7:33 - 8:06)
That could be true. I suppose that. But sometimes I think there is stuff going on under the hood that makes things appear simpler than maybe they really are.
Like even like a chat GPT, for example, like the amount of engineering effort that goes in from building the model itself to being able to run that model on production, to being able to make it secure, all these different things like there's a lot of stuff going on in there. And often the AIML is only a small part of it, granted, but it's kind of the key enabler, if you will.
[Speaker 2] (8:06 - 8:10)
Yeah, I guess that makes me think about just design in general, right? Good design is invisible.
[Speaker 1] (8:11 - 8:12)
There you go. Exactly.
[Speaker 2] (8:12 - 8:30)
So, you know, you really only notice bad design because something didn't work or you didn't enjoy it or something happened. So, yeah, when a light fixture works like it's supposed to, it just works and you don't think about it.
[Speaker 1] (8:30 - 8:31)
It's true.
[Speaker 2] (8:31 - 8:33)
You're doing your task and you have the light you need and that's it.
[Speaker 1] (8:34 - 8:34)
Yeah.
[Speaker 2] (8:34 - 8:46)
So, yeah, I guess that makes really good sense in that my perception of what I'm getting from AI or what's available is that it's all really simple, which means they probably did a good job designing it.
[Speaker 1] (8:46 - 9:02)
Oh, yeah. Well, some things aren't that simple yet, as I'm sure you're aware. But, yeah, there's a lot of things that go into making something user friendly.
And, you know, even a lot of questions around products are, could you ever get someone to use this? If it's too complicated, people just won't use it.
[Speaker 2] (9:03 - 9:37)
Yeah, I actually saw something about that recently in all those seemingly smaller tools. Some companies are actually getting themselves and potentially in trouble because they're just implementing all kinds of things that they can get their hands on and just making kind of a spaghetti mess of processes or just additive processes in order to use these tools, which I thought was an interesting take. Do you mean like in the way that they're building their tools?
[Speaker 1] (9:37 - 9:39)
Kind of like putting together other things?
[Speaker 2] (9:39 - 10:00)
Like piecing together multiple agents with multiple different platforms because this one's good at this thing and this one's good at this thing and trying to get, trying to make a more complex system that does larger, more complicated tasks, but potentially in a way that's going to be difficult to manage in the future.
[Speaker 1] (10:01 - 11:30)
You know, there's actually a term for that in computer science and engineering. It's called technical debt and it's this idea that, oh, we're just going to string this thing together. We're just going to connect these two components because it'll allow us to do this really fast.
And then as you kind of allude, it's like engineers a year later or even six months later, but typically a year later, two years, two years, three years later, they like, look at this and like, how is this manageable? Like this is not manageable. Like I literally was working on a project last fall where I actually had to quote unquote fire the client because it was just like such a nightmare.
It was like such a nightmare and they just didn't have the resources to actually make it do what they wanted it to do. Moving forward, they kind of almost hit like an innovation bottleneck because they had so much debt in their product. It was like, you can't practically get those two pieces of information that you're looking to put right next to the screen.
And it sounds silly, but like it's exactly because what you said, it's like, well, no, because we've got to query two different servers to get that information and then use that to query this other service that then returns the guy's birthday. And you're like, how does that make sense? And you're like, listen, how much time do you have and how much do you really want to know?
And you're like, I don't want to know about that at all. And you're like, exactly. And that's where it's a thing.
Yeah. But it's definitely being accelerated with kind of plug and play tools that are easy to spin up for sure.
[Speaker 2] (11:30 - 12:37)
Yeah. So I'm bringing things back around to architecture a little bit because I think that's what we're supposed to be talking about, at least loosely. It's a good point.
I'm just curious from your perspective and actually having been trained as a structural engineer, you might have more insight than I actually anticipated. But, you know, we've had a couple of architects on here, but I'm curious about your perspective as to kind of why the architectural industry should be interested in AI or leveraging its capabilities. I think the industry in general has, you know, there's always people operating on the fringe and doing new and cool things and trying to use brand new tools and be really innovative.
But I think in general, the industry is pretty slow for uptake. So talking to architects or like, why would AI be something that they want to leverage in their daily life?
[Speaker 1] (12:37 - 13:06)
Yeah. Well, I think my take on what you said is 100 percent correct. There are certain industries, architecture and kind of, you know, building construction certainly comes to mind that are much more risk averse.
And what does that equate to? It means slower adoption of the leading edge and cutting edge of tools in the industry's defense. I think when you're dealing with like high capital projects there, you need to be more conservative.
