The Impact Of Tech And AI On The Built Environment And Profession Of Architecture, Ft. Jordan Harris

The Impact of Tech and AI on the Built Environment and Profession of Architecture, ft. Jordan Harris
===

[00:00:00]

Stephen Drew: Hello everyone. Make sure that Facebook, Google aren't watching. We're gonna be talking about ai. We're gonna be talking about tech generative design. Is your job gonna be replaced? Who knows? Maybe your employer will have a script by the time this podcast ends, which will replace your job. I'm only choking or be revealed in 15 seconds.

I'm joking. No one's gonna lose their job. Promise the robots, the saintsbury can't get a self seven checkout tail, right? You've got nothing to worry about. 3, 2, 1. Woo. everyone and welcome to the livestream. I'm just joking around. You are not gonna lose your jobs, not just yet anyway, I don't think [00:01:00] so.

What am I talking about today? I'm talking about computers, tech, ai, technology in the Architecture. Profession, but the built environment we're all experimenting with. Chat, G P T at the moment. And last year, mid journey took off. Everyone was generating all these cool buildings. So it's probably on trend to talk about now the future.

And what we think will happen with all this tech and how it's gonna impact your jobs potentially. But rather than just me talking about it, it's always better if I've got an industry expert and that is why I've got the all awesome Jordan Harris here. Jordan, how are you sir?

Jordan Harris: Not too bad Overall.

Stephen Drew: Jordan, you get an A star for getting here, especially when I almost messed up the invites to get here. So you've zoomed in for this live appearance. I appreciate it. But Jordan, while I know you from the industry, for anyone that [00:02:00] doesn't know you and what you do, can you tell us briefly about yourself?

Jordan Harris: Yes. Okay. So I'm my name Jordan, obviously. Job description's a bit weird. My technical job description is Systems Architect and Digital Operations Manager, which is basically a made up term for they couldn't find a good role for me. I start pretty much I started in Architecture and I did that for many years.

I did the conventional part one, part two undergrad, masters. But I've always been more tech oriented and I began to take a shift more towards the tech industry. Learning the basics of coding and where that can go. From there I expanded more into sort of the BIM side of things and also taking courses in AI and in data science, basically building ai, machine learning algorithms.

And where I'm at right now is I'm trying to take all of these technologies and Architecture, Architecture being something that's always been a little more technophobic and bring them together and see where we can go with those. In the future and see where [00:03:00] that's gonna be going in the next 10, 15 years.

That's my role at the moment, as well as the general day-to-day fixing people's computers and other such sort of more, more mundane areas of technology as well.

Stephen Drew: Oh, there you go. You gotta, we've gotta have all angles, isn't it? And unfortunately, like you said, as being an Architect, you've got Outlook, you've got read it, but also you're trying, dealing with the procedures and the all. This new cool stuff as well. So I find it really fascinating. And one of the things that we talk about on the Architecture Social as well is alternative careers.

So maybe just before we talk about I've got my own thoughts on AI and chat, G P T and all that stuff, and we're gonna save that layer, but also the, in the audience here, if you do want to share your thoughts on anything that we're talking about. If you wanted to ask Jordan a question, you can do as well.

But before we do that, Jordan, I'm always curious cuz like myself, you are a part two. You are a qualified part two. Architectural. Assistant. You're not an Architect. Neither am I, which is [00:04:00] totally cool. Now I

Jordan Harris: quite gone down that one yet.

Stephen Drew: No may, maybe one day you can do it, but you've got a lot on your plate with the it, the digital side of the BIM.

You, you're doing enough. You don't have to worry about qualifying for now, but tell me why did you choose to go down this avenue of an alternative career then the typical part three.

Jordan Harris: Yeah, it's many reasons. One of the reasons I. Quite like tech is that it's a little less prescribed and regulatory than Architecture. I think when you do undergrad, you you see all the pretty drawings, the amazing designs, and you have, and you envision yourself as, the next Renzo piano or VR gingles and or any of these sort of very influential architects.

And then you get into industry and you realize that is absolutely not the case. You spend most of your time doing schedules and writing design and access statements, which is fine. It's all well and good, but it's. Everything is very forceful. You have to fight for anything, any creative license, and it's very draining, very, you see a lot of burnout in Architecture.

I think people that sort of become complicit with design, [00:05:00] they just give up a little bit and they go, that does the job because it's, it settles, we've done it, it meets the regulations and there's a little pushing of the envelope I think there, whereas in tech, If I want to go and make a new website, or if I want to go and develop a new technology, I can sit on my laptop and I can just plug away and do it.

And there's no restrictions or requirements and there are more coming in over time in terms of, GDPR and privacy, but very limited. So it gave me a lot of options to expand and grow and it still scratches that same itch of creativity that Architecture always did. I think a lot of people think of tech as not massively creative, but it absolutely can be.

It just depends on how you take it and where you go with it. And I also feel in the same way as we used to have drafts persons as massive part of the industry and then Auto Cloud and MicroStation. Things like that came to the industry and those began to phase out. They haven't gone entirely, but they're mostly removed or translated into newer roles.

And we are gonna, we've seen that again, a little bit with BIM. I feel like it's probably gonna be the same again [00:06:00] with ai where we don't, it won't. Knock out the whole industry. I think a lot of people are over exaggerating its impact. It will be a big impact, but it's not gonna be the impact that I think a lot of people think where it's just automating, designing.

It'll more be the tools will be more efficient, so you need less people to do the same job

Stephen Drew: Yeah, said. It's it is interesting. And just touching upon that, you're right because, so 10 years ago in the uk all the practices were moving over to Rev. And some practices still are. That's the crazy thing. It's not clear cut. And as Jordan as well it's one thing to say, oh, we are gonna get we're gonna get Revit in the office.

