From Measurement to Model: The Evolution of Scan to BIM

Foreign.

Welcome to another episode of Design Under Influence.

I forgot my own podcast name for a second.

No, I did a long week.

No, I didn't do it.

Restart.

It's dui, which stands for not what you think or what.

It's a pun.

Right.

Design under influence.

Because if you're not learning and growing, it's always, almost always like Designing under influence, which is, you know, it's not ideal.

So we're here to help you learn and grow, answer some of your burning questions.

And with me, my culprits, my co conspirators, Megan and Liz.

Good morning, good afternoon, Hello.

Good evening.

Hello.

Depending on when you're watching, you know, we want to be inclusive.

Good night always.

So welcome to.

It's always great to see you.

Every Thursday we meet and we talk about all these interesting things.

I learned a lot and hopefully our audience does as well.

Today we're going to cover scan to bim.

I am somewhat familiar with that technology.

I'm curious to explore it a little bit and maybe you ladies can tell us what it is and what is not.

But before we go there, let me just.

I think we kind of fail to introduce your company, what you do initially and your expertise level in every show, assuming that everybody watches everything, which is not necessarily true.

True.

Right.

Audience, whoever's listening, maybe it's the first one.

So you guys, can you briefly kind of say where, where you work and how you plan to dominate the world?

I'll let Megan take this to start.

Well, essentially, we help architects and engineers with their technology.

Basically all technologies related to their drawing, production.

And with that, there's cloud based software that we help them with.

We help them with standards and templates and converting from old systems to new systems.

Basically, anything that we can figure out, we help them do.

Yeah.

AKA BIM Consultants.

BIM Consultants.

Yeah.

Searchable term is Aurora BIM on this side?

Aurora bim, yes, exactly.

You guessed where it is, actually.

Yes.

And Aurora is Aurora Borealis.

So many colors, many facets to what we can do.

Very fascinating.

And we are not as fascinating, but we are mission critical.

We do it for architecture, design, engineering firms.

We go deep on this industry.

So we'd like to solve problems before they become problems or create issues before they become problems.

And that, I think, requires deep knowledge of the industry.

And that's why we're always learning, always growing.

And we work together with Aurora BIM with a lot of clients.

So anyway, it's a happy connection.

But let's go, let's dive back into the subject so, Liz, you were starting talking about how you use the scan to bim, or I guess maybe a version of it.

Let's kind of introduce us to the whole concept.

Yeah.

So I think part of what makes us, all three of us successful at what we do is we're super nerdy and curious.

So.

So we always want to try out the newest technologies, and why not try them out on our friends and family?

And so I am married to a fantastic, very intelligent man, but he cannot read a drawing to save his life.

So when we go through our various buying houses and renovating and adjusting things and doing stuff together, I've had to learn different ways to communicate with him about what my vision is.

And it started out with, well, I can make models in 3D and I can spin them around.

And that wasn't quite good enough.

So then I also moved on to, like, well, I could actually produce these kind of VR images, so you can stand in the room and look at it.

And so he's been my guinea pig along the way, but he's also super into technology.

He's in the IT world, so he loves also testing things out.

And when we bought the house we just moved into, he downloaded an app to his.

His iPhone, and he actually walked around and scanned the whole place into his phone, created a file, and then asked me, what do I do with this?

And he was able to create.

Someone who isn't an architect, doesn't know how to read drawings, was able to create not only floor plans, but a 3D model he could send to me that then I was able to use in revit to as the basis for playing around with, what can we do with this property?

Do we want to buy it?

How big is it?

You know, what do we think about it?

And that's a great entry level for just about anyone to understand what we're doing and what this technology can provide.

And that's been interesting to see.

The evolution I was mentioning before from super big, heavy, clunky survey equipment that has to be precisely calibrated in order to get any data out of it to, well, I'll just download an app on my iPhone.

So that's the intro to what scan to BIM could be for some people.

Gotcha.

But that's more on a consumer side.

And, Liz, you're not going to use these drawings for remodeling, right?

No, I'll completely redo them, but it's a good start.

Yeah, yeah, no, I. I get it 100%.

So, like, people who hasn't really faced that technology.

I think that's a great intro and great consumer application, you know.

But as far as going a little bit more next level to pro Megan, where.

Where is it at now?

I mean seems like AI is here and is.

Can you give us, can you open up, give us a little bit more on how people are using it in more professional settings like to actual design stuff.

Yeah.

And I think what I'll do is kind of break down the.

The three parts of the phrase, right.

The scan and the TO and the BIM which I think are all equally important.

So the scan like Liz mentioned, you can now use your phone.

There is survey equipment that is expensive and more highly precise.

There are now which has been a development in the last like 10, 15 years.