It's not a little software app that you can just throw together and tear down.
[Speaker 2] (13:06 - 13:08)
If it falls down, it's OK. Exactly.
[Speaker 1] (13:08 - 14:22)
Right. So I remember, you know, speaking of my structural engineering days, I remember we had like very frank conversations when I was getting taught, like, hey, mistakes have very serious implications in our industry. You know, the safety factors we were using were like three, like it needs to be three times safer.
You know, I mean, I know people do different design, like was it LRFD, load reduction factor design now? But, you know, it's just this idea of like there's a reason certain industries are more risk averse. That being said, I think it's like anything.
It's like if you don't keep adopting the latest things and keep moving forward, you're going to eventually find yourself being outcompeted and you're going to fall behind. I think that so that could be one very obvious practical reason. Another reason is you might be surprised what these things can take off of your plate.
And that's something we're looking into with Redmark, specifically on like project communication as kind of a first step is like, hey, you're an architect. You enjoy creating these cool spaces that have these awesome like feels and flavors when you walk into them and have this like wow factor when you're walking down the street. You're like, whoa, look at that new building going up and then you watch it get built.
Right. In my experience, that's what architects love. Right.
[Speaker 2] (14:23 - 14:41)
Yeah, we I'm it's all about the problem solving for me. And and the interactions with people, whether it's staff or clients, consultants. Sure.
Like working with people directly is really fulfilling. And then being able to see the outcomes.
[Speaker 1] (14:42 - 14:42)
Yeah.
[Speaker 2] (14:42 - 14:48)
And but I love solving puzzles. So every design problem is fun for me.
[Speaker 1] (14:48 - 14:58)
Exactly. But you probably don't like reviewing the last, you know, three quarters of meeting notes to try to determine what, you know, client X's opinion was on something.
[Speaker 2] (14:58 - 15:09)
Correct. My current job entails me making those meeting notes. And the number of times that doesn't actually get 100 percent completed is probably not something I should admit publicly.
[Speaker 1] (15:10 - 15:11)
It's this is amongst friends.
[Speaker 2] (15:11 - 15:11)
Right.
[Speaker 1] (15:12 - 15:52)
But yeah, I think that that is the reason, though, to start looking into these tools is a lot of these tools right now, especially at the stage that we're in, are in just making you more efficient with some of these annoying tasks like processing, you know, a meeting transcript and making a brief summary of the meeting notes. That's something one of the things we're doing with Redmark right now is making a meeting notes analyzer. Right.
Another thing is making like to do's based on those like auto filling to do's in your calendar, per se. Like, for example, or in, you know, in software, we use tools like Asana and Jira. But I'm sure there are similar tools and architecture in terms of like just literally to do lists.
[Speaker 2] (15:52 - 15:56)
Yeah, I think a lot of architecture firms use Asana or other similar tools.
[Speaker 1] (15:56 - 16:14)
OK, there you go. Yeah, so exactly. So, you know, basically taking like a meeting you just have, you know, for Redmark specifically, you have a recording device in the room or on the call and then it just takes that information at the end of the call, processes it and gives people to do's and a summary of that meeting that people can review who maybe weren't there.
[Speaker 2] (16:14 - 16:37)
Yeah, I've used, I've personally used Asana in the past, but similarly, I think to what we were talking about, about some of these tools previously is it's like managing it requires a fair amount of effort if you're really going to stay on top of it and it's going to be a meaningful tool for the team.
[Speaker 1] (16:37 - 16:37)
Yeah.
[Speaker 2] (16:38 - 17:04)
So for me personally, I really need something that that does exactly that, that allows me to focus on the project and my team in a more personal way. And but I think my teams would be really happy if there was better task lists and something that they could go back to. Yeah, personally, I could use something that did a lot of that work for me.
[Speaker 1] (17:04 - 17:48)
Yeah, no, and I think that's that's exactly what we're talking about right now is, you know, there's this explosion of all these kind of toy tools, if you will. But at the same time, what that actually is, is engineers and entrepreneurs trying to explore the space and figure out how can we help you? You know, how can we help architects?
How can we help a master architect like you versus a junior architect? Right. These are, I would imagine, I know, very different problems, right?
I mean, that's one of the also the long term goals is how do we make it like an information bank where instead of junior architects coming up to you, they can just ask the tool. And maybe it has some insights because it's been, you know, kind of learning about you and your your career and your experience for many years. Right.
Sure. Yeah.