There's procedures, there's workflows and you touched upon it a bit and I agree that actually. There's an art form in designing the processes to enable designers to design. So that's a bit of a mouthful, but you get what I mean? So is that one of the things that you really enjoy in your role then, Jordan?[00:07:00]

Like optimizing the workflow, trying to get everything as sufficient, as easy as possible for

Jordan Harris: It sounds as boring as all hell, but it, yeah, I really enjoy that. I spend probably more time than I'd like to admit, sitting there designing file structures and how do we make sure people do this and how do we make sure that, so no one's it goes rogue and starts using random line types or random families in models and breaking stuff.

Sounds very boring, but it's one of those pet peeves I actually quite enjoy.

Stephen Drew: Yeah, fair enough. And another bit that resonated with me, so I I put out an advert recently for the Architect social, if people want to do a bit of. Writing in terms of content, then please get in touch for the newsletter now in theory with Jack g p t and we will talk. I want, I'd love to get your thoughts, OnIt potential Architecture applications and that in just a second.

But what the thing is, right? I'm like you, it's a really great tool. For maybe some mundane staff or some basic things or some [00:08:00] quick directions, but that's the key thing. I feel like I still need to direct it on what to do. I don't see it necessarily replacing someone's job yet. I don't necessarily seeing it.

Perhaps designing a residential building. However, I can see it being really useful for perhaps an email that, if for instance, you're tired you're not sure you want to just check things and maybe it suggests an email, there could be a practical use for it. What's your thoughts on everyone freaking out about chat, G p T and AI at the moment?

Jordan Harris: So there is actually already an application out there that designs buildings using ai. It's called Finch. It's currently due to be released in the next few months. It's in a beta program, I think. And I, it plugs into Rhino, if I'm remembering correctly.

Stephen Drew: Oh wow.

Jordan Harris: and runs through. Yeah, it's it runs.

I can't actually remember the model it used off the top of my head, but it's essentially utilizing a Python script through [00:09:00] Dynamo that, yeah, and it's able to automate design. The thing with this, and this is actually something I was saying to one of my directors the other day, it doesn't matter how advanced the stuff gets, because at the end of the day, the only time this will replace your job, if you are the one thinking about design.

So if you are a CAD monkey or BIM monkey, Then you might want to look into this a bit more. But otherwise, if you are the actual creative, if you are the person that's making decisions, you're understanding the design and the user, and you are translating that into the built environment, you are fine.

Because at the end of the day, the only time this is an issue is if a client comes to you and they know exactly what they want from the get go because they have to ask the question. And that's the thing a lot of people forget is you have to know how to ask the question if you don't know that you can't use the tool.

And there is a very lack of appreciation. There's a considerable lack of appreciation for that as a skillset. You have to understand why do we do it this way? And also understand when the AI is wrong. [00:10:00] AI are not perfect. They're precedent based. So the, this one here uses something called a recurring neural network, I believe, which is essentially means you have Nodes called neurons.

And what happens is each neuron looks for patterns. You give it the training data set and the test data set, and you split it. So what you do is you essentially give sort of 70, 80% of it to learn, and then you give it the correct answers. And what it's gonna do is it's going to try and build these neurons and look for common patterns until it spits out something that is correct.

Now ais can make mistakes they can look for. For instance, if you, let's say as an example, if every single design you scanned was from the same practice. It might see the footer or a particular detail or a particular line weight or pattern as a correct design element because it's common. But it might just be that is a graphical choice

Think that is actually something that should be InDesign.

But it could be a fluke or it could be. There's you lots of examples of this. For instance hospital very recently used a similar sort of [00:11:00] system for this to try and detect lung cancer. It was very successful. I think it was like 99.9% successful. In testing, it wasn't, the reason it was successful is cause it turned out every single one of the people that actually did have cancer, there was a small dot on them that the doctors were putting to mark, and it would see those on every single scan and realize that those are common.

So

Stephen Drew: Oh wow.

Jordan Harris: yeah. So it, look, it's just looking for common features, but it isn't that it's looking. For it doesn't understand I need to look for this feature. It just knows I need to find something common between these elements. You can do that with buildings as well, so you can, if you give it 300 floor pans that are good and you categorize it and you say, this is a good house, this is a good, these common elements need to be there, it can probably figure that out.

But you also need to have to look at that and understand, actually we probably shouldn't have a 1200 by 1200 bathroom. That might be a little on the small side. And it might have figured that out because it might have gone okay, actually most bathrooms maybe are long and thin in these house designs.[00:12:00]

That's long and thin. That's long and thin, but in a different orientation, I'm gonna crop the two down because both are correct. So you can easily get mistakes like that form. So it's a tool to speed up the process and to try and remove a lot of that variation, InDesign. But it isn't inherently just a designing tool.

It's a. It increases the workflow rather sorry, improves the workflow rather than simply just replacing the designer. I think it's the most wordy way of saying, don't worry too much.

Stephen Drew: I agree with you. And and even a very basic example. So for instance, I do a lot of recruitment with EPR architects who do their internal recruitment. And I thought experimenting with it, get the bio that see what what AI can generate. And it was very interesting, Jordan, like you say, it was a very convincing output, which the AI had.

However, there were wrong dates on for when, for instance, when it was founded, it wasn't quite accurate. And and that's partly [00:13:00] as well because AI is built, I'm guessing on all this information out there, isn't it? So some bits it's hard to say an accurate answer. I know that Google cuz Google have released their AI to go against.