There are these middle of the road scanning devices that now architecture firms, they don't have to be huge, but they can own a scanner and they can go in and actually do as builting and in a way that is a little, a lot more efficient than going out there with a tape measure and taking pictures.

So it has kind of in a way revolutionized the way we do as built drawings just in that little step that it's taken over with going out on site with your camera and your, your ruler and your pencil.

So that so scanning.

There's many ways to do it.

You end up with a point cloud in most cases from most of these devices, which is a super, super dense amount of data that is basically a laser that's shot out at the walls and the floors and the ceilings and all the things in a room and it comes back and carries all that data with it.

So it's little points all over that represent your entire building.

Those points can carry data in terms of color and you can overlay images on top of them.

And in order to look at some of these point cloud projects as like a walkthrough, kind of like you're similar, you're used to seeing with these Matterport or iPhone applications where you can walk through them.

It overlays the images that the camera's taking on top of the point cloud and it makes it easy to really view what the data is that you're getting.

So that's kind of the scan part.

There's like all of these pieces to it and then the two which is getting those points into a usable format in order to create a BIM model is the part that everyone skips over.

They think it's a one click thing that you can do where you take this group of points and Say recognize everything that's in this point cloud and make me a model.

That's what it is.

Not yet.

Because things as you know in real life are not completely flush.

A wall is not completely vertical, A floor is not completely flat.

But when we do a model of it, that's actually what we represent.

Most of those elements with is families in Revit that are simplified for design.

So the actual process of the two involves taking these point clouds and overlaying them in a model and actually having a technician go in and model beams and floors and walls within this shell.

And that takes a long time.

So you're slicing and dicing this point cloud and you're looking at outlines and you're looking at the whole building like this, like taking like thin slices through it and modeling all these elements manually in most cases.

And because of our inability to have X ray vision into inside of walls, you don't actually know what's going on in there either.

Or below fireproofing or you know, into areas where your skin can't reach.

There's.

It's got a long ways to go for that.

Like two to actually be this click of a button.

But I'll get to this later.

But I think AI is going to help us get there or at least get closer.

Yeah, that was really long.

And bim, you forgot you didn't say.

Anything about the BIM part is the modeling part.

So that's the part I was mentioning in the end there.

So it's the two is the like actual mechanics of creating everything in bim.

But the BIM final product would be the whole model being built.

So like you have your point cloud and then the two is the work that it takes to get this model in the end where normally we would just build that from scratch.

Right.

Or we would build it based on a set of as built AutoCAD drawings or pictures and measurements from sight.

Great, that's very detailed.

I really appreciate it.

I did not know about the point cloud.

I thought it uses just like Tesla, it uses camera to represent reality and then it sort of.

It like vectorizes everything.

Right.

And that's actually interesting versus lidar, which is.

LIDAR does use ray.

I think that's what.

That's what the same thing, isn't it?

The similar technology to LiDAR.

Do you know?

Yeah, yeah.

I think LIDAR can read through more stuff.

I know that they do for like infrastructure as built.

They'll send airplanes out with LIDAR scanners to get the topography and it can read through the trees.

But it will also get the trees.

So that can be quite important too.

But that won't tell you what's under the ground or what the ground's made of.

It'll just tell you where the ground starts.

I wanted to sort of take, take a step, elevate for a second and say, okay, what is the advantage?

First of all, who are the main people who will benefit from this, who are benefiting from this technology and how the advances in this technology will improve their lives?

In other words, give me a little bit on who's right now are so deep in this industry, who is it benefiting right now and how much time they're saving versus the old school way.

So I have a little story and when one of the projects I worked on at the hospital where they were renovating an old historical building, and I would say historically, not to reuse that word, historical preservation has been probably the place where a lot of this got pushed forward, especially in Europe, because there's a lot of rules about specific buildings that you can and cannot take touch or do things to.

And so getting those scans have been.

I've.

That's where a lot of, at least in my experience, the architecture world has started entering this.

But in this case it was actually because they're going to convert an old building from 18 something into an administrative building on a hospital campus.

So they needed to scan the whole thing.

And the roof was a very, it was a pitched roof and very tight.

And we're going to put a whole bunch of modern mechanical equipment up in there amongst all of the 200 year old beams and chimneys and nonsense.

And all the drawings, as you can guess, were hand drawings converted to cad now converted to maybe revit.

And they had no basis in reality at all.

So a lot of the coordination meetings were actually everybody's drawings laid together plus this point cloud to try to fit all the pipes and pieces in to the space that was there without actually climbing up into the attic, all 20 of us and trying to look with a flashlight to see can we fit this or not.

So I would say project teams on as built projects or renovation projects benefit a lot and save a lot of money and heartache when they can play with these models.