[Speaker 2] (17:48 - 18:57)
We just had this conversation in the office yesterday. People were asking about mentorship programs and other things. And it occurred to me that if you have.
So I'm really excited about this opportunity to have like a comprehensive communications database for projects. Right. So that it's all.
In a place where you can interact with it, similar to chat, GPT or anything else. And get real responses based on that data. I get really excited about from a mentorship perspective.
Once you get enough data, junior staff can ask questions and get answers that are based on real project data from the firm. So they're going to be able to to understand processes better. And actually, I'll be able to understand processes better, too.
But I think giving people access to that data is going to be really interesting.
[Speaker 1] (18:58 - 19:33)
Yeah. Yeah. And I think that also, you know, something that just came to my mind, I was like, man, I remember when I was a laborer growing up and there were these different terms.
And I was always like embarrassed to ask the foremans or the PMs, the project managers, like, hey, what is what actually is glazing? Right. I guess that one for me always stuck out.
And it's like, dude, it's this glass, bro. But like they're like, you know, I don't care what they say. There are no stupid questions.
There are. And people will 100 percent judge you if you come up to your boss and ask something totally dumb. You know what I mean?
Yeah, it's not good.
[Speaker 2] (19:33 - 19:37)
It's not good, especially in the trades. Like it's not. Oh, it's terrible.
Those guys are ruthless.
[Speaker 1] (19:37 - 20:00)
It's mostly dudes. Right. They are.
They are. They're looking for someone to just like unleash on because that's the human nature. Right.
So even things like that, I think, would be really helpful to have. You know, maybe you can ask Chachi BT that. But in my experience, as soon as you get one click beyond like what is the generic information on architecture, these general models are worthless.
[Speaker 2] (20:00 - 20:44)
Yeah, I use Chachi BT quite a bit, but the. Interestingly, even the language like the it tries to pretend to be, you know, an architect or a project manager or whatever, I try to prompt it to be. But a lot of times I have to go back and fix the language because it just uses terms that we don't we don't use.
I mean, it's trying to be it's a great first draft, which is very helpful. But I think being able to leverage a real project communication in a meaningful way is. I'm really excited about it.
[Speaker 1] (20:44 - 21:18)
Yeah, I think I think just as maybe a final thought on that question of why would, you know, maybe a little more of apprehensive architect want to explore some of these tools? Is this exactly what you just said? Like, you actually are smarter than the thing, trust me.
And it can give you still like a very nice outline to get started with. And then you can tweak that and that can save you. Let's just say it saves you 50 percent on a certain report.
Right. That then allows you to either what? Go get more clients or help a junior architect level up or just go home with your wife and kids or whatever.
[Speaker 2] (21:18 - 22:44)
You know, that's that's one of the things that I get excited about. Redmark, too. And I could you get me too excited and I won't stop.
But please do just the idea that it's it's something that works on a multitude of levels. Right. It can be more revenue or it can just.
Benefit employee personal health and well-being or both. Right. And or or employee satisfaction and retention, like it doesn't have to be any one thing.
But that also makes it really hard to figure out how to market it, because which one should I be talking about? I don't know. Maybe all of them.
But it's really hard to talk about all of them because there's a lot. So, well, I think since you brought up apprehension, I guess another question that I had and I think this is something that gets talked about pretty often, even outside of architecture would be like the the person AI balance. Right.
Like there's people that are afraid of losing their jobs potentially. Yeah. And then so like, do you have an idea or of what an appropriate balance between AI to augment?
And it's probably helpful to stay in the architecture space just because if we go outside that, the conversation gets really interesting.
[Speaker 1] (22:44 - 24:31)
Yeah. No, totally. I think that I mean, specifically about architects, you know, it's it's what are your use problems and it's going to be different for different people.
Right. I mean, my understanding right now is a lot of the drafting tools are still just not quite where they need to be. Again, you know, I'm not an architect, but that's that's my understanding.
A lot of the image processing to actually go through plans is just not there yet. Could be wrong on that, but I think that's still pretty accurate. But I think in terms of, you know, it's like everything.
You need to start to adopt the tools where you see that they're actually helpful for you personally. And that's kind of where you're going to strike that balance. You know, if you think it's helpful just for drafting emails, cool.
If you think it's helpful for taking a transcript and processing it from a meeting, cool. You can do both of those things. If you are trying to, you know, have some kind of a design session with it, like there's a guy in town who I had a conversation with who is building a like virtual CEO, like CXO suite.