Chat, G p t and there was a big hoo-ha Jordan because they put in an ad a question and the AI generated and it was wrong. It was the wrong answer.

Jordan Harris: It's like anything. The problem with AI and the reason it's dangerous is because okay. I actually probably should clarify something cuz a lo this is very common thing. AI really isn't a thing yet, even though it's a buzzword. True AI means it's co essentially conscious. It's able to understand context.

Very few are at that point, yet we have a few that are in development. That might be true. AI isn't really a thing yet. Chat DPT isn't a true ai. There are two main systems I used, which is machine learning and deep learning. Machine learning is what we've seen for the majority of the [00:14:00] last 10, 15 years When we've seen these clever automated tools that is very prescriptive.

Someone is literally sitting there telling it the right answer and telling it how manually adjusting it till it gets the correct answer. So things like fingerprint, sensors, fake recognition, most of that would've been. Machine learning, deep learning is the new one. That's also why we're seeing everything coming out all at the same time.

They're all suddenly coming out cause it's very recent technology and it's relying on reductions in cost of running the machines and also the advancements of technology in developing these out, the algorithms they actually use. So they all more or less run very similar algorithms and methods of working.

It's just slight various, it's between each company and each. Technology, but deep learning is different than machine learning because it is, no one is actually sitting there telling it how to solve the problem. You are giving it data. You're giving it general prompts to push it one way or the other, but you're not exactly [00:15:00] saying that is what you need to be doing.

You're roughly nudging it, and then it has to figure itself and learn by itself, which is very clever and very effective, but it also means it makes mistakes and you are not sitting there granularly changing it. So it can be much faster, but also much harder to control. And that's why it can make mistakes and that's why it can seem to do weird random things.

Stephen Drew: Very interesting. I I agree with you as well on the that AI tends to be a very marketing buzzword, doesn't it, Jordan? And I'm like I've you AI for a few years, especially when it comes to web design or mailing list. And you're like, is that really true ai? I'm not I'm not convinced.

What I was gonna say, Jordan, while we've been talking, Patrick actually says, I think it was through one of your earlier points, when he says, I think Jordan is correct, AI is just a newer, very good tool, better a tool. There is certainly an expertise in using any tool to the level of a craft person. And why should this be [00:16:00] any different?

So fair. Fair enough. Perhaps another tool in the toolbox. That's what I wanted to get your Pick your brains on this then Jordan, because AI is the buzzword. People are freaking out or they're loving it. And we've seen all these cool stories in the news. But you are doing this day-to-day job where you are focused on the systems.

You're focused on the digital tech within an Architecture practice. So the practice that you work in is a hail architects. I'll bring them up really briefly now, but what I would love to know is Okay. Rather than the AI per se, you might say, still say that AI's the most important thing this year, but what are you excited at the moment, or what do you think is gonna have the biggest impact on Architecture tech in the industry based on what at the moment

Jordan Harris: That's a good one. I think, I definitely think AI will have a big impact. In terms of new technologies? Ooh, that's a difficult one. I think Dig? Yes. I think Dig Digital Twins is gonna be a really big [00:17:00] development soon. I see that for those that don't know, by the way, digital twins is essentially the, it's an evolution of BIM, not quite, it's more of sort of a parallel technology.

Which I can explain a little bit more in a minute. But in essence the premise of it is that it does what BIM never quite managed to, which was it allows you to attach all your data effectively to a model and hand it over to a client in a form that's actually usable. The problem with BIM and the reason it never quite, I think it, it's never been as successful as it really should have been, is that, We spend all this time being a very clever model.

We stick all of our data in, it becomes parametric and it can, it sources everything. And it's essentially a database for the built environment for your building. And then you finish with it and you should hand it over to the client. What client understands Revit, what client can update that model and maintain it and actually utilize it in a meaningful manner.

So as a result, it becomes a dead end [00:18:00] product. Digital twin is intended to take that a little bit further. And becomes reactive. It gives accountability for your lights, your m and e, your all the products in the building, and it gives you the ability to say, oh, that door handle's fallen off.

Do we still have warranty to it? Who made it? Who supplied it? Where do we get replacement? Who do I contact for this? That's the theory behind it, at least.

Stephen Drew: It's I used to work a lot within real estate as well, so FM management, that's where this example that you are on about would be really useful, isn't it? The life cycle of. A building that's really fascinating. There's a little video here, so I'll bring it up if I can. Let's see. We can put it on the background.

I never heard of digital twins until we spoke before. I always got it confused, Jordan with twin motion. So that's my amateur hour. But there's two separate things, right? It's completely different.

Jordan Harris: Emotion's a good tool as well. To be fair.

Stephen Drew: That's more like visuals and where twin motion, it gets me interesting. [00:19:00] Cause I, I wrote my dissertation years upon it is the video game technology. So like years ago I used to do a blooming render in 3D max and I'd be there for six hours on the best settings. Whereas now you've got the feedback from scape and all that in Reddit.

Using the 3D game engines, maybe quickly in that foray, do you see that being used more and more in Architecture practices, or do you still think there's a resistance to use those kind of tech?

Jordan Harris: Yes I do. For me, and this is probably a very controversial for many people, I actually see it as the modern day evolution of the scale model in that a lot of people treat renders as an output product, that is for, given to your client. It's a pretty sales thing. Shouldn't be, it should be interactive.

You should be using it in the design process because you can't really sit and Revit and understand, how the light is affecting a space or how a facade is affecting the space. And Twin Motion is a great tool. Unreal. Engine five is also a really good one, which they're both built by the same [00:20:00] company.