But then I would bring up what Megan said, which is the two part, because you can scan a point cloud, but unless you have a professional BIM person or someone who knows how to handle them, help integrate it into everybody's model, it's kind of useless.

And actually when I started in that project, no one was using the point cloud.

So I had to.

I was like, what about this, guys?

Oh, yeah, that could be useful to take a look at.

And I know Megan's been struggling with that too.

But to get everyone to use the data and be able to manipulate it and rotate it and look at it and understand what they're seeing is then, to answer your question, who does it benefit then?

It benefits everybody.

Who does.

This person.

What sort of.

Is this the.

The BIM consultant's job to interpret point clouds and add them to the models.

Who does this?

There's actually companies.

Yeah.

And there are companies that like, went out and bought scanners.

And this is all they do.

They do the scan and they'll scan and then they will create a revit model for you.

But that's typically like the actual, like, scan data.

And I'm getting off topic here, but jumping ahead, but the scan data is so heavy that you have to be set up to deal with that data.

You can't just hand it over to an architect or an engineer like Liz was mentioning.

Like, it's.

You have to have the right computers to set up to do it.

They have to be the solid state drives in order to work with those.

You cannot work with it off of a cloud.

It's gotta be downloaded locally in order to load it into revit.

So that requires that you have that much free space on your local machine to work with these files, which is a limiting factor for a lot of teams.

You just don't have hard drives that can fit these point clouds or multiple versions of them in order to work with the files.

So that's why the specialized companies exist essentially, to be able to scan and interpret and then just hand you the ready file.

You can work on that.

Exactly.

The viewers are great online.

You can look at the data and you can view it online without having to download it locally.

But in order to actually bring it into a revit model, turn it into a BIM model, you need to have the space locally.

So before this scan to bim, how was it done?

The way I mentioned before with your.

You will take pictures on site and you are looking at that and you measure and you're looking at maybe old drawings, the AS built drawings from the city archives.

You're looking at maybe a scan, maybe a survey company went out and did a survey and I gave you a cad, a set of CAD drawings of the building that I was built.

But then you're extruding all of that and you're creating the model from those.

So it sounds like it was a lot more labor intensive and less.

Maybe even less exact.

Yeah, a lot more guesswork.

Yeah, a lot more guesswork and a lot more trips back to site to.

Check one more thing.

You know you forgot to take one measurement, right.

With these scans too, you can measure in the viewer.

So if you forgot to take a measurement, you just go and use your measure tool and it's all there.

Exactly.

Okay, so there's still an.

I mean, you're saying that we're going to cover it later, but maybe later is now.

I'm actually infinitely curious.

How do you then if there's one problem choke point right now, and that is converting that point cloud into the model.

Why is it so complicated?

Is it because heavy.

The data is just heavy.

A lot of data.

Let's say you're going to create the structural model.

We care about the columns and the beams, the stuff that's sometimes under the surface.

The points don't capture that, but they capture the shell around it.

And we can manually tell it what's inside.

But I do see.

So right now, like it's manual.

You have to look at the point cloud and then you're placing the element inside of it.

You can take the whole point cloud and make a mesh which is like just a thin surface that connects all the dots and that is used.

And if you want to call that bim, some people do, because then you can bring that into a model and you can turn it on and you can see this is where the as built outline is.

But it's not straight and it's not what we typically see in like a design model.

Right.

You'd see your walls kind of jitter a little bit.

The point cloud is not straight because real life is not straight.

Right.

You've got finishes and your walls are.

And you have a lot of junk sometimes in the room.

There's like chairs and tables and somebody's jackets hanging up and.

And the person taking the scan.

The person taking the scan sometimes in there.

Yeah.

Cobwebs.

Everything.

Yeah.

So I forget what your question was.

My question is converting that into a model.

So I guess, yeah, there's two pieces.

There's another piece to the question, the piece of unknown, you said obviously it's not scanning through walls.

So you're still getting that information from other drawings, other surveys, whatever.

Right.

You're still laying in that over that point cloud or into the point cloud.

Yeah.

I mean, how much faster is this versus the old method?

It sounds like it's a little bit faster, but Is it like a lot faster?

And what can make it a lot faster?

It's more accurate.

More accurate?

Yeah.

I would say the collection of the data is faster than like Megan was saying, going out and measuring and writing.

There's less of a chance of error because someone doesn't write as six instead of a nine.

Or it's too sloppy to read, or you missed something.

There's also the fact that everything is scanned with a picture.

So you don't have that.

Oh, we didn't get a photo of that thing because everything is photographed.

So you get all the data.

So you.

So it's funny because it's like on one hand you have now collected all the data in a very accurate way and now you have too much data.