Right. So like CEO, COO, CTO, these kind of folks. And then as the CEO, you can use that tool to have like conversations about it.
So maybe you want to have a similar like design session with CEO or with a chat GPT or plot or whatever about the actual design process. Let's say, you know, not like actually going through the design per se, but like, you know, kind of the the actual building program you're trying to figure out. Like, OK, what you know, what what is a typical hospital in the Midwest that is a level one trauma center have?
And, you know, it's like, OK, and it will give you just a list to start with. And you're like, OK, yeah, all those I was kind of thinking. But, oh, man, maybe I wasn't thinking about, you know, I don't know, a parking deck or something like that.
[Speaker 2] (24:31 - 24:31)
Yeah.
[Speaker 1] (24:31 - 25:11)
And you're like, oh, OK, that's interesting. It's getting me thinking now. Now I can kind of ask follow up questions.
And that's a great way to get started. I think other things around that could be even processing like submittals. I remember doing that when I worked for a construction company.
We would you just have to read like mountains of like doorknob submittals. Yeah, it's rough. You know what I mean?
Like what if you were able to just like scan those in, do like optical character recognition, like OCR, like having it converted to text and dumping that in like a tool or even not having it converted to text, just having the image, just read it and say, does this meet the specs currently in our project? Yes or no.
[Speaker 2] (25:12 - 25:20)
Yeah. And I think I think some of those tools are tools are starting to show up. Yeah.
So we use a program called Part Three in the office and it's for construction administration on the architect side.
[Speaker 1] (25:20 - 25:21)
Oh, OK.
[Speaker 2] (25:21 - 25:28)
And I think they're I think they just they launched it recently. I haven't had a chance to test it.
[Speaker 1] (25:28 - 25:28)
Yeah.
[Speaker 2] (25:28 - 26:22)
But you can upload your specs and then you can throw submittals in there and it'll say, like, yeah, these ASTM numbers match. So you're good. Or so I think there's definitely people working in that space.
You just gave me an idea for Redmark, though, OK, that having a client portal where they can have a conversation with it. Because we're basically talking about logging communication anyways, but the opportunity for clients to just talk about their needs and their wants openly and potentially have some some questions that help them think about their project. But then all of that information would end up.
Available to the project team, yeah, to query and to help develop program or to to help further the project. You know, I would imagine it would be really interesting.
[Speaker 1] (26:23 - 26:51)
Yeah. No, I would imagine that, too. It helps with like kind of those language differences that you're talking about, you know, like, yeah, sure.
As most I'm guessing most of your clients don't have like the architecture vernacular around certain building structures and that kind of types and themes. But they kind of know what they like and you can kind of really have like an iterative conversation where you can really kind of dive down and then you as the architect could take that and be like, oh, OK. They're talking about this kind of a style.
[Speaker 2] (26:51 - 26:51)
Right.
[Speaker 1] (26:51 - 27:45)
Yeah. Yeah. There's a lot of cool applications like that.
But to get to your broader point, though, is, you know, if as an architect, you're like, I'm a designer, I don't need these tools, I feel threatened. You're leaving a lot on the table and a lot of the other architects who maybe are more open minded to handle these tools like submittal, you know, processing or, you know, meetings processing or these that and the other. They're getting all that out of the way in maybe 10 minutes of their day, 20 minutes of the day.
You're spending three, four hours of it on your day. OK, that means they just get to do the fun parts about architecture more. And you're like, you know, kind of looking down and being like, oh, it'll never work for me.
And it's kind of like, yeah, but, you know, there's always going to be a place for humans in this world, like this idea that these AIs are just going to like, you know, take over and this and the other. It's like it's just not happening right now.
[Speaker 2] (27:45 - 27:56)
Yeah, certainly. And especially in architecture. Right.
We talked about how slowly things can move in architecture in the construction industry.
[Speaker 1] (27:57 - 27:57)
Yeah.
[Speaker 2] (27:57 - 28:09)
I know that Revit as a program was supposed to change the way we do architecture because we're building models. We're not going to need drawings anymore. We just give the model to the contractor and they build it because they have a digital copy.
[Speaker 1] (28:10 - 28:10)
Right.
[Speaker 2] (28:10 - 28:26)
And still not happening. We're still giving out drawings. I do think that by and large, we don't print them as often as we used to.
So printing is actually gone down. But that's even been for me just in the last couple of years.
[Speaker 1] (28:26 - 28:26)
Really?
[Speaker 2] (28:27 - 28:36)
And we've had, you know, computers and PDF technology or, you know, the ability to look at PDFs or images on computers for a long time.