I use this myself sometimes. And it's also nice it comes with Rabbit now, which is a nice freebie. But it's got a really nice it has the ability of actually sitting inside of a model, sitting inside of a building and testing different materials, different textures, different feelings.

In the same way as you would've done in the past with whole scale models, you can look inside of it, you can see how lighting affects it. You can see how slight material changes and slight textural changes will actually affect the lighting. And because it's so accurate, because it's so close to reality, you can actually be quite truthful there.

You can set it to be geo-located. You can set realistic suns, you can use realistic backgrounds and materials that are. Quite accurate in terms of real life materials. Going forward, there are you can use materials from a company called Rexel, which again ties into the whole unreal thing that Epic games have basically just gone and bought up everything in, in one of the industries at this point.

And those materials are scanned. They're hdr I, so they're very high resolution. I think they're like eight or 16 k, if I can remember right. Very high resolution [00:21:00] textures. They're literally scanned and they. Very accurately model texture and grain and reflections.

Stephen Drew: Wow. It's I'm amazed at how far it's going. The thing that comes to mind, Jordan, when you're speaking is that, I remember years ago and you laughed. So show my age here. But in 2015, I went to BIM Show Live, right? It was a big show and they used to have sometimes Architecture practices coming around being like, We were adapting BIM and then we used to get some practices perhaps challenging it, we are really good in cad.

You can't do these kind of buildings in BIM. And there would be resistance Jordan. And I wonder and I'd love to get your thoughts, Nick, cuz you're in an Architecture practice and typically you'll have people who will love. I'm embracing the tech, but equally you're gonna get some people which are a bit more hesitant.

What would you say to someone right now who's very hesitant towards the ai, hesitant towards or, real [00:22:00] time rendering and as you say, those scripting AI and all the stuff we talked about. Do you have any advice for someone hesitant?

Jordan Harris: Oh. At the end of the day, they are just talks. That is the base premise of it. So you are just adding. New tools to your tool set? I would say in terms of the automation and the AI and the visual side of things, you are, all you are really doing at that point is you are trying to use your knowledge and translate it into a more efficient means of creating.

Stephen Drew: Yeah.

Jordan Harris: At the end of the day, you are the person that's dictating the tool and controlling the tool to represent. Your understanding of the space, understanding of design, understanding of the regulations of the requirements of the clients. Doing that by hand might be nice. Doing it by CAD is nice.

Doing it by Revit or an AI is nice, but it's all choose the same output. It's just how you get to that output. If you, if I give you a complicated site with a strange envelope with lots of design conditions, maybe it has to be fully. [00:23:00] Part 3M compliant, and you have to go to access it fully with wheelchairs.

Maybe you need to have changing spaces, bathrooms, all of these things are quite difficult to design around. You can spend hours and hours and hours and hours trying to figure out the best layout. Or you can give an AI the rules, design the bits that are important to you, the spaces that are actually enjoyable or the spaces you want to design.

And then let the AI figure out, actually I can fit, this here and this there, and I can put a ramp along here. And that's where these tools I think are gonna be really useful is it's actually gonna give you more time and more excuse to actually do good quality buildings rather than constantly trying to chase your tail and be attempting to fix problems that are unnecessarily difficult.

Humans are not very good at solving sort of patent problems in terms of spatiality. We're very bad at that. It's why it's always been a skill, if I have to split a building that's complicated and figure out how I can put everything into it, That is a complicated skill. AI's very good at that.

They're very good at that sort of thing. And that gives you time and space and [00:24:00] excuse to sit there and go, actually I really want that facade to look this way. Or actually, I would want this really nice, glass facade and glass atrium with cantilever elements all way long.

It, you can actually afford to spend that time. So it's just giving architects, I think, a little more of the domain back. I actually think we are gonna have more control. Going forward because it's going to get, but it also depends if the industry adopts it or not. If we don't adopt it, what's gonna happen is we're gonna have another job merged my job.

Probably that's gonna be its own role that then takes even more of the architect's role like we've seen with, every other part of the architect's job over the last sort of a hundred odd years. But I also think it's a great opportunity if architects do embrace technology and they do start adopting these new systems.

It will give us more design control back and will give us more of the industry back.

Stephen Drew: Yeah I couldn't agree more. Even in my business, I'm always looking at optimizing stuff I hate. If there's an opportunity to avoid manual entry, then I'm up for it. I've got some horror stories. Jordan doing my part [00:25:00] one and part two where I was just in CADs. It was at mic station at the time.

Things are all on the same layer. You're just doing the stupidest, rubbish stuff, thinking what am I doing? Or counting areas in this crazy. Drawing late at night for a submission. And I think that you're right, there are all opportunities to change and the best bit about AI, in my opinion even now, is I try to get it to do the stuff which I find mundane, the stuff that is repetitious, and it allows you then to think about.

The key things which are really important. Designing that building, get the, what we really want to do is it frees us up to do. Then what I was gonna say to you is me and you both part twos, right? Unfortunately we're getting older. But there's a lot of students as well, and we focused on the ways you can use these tools within practice.

However, Jordan, if you were a student now, right? Say now someone's [00:26:00] watching this video and is thinking, this is really cool ladies for practice. I'm interested in that stuff. But if you were studying Architecture again, Is there anything that you'd keep an eye on now that's cool, stuff that you think would be good to look at?

Jordan Harris: If you want to do, if you want to prepare yourself for the ability to do ai and that's an area of interest, you grasshop and Dynamo are great if you just want to dip your toe in into automation for design, that's a good one. If you want to have a little more control if you wanna go a bit further and you really want to have some impressive control, Python, if you learn Python, isn't that tricky.

It's not any much harder than. It's not AutoCAD, it's just a command based scripting language. It's just understanding the basic premises of it. That really opens a lot of opportunities. Cause with Python you can then use things like TensorFlow, which is Google's sort of machine learning engine.

You can then, they're all, virtually all AI nowadays is re written in Python derivative packages. So the main one is SK [00:27:00] Learn TensorFlow spi And then things like pandas and NumPy are really good for sort of data and analytics and data processing. These are all really useful tools that allow you to build ai.

And if you learn Python and NumPy and Pandas and TensorFlow, you can sit there and you can build ai. It's in minutes. So that and yeah.

Stephen Drew: no carry on. Sorry. There was a bit of delay on my end there.

Jordan Harris: No. I was just thinking so in terms of. If that's where you're interested may be going, that would be a good system to learn. If you're more interested in the visual side, you want to be a bit creative, obviously unreal Engine five is fantastic. And from there you can also learn Python again, or c Python.

If you wanna learn coding within Architecture, Python is the only really good choice because Revit utilizes Python. So does Unreal Engine five. So both of those you can actually script into them to make them do things. So for instance dynamo actually is built on top of Python. So you can actually write Python scripts within Dynamo and save a lot of time and energy and make it do things it's not really supposed to do.[00:28:00]

Stephen Drew: Yeah, it's it's such an important skill, isn't it, Jordan? Because I remember even when. I was looking at about Dynamo and Ari and stuff. It was, I think it was iron, Python at the time. No, I'm really rusty. But programmer language I think is so useful cuz even like me, right? And I don't practice Architecture anymore, but with my website.

So I found that I've, there's a lot of PHP involved, CSS and bits of h t ml, but having even a basic understanding is really important. And we talked about in the past web design as well, and it, okay, it's another foray, but once you start to understand one, Coding language is a lot easier than to learn the others or similarities.

It's a bit like AutoCAD, MicroStation Vectorworks or now acad and Revit. I think it's a really important skill for anyone listening here to start experimenting in that. So it's just nice for to hear you confirm that as [00:29:00] well,

Jordan Harris: oh, absolutely. And everyone thinks of coding as being this really challenging thing to learn. It's not the easiest thing, don't get me wrong, but it's a lot more understanding how it works rather than.

Stephen Drew: Yeah.

Jordan Harris: A lot of it isn't just, oh, I have to remember all of these different commands and different systems.

A lot of it's more understanding how code interacts with a computer and also understanding what you want it to do. The actual coding side of things isn't that tricky because at the end of the day, there is a lot of documentation support, millions tutorials out there. You can go onto YouTube and learn everything you could ever want to learn in Python for no cost at all.

You just sit there and. Watch a tutorial or build something from there. It's not all that tricky. And it does it is more about control, learning it, learning how to code in terms of within Architecture lets you automate a lot of things, but it also gives you a lot more control over the tools you are using.

I can make edit, do things in 10 minutes. That would probably take someone doing the conventional click and [00:30:00] drag a day to do. Now I have to know what I'm doing. But it gives you that level. Finch is built just using dynamo Python. That's what that's built on.

Stephen Drew: Okay.

Jordan Harris: So that's the sort of thing you can do with it.

Stephen Drew: Yeah, it's another thing that I'd like to add is cuz I've always, when I was like I'm gonna learn programming, and you sit down and you've gone and got the command line or whatever, you think, wow, I really don't know where to start. And I always tell people, I, where I've done it.

Is that, so online with web design and stuff, you get snippets of code and I'm sure it's a similar thing where you've got like people's got some scripts which go into Dynamo or Grasshopper and you load it up and you can always backwards engineer it from there you go. Okay. So logically this must be doing this must be doing that.

So if by change these variables, then you get a little bit of a different output and then over time you get more and more confident that you start writing little bits. Yeah, so that's what I always advise people to do is look what other people have done [00:31:00] first and tweak it for yourself in that theme that we're talking about, like a tool.

You don't need to just learn, maybe Python per se, but if you've got a task and then you learn Python enough to get the output you want, then that's a success story,

Jordan Harris: oh, absolutely, and it's, it. Yes. It is a simple answer. Yes. Learning to code is a strange thing and it doesn't feel like something that maybe architects should be doing in terms of a classical definition of the industry. But the industry's changed a lot since those classical definitions were creative.

We now use 3D modeling softwares where it, in many ways closer to animation in many areas and the movie industry or, other, such sort of high tech industries than we were a hundred years ago where we were drawing everything and everything was hand done.

It is an evolving art form and we've evolved our skills over time and coding. It's not, I think, a prerequisite for Architecture. I don't think it ever will be, but I do think it's an important skill. In terms [00:32:00] of improving our control and improving our workflow and how we produce our outputs and how we design.

But again, like if you just want something simple to do, Google tutorial, how to do this, or how to build a website or how to even simple something simple learning how to run basic scripts on a computer is a really useful thing to do because you can use it to create new projects. You can use it to clean up your files.

You can just. There are lots of very simple things you can do. And odds are if you google approximately what you want, find what someone else has done and just modify it and you'll,

Stephen Drew: to go. Yeah. I agree. And the the last bit on, on this theme of script then that I'd like to add is Architects are typically in Excel and as well as BIM and Revit and I'm amazed in the theme of what we say, John the amount of manual stuff that I see going on where equally, if you spent the time just to get that one visual basic script or even the complete kind of complicated formula, the amount of time that you can save in [00:33:00] Excel is phenomenal.

And it's the same principle, isn't it? Whether it's Excel. Or Revit or whatever, you can optimize this stuff, but it's typically the learning curve to do it is difficult so that people feel like, oh, I can't bother. And actually, I think that's one of the deceptive things with Chat, chat G P T, is that I think people just put in a prompt and you get something that generates out and it's wow, there's no barrier to entry on it compared to the others.