[Speaker 1] (28:36 - 28:46)
So no, totally. And the even things like iPads have been around now for what, 12, 13 years. Yeah, that's crazy.
So, I mean, you know, if you're saying it's just happening in the last few years, you know, yeah.
[Speaker 2] (28:48 - 29:06)
So I think the next thing and maybe we were already there a little bit, but it's just diving in a little bit more about Redmark and what's under the hood. And hopefully it's exciting for you, but also just, yeah, just to let you talk about what you do and learn a little bit more about Redmark.
[Speaker 1] (29:07 - 30:53)
Yeah, totally. So, no, Redmark is my favorite project right now that I'm working on. It's I'm not just saying that to promote the podcast and architecture, but it genuinely has a lot of cool things under the hood.
So basically what it is, is it is a way, like you mentioned, to handle client communication and project communication kind of at all levels. So like between the teammates in the architecture firm actually building stuff, between the client itself, you know, communicating that way, kind of all internal communications like emails, maybe Slack or Google chat, things like that. And basically kind of processing all of these things and trying to derive insights from those.
So imagine you have a transcript from a meeting and it could be as simple as like, you know, we've mentioned this one, but like give me the highlights from this meeting and give me the meeting notes. And that's a document then. Then you can automatically fire off an email to everyone who's in that meeting.
Hey, here's the meeting summary. You know, there's people are starting to do that now, but how do you do that in a industry specific way and then be able to derive industry specific insights from that? That's what Redmark is, is kind of collecting all of that information in one place and then using kind of more advanced data structures or data science around that.
So then you can start to make, you know, not only kind of like spatial learnings, like what's going on between different team members, like, you know, for example, the meeting that we just had on Tuesday, like you and Josh talked for a while. And then I showed up and we all talked for a while. Like, what if Josh left and you and I just kept chatting?
Like, is there a way to kind of merge those two meetings so we can all understand what was communicated, right? I was, I was thinking about that this morning. There you go.
[Speaker 2] (30:53 - 31:21)
The, just the idea that, you know, inevitably there's somebody that says, oh, I really wish I was in that meeting because I didn't get that information, but you record the meeting there. In reality isn't enough time or desire to watch the recording or listen to it. No.
Um, and meeting minutes are valuable, but they don't necessarily give the person the background or the information that they need.
[Speaker 1] (31:21 - 31:21)
Yeah.
[Speaker 2] (31:21 - 31:40)
So the idea that you could have a team member that wasn't able to be present for a meeting, then, uh, interact with the transcript from that meeting in a way that allows them to quickly access the information that's relevant to them. That's also very interesting to me.
[Speaker 1] (31:40 - 33:47)
Yeah, totally. And I think what's even beyond just like one meeting, like we're talking about, like a whole series of meetings. So like, imagine like, you know, a big kind of like network of meetings and people who are in those meetings and kind of all these lines connecting them all.
Right. You can imagine that kind of in your mind as maybe a big spider web of meetings who was in that meeting and then kind of through time. Right.
And then you could build this kind of, you know, the red mark, um, kind of data structure, if you will, in the background to actually say like, Hey, what are common themes with this client? You know, and they could be everything from like, Hey, it seems like this particular person on the client team has like outsized influence because everything that they say three months later ends up being implemented some way in the roadmap. Right.
Sure. And you're like, Oh, but that's not the decision maker. And it's like, well, maybe they are.
Right. I mean, there's all sorts of team dynamics and things like that. Those are like the super exciting things.
Right. Where it's like, Oh, it's not even just the CEO of our client said this and they're the decision maker. And it's like, actually we determined that this other person who is, you know, maybe his administrative assistant, right.
But has his ear in some way and they're able to influence these designs. Right. So, you know, I don't know.
These are, these are the interesting things I think that a tool like Redmark potentially opens up because you could be like, Oh, you know, another, you know, an example I've heard of like, why is the current CEO of Google? The CEO of Google and not someone else. And it was literally because Larry Page noticed that none of the people would agree on stuff in the C-suite unless this guy was in the meeting.
And he's like, yeah, I don't know if this is apocrypha, like I've heard this and it's kind of like, that's fascinating. Yeah. And now he's CEO of Google.
Right. So like, are there insights like that? Like, hey, whenever this person is in the team, we look back in the meeting, we noticed that like decisions were made in that meeting, but it wasn't the decision maker that we thought, right?
Right. I mean, that could be like, you know, an insight that we're potentially going to be able to gleam with a tool like Redmark.