But then, like you say, You can only get so much results from it. Whereas if you go deeper and deeper, you can, over time, you can start making these tools extremely beneficial to your output.

Jordan Harris: Yes. Yes. And, but it's also bear in mind that having understanding how things like Chuck g PT work even at a very low level and just what's it expecting, how is it spitting that output out means that you can actually then [00:34:00] use those tools more powerful if you understand. Why chat g p t spits out this answer.

You can then change your prompts to give it a more effective output. So even things like chat g t, while it feels like there's no learning curve, there is, it's just that it's spitting out the first answer. But that might not be the answer you want. You might want to change the answer, you might want it to give you a more accurate answer.

So there, there are still even there variables that you would need to know over time. I think that goes into tools like Chat GT and ai. Search engines effectively, I think, are gonna become a more useful tool in the future. I think it'll be a little while before it's in everything, but it will eventually happen.

We'd probably get to a point where Revit has the ability of specifying an area, giving it parameters and it will auto arrange it. I'm very sure that will be a feature that they roll out very soon because otherwise someone else will then, they'll have some interesting problems at that point.

But Yeah, a lot of it's just the learning curve, I think, for people. And, but it's also worth bearing in mind that everything has a learning curve. It's [00:35:00] just that you've already put your time and engine to learning your current system and there will always be a reduction of productivity.

When you learn something new, there's always gonna be that annoying time that you feel like it's really hard and you can't get your head around it, and it takes your time to get up to the speed. But when you're up to speed, when you know if you commit to the system and it does work, and when you get to that point, it's fantastic.

Because then you can actually utilize it. And a lot of people, I think are quite, stubborns are bad words, but they drag their heels. They're comfortable. They've got this little container that they use that works well enough for them, and they don't see the point in investing time and actually going down first to go up again.

Stephen Drew: Yeah. I agree. And that's why I just wanted to add it. It's not even just for instance, while your role is fascinating, because like you said you teased at the start that you made your own job title, but it's a newer space. That being said though, if ever gonna have a little bit of this attitude in their jobs, then hopefully they.

This is not [00:36:00] real opportunity to remove the mundane, the o. The last point I was gonna talk about, so is last year around October mid journey came out ai, which is AI pictures and there's a few others. The open AI have their own one as well. But remember these videos came out and everyone was freaking out.

Oh my gosh, they're gonna replace the architects, yada yada, yada. And I think it's all gone quiet. And that's how I feel like it was the fad. And listen. Okay. I understand if you are an Illustrator and you spare time and you spend these beautiful you take hours doing these portraits or whatever, and this AI does it really quickly.

But in Architecture, apart from maybe a mood board or a feeling, I haven't really seen anyone use. These mid journey images in, in, in a project apart from one person, which blurred mid journey [00:37:00] with traditional methods to make some art. Have you seen any ripples in this space using it, or do you think it's been a bit of a fad?

Jordan Harris: I the problem with Mid Journey, I haven't actually seen anyone that's used it in a kind of, Real world concepts. I've seen a few people play with it. But the problem with something like Mid Journey is it's control. Mid journey again, is it's a recurring neural network. So it's, it can only take from the past.

It can't make things that are new. So you are, it's fully relying on whatever it's been fed beforehand. So you, it's great if you want to create something that's abstract or conceptual. So if you were gonna go for maybe a very stylized image, that could be where it would be quite useful. I see it being far more useful as an accent, potentially taking an existing building or a sketch and then elevating it with new visuals or conceptuality or allowing us to put layers on top.

There are some really interesting use cases for it. For instance, we could use it for [00:38:00] replacing superimposing elements of a building. The problem is it's just not efficient. And what's the difference between something like Mid journey and a render? At the end of the day, it's not real. You are just sticking textures.

You click render, it's done. You could make the same argument versus previously when people would hand draw and hand color and hand render it. It's just a slight evolution of the technology. We will definitely hit a point where, again, that becomes more useful whereby it's built into a render engine and maybe rather than having.

Pre-made textures, you can just specify what you want and it can generate texture, it can generate background. The problem with technologies like this is that they're more showpieces. They're they release these kind of products not to make money. They're trying to actually get hype because then they can go to, A big investment firm, something like SoftBank, and get a ton of money or get bought up by a big tech company.

That's more what they're actually trying to do. They're not really trying to, the great example is look at Chat g PT was rolled out and [00:39:00] now it's being rolled into Bing. So that's going to be used for Bing for their search search engine now. So the reason they're big deals initially is because they're meant, and they're portrayed as a toy, something to play with, something that can do something unique.

In reality there, there is a business intention behind it. They will be roll out something business oriented in the future. I suspect something like Mid Journey will probably see its way into, things like Photoshop and rendering engines and other such tools like that. It won't be a standalone tool and it isn't really useful as a standalone tool.

It's, if you're an artist, maybe you can argue it's somewhat useful, but again, it's only taking existing images and tweaking them. And it's the same as, it the same. It's pretty much limited to a color by dots. Effectively, if I gave you a color by dots and I said making a picture, you can make a very nice picture, but it still wouldn't be original.

You can change the colors, you, maybe you can ignore some of the dots and try and draw your own stuff, but at the end of the day, still based on that original image, and you are limited by that [00:40:00] factor.

Stephen Drew: Yeah, said, really well said. I for example I've paid for mid journey a few times in the novelty. Wor wore off quick, where it was good though for me. With stuff like copyright infringements. So say, now I want to make a thumbnail of a building which doesn't exist on a website, that's great because then I won't get sued, which is nice because if you take images then you know, pe you gotta pay for that.

And that's fine, and I agree with it. However, that's a very niche purpose, isn't it, Jordan? I'm not building a building from it. I can't really claim and what I mean is in a thumbnail. It's like an impression. It's quick. That's it. There's not much more substance apart from that. So that's been my experience on there.

I was gonna Ask you one last thing and the audience, if you've got any questions. For Jordan, now is the time cuz he has gotta go back to work before there the BIM models crash and all that stuff. I don't want to be responsible, although I think you've got a well [00:41:00] over machine. I don't want to get you in trouble or anything.

What I was gonna say, Jordan, is last year, depending on your passion, Pete, I was, did a lot of talk about the Metaverse. Okay. Some people loved it. I think I drove a lot of people insane talking about it. They're like, ah, one more talk about the Metaverse. But I think it's gone a little bit quiet this year on the Metaverse.

There's still stuff happening and you said it quite rightly, there is a business case for it. But that being said, where is the Metaverse going?

Jordan Harris: Oh

Stephen Drew: I have a big question mark on it. What's your views on the Metaverse Jordan?

Jordan Harris: Based on current funding for Facebook and where it's going. I think it's, I don't think it's gonna just completely disappear, but Facebook are already looking like they're gonna pull the plug on it a little bit. The whole, the whole premise of the Metaverse was the reason why Facebook's invested so much in it is because Facebook is facing a reality where they make most of their money out of advertising and they can't control the current internet.

It's [00:42:00] mostly controlled by companies like Google. And they don't like that they have very limited control. Apple, for instance, released privacy update. Basically it removed all of their funding from Apple devices overnight, which is 50% of phones in America, for instance, I think more actually at this point.

So they basically wanted to create their own controlled ecosystem that they had full control over. That's primarily why they've invested so heavily in it. The problem with the Metaverse is that it's. It's very over controlled by Facebook. They don't really want anyone else having too much control with it.

The other problem is that it's basically relying on technologies like VR headsets and systems like Oculus to actually be able to use it. So the barrier for entry is really high. So I think that would have to come down massively. And the other problem is it is just. Technically speaking, really struggling.

It's very heavily reliant on these very sophisticated [00:43:00] 3D augmentation devices. Things like glasses. Those are very expensive and less everyone has it. It's just not accessible. It's the same reasons why anyone adopts any technology. The barrier for hand entry has to be low enough that the average person to do it.

Look at laser disc technologies like that were arguably better, but, Never takes off because it's the barrier. French is too expensive, too high. They look like they're gonna be spinning towards more business oriented models. Soon they have some interesting ideas of a virtual office. I dunno how I feel about that.

It's

Stephen Drew: Yeah,

Jordan Harris: quite clever. But the other argument is, would you wanna sit in a virtual office all day, every day with a headset on? I think the average person would say no. I could see for Architecture there could be some very good uses cuz you could actually have a real built, could have an actual building that you can walk around, look at and understand.

And I've seen some, they're not truly Metaverse, but I've seen some really interesting ideas of how. [00:44:00] Buildings could actually be augmented by ai, for instance. We sounds quite scifi, but we are actually having the idea of a hollow deck is becoming quite close to reality at this point. With the idea being that you can have spaces that are accentuated by ar.

There are people now who are developing kinesthetic suits, kinesthetic gloves, where if you touch something, it puts pressure back so you can actually feel it. So the, a lot of it's tricking the senses in with these new technologies. The big limitation with the Metaverse is it's basically just one company trying to take the whole industry when reality, all of these other very clever technologies they've been there around there for a while and they do their own thing and Facebook's just trying to essentially they are very heavily invested in, they want to be.

The defaults, they want to be the Google of the web three effectively. So it's very difficult to say when it'll be, because it depends on how you define it. If we're talking about Facebook's [00:45:00] Metaverse, I don't think that's gonna have a very good lifespan. It's just not, doesn't work.

But if we're talking about the Metaverse as a concept, cuz they are actually different things, I think that will definitely take off. But I think it's going to be a few years off before we actually start seeing it because it's going to require things like, VR to come down and price even more augmented reality to build more mainstream.

But there are some really clever bits of technology. We're already starting to see if you have an iPhone and you use Apple Maps, the three people that do that you actually, they actually have some very clever AI in there. So you can actually hold it up against a street and it will in 3D project signs on the street to tell you where you're going.

Stephen Drew: Yeah, said. I I completely agree with you as someone that owns a HTC Vive for the first generation, I absolutely love it, but it's not practical for everyone. It's not the kind of thing you can whack up in, in, in an office or transport that easy. Possible, not impossible, but as you say, the more portable [00:46:00] the technology is, the kind of the trade off in the immersion, is that, that, that's my feeling.

It, and like you say, It's a very niche case. Use case point or whatever, because really we're all using screens and stuff. I agree with you. I was gonna say though, Jordan, I've been asking you all the questions now. Feel free to fire a few up me, if you have any. I haven't practiced Architecture for a while.

But what I do, I guess is I recruit. With Architecture practices. So I'm aware of the trends of current employers with tax and stuff and what has filtered down the trough, what hasn't. Do you have any questions for me at all?

Jordan Harris: Ooh, that's difficult. Thought that beforehand here's a good one for you, as someone that sort of does work in the recruitment industry, what sort of technologies are you seeing employers ask for at the moment? Mostly.

Stephen Drew: Good question and always changing. That being said though, so 10 years [00:47:00] ago, Revit was a novelty that not many people understand it or it was. On the wishlist to get just had to cut off a phone call. Sorry, I should have turned that, should have turned that off before it. So Revit before was like, oh, that would be a nice to have.