[Speaker 2] (33:47 - 34:03)
Right. And that's another thing that I get really excited about when we talk about like aggregating this data is I think there's a lot of possibilities that I'm sure we're not even thinking about right now.
[Speaker 1] (34:03 - 34:03)
Yeah.
[Speaker 2] (34:04 - 35:33)
Another one that while you were just talking popped into my head was the idea that like I had heard that call centers are using AI. So they're actively commenting to the person on the phone call. So they were using it to augment like human interactions.
But they were, it was more than 30 percent better or more successful. Oh, wow. And it was because AI was like essentially reading the room and then giving prompts to the person, to the person at the call center.
Like this person seems like they're getting upset. Maybe you should try this or like that question didn't land. So you should try this.
And, but you know, dealing with, dealing with people, there's always a potential for people to butt heads or misunderstand each other. So even there, having insights into client reactions during meetings and being able to figure out why, right, or what's possible or what we could be doing better as a team is also really interesting to me. Because typically, you know, you get a couple of people in the room and you're like, boy, that meeting didn't go very well.
How do we make sure that doesn't happen again? But when you, you might be able to get much better insights with the kind of data sets we're talking about.
[Speaker 1] (35:33 - 36:58)
Totally. I think so. Yeah.
I think there's a lot there. And I think just to kind of assuage a lot of the adoption fears, especially in an industry like architecture, it's like we humans really crave human connection, like really a lot. And so I would never, I'm just going to tell you, like there is never a world where like people are building buildings and humans aren't involved.
In some capacity in the design process and kind of like actually enabling that. Like, I think COVID taught us a whole lot, but like one of them was that like isolation is really bad for people, like really, really bad. Um, so much so actually, I heard this recently that like one of the only tools that, um, like maximum security prisons have is isolating people.
Cause like even the most like deranged kind of, and I don't use the term lately, like psychopaths, like still derive some need of being around people. So it was like literally the only punishment. So it's kind of like, look, the human element in being able to mediate these communications and sales and that kind of thing is always going to be a thing.
So why wouldn't you want to use these tools that enable you to do a better job at that or enable you to even identify like maybe your own blind spots? You can be like, Oh, like why, when I'm in the meeting with that person, does it always seem to not go well? You know, it could be a communications issue, you know, and those are the kinds of things that a lot of these tools are going to be able to start to suss out.
I think as well.
[Speaker 2] (36:58 - 37:57)
That's, that's really, that's cool. Yeah. And I look forward to seeing it happen and figuring out what we can do with it.
Yeah. Uh, we say data a lot in this. Uh, I think probably the most frequent question I get is about data security.
Uh, and the idea that, you know, recording meetings and which has gotten, people have gotten pretty comfortable with meetings at this point. Um, it'll, you know, you throw email and Slack conversations and other things into the mix. Mm-hmm.
Um, yeah, I think people start to get less comfortable, um, way less comfortable. If you talk about recording just regular conversations on a regular basis, but how does, um, and just the idea of, um, you know, intellectual property and the things we're talking about have value and where does that go? And does it end up being open source at some point or is it secure?
[Speaker 1] (37:57 - 40:45)
Yeah, well, so I can comment onto that on multiple levels. So with specifically with the Red Bark tool, you know, we're doing all the standard kind of web technology stuff. Like everything is encrypted, everything is, has, you know, kind of permission to access, like not only kind of internally, like amongst our team, but how you can share information.
All that is like kind of very traditional, like data management or data governance is the technical term. Right. Um, and that's, there's a lot of freedom in how you want to do that.
Like anybody could access anything if you're on a small team, maybe that makes sense. Or if you're on a larger firm, like, Hey, only the principals can access this particular information or only people who are in a certain call can access that transcript, uh, unless you have permission to share that, something like that. So there's a lot of different kind of granular permissions like that.
Um, something that we're doing, uh, on the more technical side of thing is what's called a single tenant architecture. So all of your data for your firm is totally separated, like on a totally different server from any other clients that are using the tool. So, you know, if there's architecture firm A and architecture firm B, they both have completely separate sources of data all the way down to, you know, what the engineers would say is like to the metal, like all the way down, there's no like sharing across the information.
Um, so that kind of, you know, starts to, you know, alluse a lot of the fears that people have, um, around data, but, you know, at the same time, I mean, there's, um, there's also the risk of, you know, people hacking your servers and these things do happen, but, you know, that's no more risky than say putting things on like Google cloud or, you know, one drive with Microsoft. So I often try to answer the security question with a kind of, as compared to what, and ask it back. And so it's kind of like, well, do you feel comfortable putting information into, you know, your Google drive?