That's simple. Let's see if we can have someone with that skillset or we're not really there yet. MicroStation is still the main one, but now Revit is the defacto. And for anyone that's resisted join Learning BIM up until now, then. You need Revit. There is no be, there is no maybe unless you are a small Architecture practice.

They, that's perhaps in rural England and they're happy with that, fine. But if you want to go for any of the big players rabbit, it is the defacto. And the other one that I see more and more popular is Rhino is always been a nice to have in the larger Architecture practices, especially if they do design.

But [00:48:00] now I'm actually seeing Grasshopper be a desirable skillset. And before it was like, No one even thought about it, but now it the magical combo is rabbit rhino and grasshopper for a mainstream Architecture practice. A large Architecture practice in the u in London for a medium-sized Architecture practice.

I reckon that a lot of them are probably looking for Rev, and if you could bring a bit of parametric design or whatever, that's cool. Because actually you've got a lot of skillset and more than anyone actually in a smaller office you fast. If you're the tech guy, if you understand BIM a bit more, you become that BIM champion.

And

Jordan Harris: Yeah.

Stephen Drew: Yeah, I think that, those are the skillset that's popular. I never, as a recruitment consultancy get a brief from a client anymore, which says AutoCAD. I have one client. I know that he still uses Vectorworks, but they're moving to Revit. So it's if the [00:49:00] person has Revit, that's important, but then if they understand Vectorworks, then that's useful too.

But that's a minority. And I do think though, if you don't understand Revit, it can hinder your career. Even if, I know some practices should train people up, but I think employers generally are gonna look for some. BIM experience. You don't need to be the BIM guru, but some initiative to learn it.

And I do think though, that be, it's a old conversation I've had for the last 10 years about should I learn BIM? And now it's not even a should. It's you must learn Rev,

Jordan Harris: I was gonna say it is probably worth for anyone that's watching that is trying to learn or wants to learn Revit. Revit is. It feels daunting, but you also don't need to know all of it. Revit is a massive bit of software. It goes off into every direction and it starts to feed into other softwares like Navis works very quickly.

The core premise of Revit learn how to build a family properly. That's what [00:50:00] everyone gets lazy and doesn't want to do. The basic, tools like walls, floors, doors, they're fine. And also learn how to locate your pro projects properly. If you can do those basic things and obviously put it on a sheet, you are probably all right.

For most basic BIM stuff. The more advanced things like schedules and making families parametric, they're great. But they are where it, it gets a lot more complicated and. Your mistakes will have a lot more of an impact. So if you can get those basic premises, learn how to build a wall properly, learn how to build a floor properly, learn how to do staircases reasonably well.

Custom staircases are nice to learn that I most tutorials recommend learning those quite late in the day. I'd say it's better to learn them sooner than later cuz staircases are something you're gonna use all the time and they are generally a more complicated element. Anyway, if you can deal with those basic things and learn how to set the basic project up, you are fine.

And put the sort of more parametric side, the more sort of advanced settings within Revit to the background, [00:51:00] deal with those later. Cuz that's also where it starts to come to a BIM Manager anyway. So you may not even need those in many practices. It depends on the practice. Some practices do like their architects in their sort of part twos, part ones to know a little bit about that.

Some don't. It seems to vary a little bit.

Stephen Drew: Yeah, you are spot on. Jordan we've blasted through, I dunno where time's gone. It's been an hour

Jordan Harris: hour.

Stephen Drew: Yeah. Super impressive. So I'm gonna give you a massive round of applause cause you've been an absolute star, A natural, it's been like autopilot for me because it's your your expertise is so interesting.

But it, it's such a vast subject, so it's hard to cover everything. I'm sure we can. Return to a few topics in the

Jordan Harris: Yeah.

Stephen Drew: However, if someone wants to get in contact with you now, Jordan and all that stuff, where can they find you? What practice you work at, and where are you online?

Jordan Harris: Oh. Probably best at the moment is hair Architecture, which is my. [00:52:00] Email for the practice is j harris hail adm.com. And I'm on the website as well. That's the best for the moment. If you want, send me any directly my email is Mr. Jordan l harris.com, which is also my website. Don't go to it right now.

It's not currently Very nice. Working. Learning to code break things all the time, mucking around with it the last few days. Probably be nice in about a week or so when I can be asked to re-upload the actual proper website. That's Hey ho. But other than that yeah. Any questions about any of the many varying topics I've talked today, I'm more than happy to chat about.

I'm, I could talk about any one of these topics in several hours. I'm very sure, so I will. Avoid that. Stay to one hour, I think is best. But yeah. Have you got any more questions for me or any

Stephen Drew: No I think that was it. We had we had the passionate Patrick Wilson that popped in, but I'm thrilled with that. I think that you've done an amazing job of giving, like an introduction. I. On the, on this moment in time, who knows? Jordan, maybe in a few [00:53:00] months we'll look back and things are changed already because the space is moving.

But that's part of it. So I'll give you the last round of applause. You've been an absolute gentleman. I thank you so much on that note. I'm gonna end the live stream a second. Jordan, you've been an absolute pleasure. If you want to get in touch with Jordan, you can find him on LinkedIn or his website.

Jordan's been very modest. It's a decent website, and it'll be all new very soon. Or you can contact him at Hale. Hail adm.com. So hail architects. Very cool stuff as well. On that note, I'm gonna end it now. Thank you so much for being here during your lunch break. We've got more shows coming up. Jordan, stay on the stage while I end the live stream and for you and the audience.

I will see you soon. Take care. Alright, bye-bye.