And most people are like, yeah, totally. I totally feel comfortable with that. It's like, okay, well, we're basically Google drive or safer in terms of your documents.
So, you know, of course there's a way that people could hack. There's this, that, and the other. But like the probability is much more likely that somebody in your team chooses a silly password and they go in the front door, right?
So that's often the question with security is as compared to what, you know, um, but I mean, the idea of there being this centralized repository of information, you know, is potentially a scary thing. And you kind of alleviate a lot of the groundwork for potentially anything from something maybe embarrassing, just like a personal quirk to, you know, potentially very heavy legal implications. So, I mean, there are risks to these kinds of things, but, you know, I think you kind of put it best, uh, in our meeting on Tuesday is like privacy is almost an illusion now.
[Speaker 2] (40:45 - 40:51)
Yeah. I didn't know if I should get into that here, but because, uh, like you said, as my take might be a little dystopian.
[Speaker 1] (40:51 - 40:57)
Oh, well, I didn't mean to put you on the spot like that. But mine is a very similar.
[Speaker 2] (40:57 - 41:28)
I do really wonder. I think that, um, I, I learned years ago now that, um, I had a real estate agent, like really successful real estate agent, um, or agency that was a client. And they had talked, they were talking about how they're able to buy insights into whether or not people are interested in buying or selling a house, um, or where they're at in their life because it all just came from credit card purchases.
[Speaker 1] (41:28 - 41:28)
Yeah.
[Speaker 2] (41:29 - 41:42)
So like if you're buying, like, so before, like even before we had cell phones in our pockets and other things, people were still being tracked just based on purchases. Yeah. So if you're buying diapers, they probably just had a kid.
[Speaker 1] (41:42 - 41:42)
Yeah.
[Speaker 2] (41:42 - 41:43)
Um, right.
[Speaker 1] (41:44 - 42:14)
So, you know that when you just said had a kid, I actually heard this story and this was at least 10 or 15 years ago. And it was a dad who found out his teenage girl was pregnant because target started sending him a baby stuff. And it was because she had her own target card and maybe she was a little older, like, you know, 19 or something like that.
But like target knew before her dad knew that she was pregnant and you're just like, Whoa, that's kind of.
[Speaker 2] (42:14 - 43:00)
But I think, I think that kind of stuff has been happening longer than people realize. Exactly. Yeah.
Uh, so, and I, so I very quickly personally get to the space where, why don't I start to benefit from that data instead of other people benefiting from my data? Right. And then that just, that requires me to, uh, either develop with Red Mark or utilize tools that allow me to leverage my own data.
Right. Exactly. Um, so that's, that's basically where I've landed is that I feel like everybody else probably has it or access to it if they want it.
Yeah. Um, but, and I just want to give myself access to it at this point.
[Speaker 1] (43:00 - 43:42)
Yeah. Or at least kind of leverage it for your own personal benefit or not in kind of a, you know, Machiavellian way, but like in a way of like, how can I help bring more value to my customers, my clients who are building buildings? And the risk of someone hacking this server, a lot of this information is already out there.
Like, you know, why don't we use that to our benefit? Yeah. I think there's a lot of sense to that.
Yeah. But that being said, we are using, you know, very advanced solutions to make sure your data is very private to only you. Um, you know, and so there are also just the, um, the reality of like, the only way some information gets out is if you release it, but that's up to you.
We try to do that, you know.
[Speaker 2] (43:43 - 43:54)
So I think tied to that, and I don't think we need to spend too much time on it, but, uh, using large language models to, uh, to process might not be the right word, but to process this data.
[Speaker 1] (43:54 - 43:54)
Sure.
[Speaker 2] (43:54 - 44:03)
Uh, there's also concerns about whether or not we're training the models and then that's a way that the data is getting out. Um, is that something that people should be concerned about?
[Speaker 1] (44:04 - 44:49)
Um, not with RedNord specifically because we're using, we're hosting our own models. So no, is the short answer on that. Um, we can use ChatGPT under the hood, which, you know, some of our demo projects do just cause it's easier to, um, use what's called an API and just kind of outsource that service.
But, um, for the most part, no. The, you know, if you, again, if you trust the service agreements of say open AI API, I mean, they're saying they're not doing it, so they might be, but they're saying they're not. Um, you know, so there's potentially opening up a lawsuit there.
I know, you know, people are saying, oh, who do you trust? And don't trust anyone and this, that, and the other. But it's like, the reality is like, in order to live in a society, you just, you do have to trust people.
Absolutely.
[Speaker 2] (44:49 - 45:08)
Yeah. This idea of a. I run into that on a regular basis.
Uh, and my attitude is, is that I exist in a service, you know, I'm a service professional. And, um, if, if I expect people to trust me, then I need to trust other people. Right.
[Speaker 1] (45:08 - 45:08)
Yeah.
[Speaker 2] (45:08 - 45:43)
Um, and that may be someday that gets me in trouble, but I'm just going to keep going until I find out. Because, you know, it's, uh, I have a really hard time with, I need to watch this person like a hawk because they're, you know, doing a project for me in my house or whatever, because they might do something bad. Yeah.
I don't, I don't want clients to treat me that way. Totally. And, um, so that's generally how I try to make myself think about it, but it's not always easy.
Like sometimes there's always a little bit of skepticism and wondering if you can trust somebody. So I understand people's concerns too. Yeah.
[Speaker 1] (45:43 - 46:52)
Well, and I think that's kind of, you know, the, the kind of the age old practical advice is, you know, trust people step-by-step, bit-by-bit, you know, and then at the end you can, you know, kind of derive conclusions around that. We have all these subtle cues in our culture and society and even just humanity, right. Of like, oh, this person is making eye contact with me.
Okay. That's tends to be, you know, it's harder to lie when you're actually looking at someone, right. It's kind of an age old truth.
Now, is it always true? No. There are plenty of people who can look you in the face and lie and it's no problem for them.
Um, you know, just kind of the professionalism of the appearance of certain products. Let's say, for example, there is evolved a standard of what a real product on the internet looks like. You know, that's a kind of a gate you need to go through.
It's like, okay, I wouldn't upload my bank statements to something that looks like HTML from 1998. So, you know, there's kind of like, uh, you know, uh, you know, don't just be rash. Um, you know, but at the same time there needs to be, um, a reasonable understanding that, hey, like you're trusting a lot of people in this society and, and, and that's enabling us all to do a lot better things.
[Speaker 2] (46:52 - 46:53)
Yeah.
[Speaker 1] (46:53 - 47:02)
Like there's, and there's, yeah, not to get too geeky, but there's a whole bunch of like game theoretic research that suggests you can be a lot more successful if people actually are honest and trustworthy with each other.
[Speaker 2] (47:02 - 47:03)
Yeah, I believe that.
[Speaker 1] (47:03 - 47:03)
Yeah.
[Speaker 2] (47:05 - 47:21)
So, uh, well, thank you so much. I think the only other question I have is what's next. Like, what, what are you excited about?
Maybe we touched on it already, but excited about for Redmark or, you know, if we get a few years from now, what do you think? Yeah. Architects are going to be working with?
[Speaker 1] (47:21 - 48:36)
No, I think, um, we have not touched on what I think is the most exciting and that's actually getting it in the hands of a few more architecture firms and actually seeing how they use the tool. Because you always end up deriving, um, more insights from users as they start to actually interact with the tool and you get more ideas, you get more, um, you know, uh, inspiration, if you will, of, uh, to even, even to just keep working harder when you actually see people using it. And you're like, no, this, someone's actually using this.
Like, um, so I'm super excited about just getting it in more people's hands, honestly. Um, on a technical side, I could bore you to death with a bunch of these, you know, graph databases and all that. But I won't even get into that.
But I think actually seeing real project data, seeing the real value derived, like, Hey, people are actually using the, like, make meetings minutes, make meeting minutes feature, right? Oh, Hey, like this is actually processing, making to-dos. Oh, Hey, we're actually like linking together summaries of meetings over the last six months, 18 months, 24 months, whatever, for a like long scale school project, for example.
Right. Where it's like, Oh, we've derived like these summaries. Like, I'm just really excited to see where that, that starts to go.
[Speaker 2] (48:36 - 48:42)
That's great. Me too. Um, so I'll do my best to get it into other architect's hands.
[Speaker 1] (48:42 - 48:42)
Yeah, totally.
[Speaker 2] (48:43 - 48:57)
Uh, well, thank you so much for your time. Uh, it's been a great conversation. I really enjoyed it.
Uh, and I look forward to seeing what's next for Redmark and watching everybody go nuts about it.
[Speaker 1] (48:57 - 48:59)
Appreciate it. Thanks so much for having me, Brian. I really appreciate it.
It was fun.