Whereas before you just would have guessed about stuff, but now you have the actual answer.

Yeah, and I'll say too, like, so when you're saying, like, yeah, we're going to overlay what we know is on the inside from drawings.

Those drawings oftentimes don't exist.

Or they're wrong.

Yeah, or they're wrong.

There's been 10 improvements or renovations.

So you use your best guess.

And sometimes they will even then chip away at the finishes and scan when it's open so that you can see.

Or they'll place the scanner up in the ceiling so that you can get what's underneath.

Yeah, yeah.

So it really is.

It's kind of putting all the different pieces together.

But here's what I.

This is here.

My dream for it would be, okay, AI, hope you're listening.

We input all the stuff we know.

We input these existing drawings.

We just upload them the input.

Then it can ask us questions like, is this a steel framed building?

Is this a concrete building?

Like, we answer basic questions about what the building is and that it can recognize the clusters of points and build the.

The model based on that.

So, like it can recognize that the cluster points.

It looks like a wide flange beam really is a wide flange beam.

And we just tell it how much fireproofing was on it and it back calculates what the size of that beam probably was.

That's where I see it going.

Because I think AI was really like, I think BIM needed AI.

We just always had so much data and now we have a way to sift through it and organize it and use it.

And especially with point clouds.

Well, I would take that dream further.

And I think this is what people are also doing in buildings in another project.

And some companies we work with is that they Will you have your design model?

You give it to the construction team, they go out, they build it, and at specific points along the way, it is being scanned.

Before you put up the drywall, before you put the fireproofing on, you scan in what's actually being built.

That gets compared back to the design model.

The design model gets adjusted so it is precise to what was actually built out on the field.

And then when everything is.

All the dirt is laid and all the drywalls up and everything is done, you have an actual BIM model that represents the actual conditions of that building.

Yeah.

Wow.

And further.

Oh, go ahead.

For planet Earth.

You mean this?

The planet Earth?

Yeah, I mean, there are cities in Europe that are doing these, like, smart cities.

They are actually requiring these as built models to be delivered to them by every single construction company and architecture firm.

And they build their city so that they can use different systems to calculate water usage or wind or solar, whatever they want.

That's so super smart.

And so I. I don't know.

I don't know how much improvement we made.

Tons of improvements.

Like in anything in specific.

You know, literally anything.

Like, hey, yeah, you know, how much snow can this roof hold?

How often do we need to do snow cleanup?

I mean, anything you can.

Absolutely.

I think that's another podcast question.

Is bim.

It's like giss, what is gifts?

What is gis?

It's bim's cousin on a larger scale.

And that, you know, all these things we do now can feed into that.

And then using.

We probably have to find a better expert.

Cause I only kind of know a little bit about.

Yes, but yeah, I mean, this also feeds into that bigger picture.

Is we gotta look.

I'm gonna leo Information spatialization or something like that.

System.

Yeah, systems.

Oh, wow.

Okay.

All right.

We were off planet.

Let's go back.

So in Europe, there are some cities, smart cities.

Yeah, I believe that 100.

I don't know if anything's happening here in America yet.

I think there's some hospital campuses and larger campuses who do that because it's in their best interest to know what all their buildings are doing.

Yeah.

And it's there this digital twin that was like.

That's a whole other topic.

So popular.

That was like a buzzword, like, what, three years ago?

And then AI now.

And now it's moved on to what?

I don't like digital twin.

That's a whole other topic we could talk about, but we could call.

It's Smart facility management is a better way to call that.

Yeah.

Yeah.

All right.

Are we straying from or.

Yeah, we did, Sorry.

It branches out into lots of different areas.

I would say it is.

It is fascinating.

I agree.

We can explore this some more, but if we stay to scan to BIM right now, what are some of the misconceptions that you wanted to dispel with this particular show?

Like what questions did you want us to answer?

And maybe let's answer them.

I.

One thing I would say is that it isn't only for large projects that you can any size project, any person, whether you are using an architect or you're just a person, can use it.

Now the other side of that is the cheaper it is, the less accurate, but it's still useful.

So the, the level of accuracy you need is going to be dependent on what are you doing with this data.

And that's always the question in BIM that we try to ask is yes, we can take this out to LOD 4 million, but what are we doing with this data?

What, how detailed does this data need to be?

And then that that informs how you collect it and what you do it and how organize it.

So I would say it doesn't have to be a big project and it doesn't have to be expensive.

And then on the other side of that, the myth I would try to dispel is that it shouldn't always be super accurate because you can have some of these companies who do scan to BIM and they will draw a building for you, but they will draw every single turn and edge of the walls.

And that is so not useful.

Like Megan was saying, we just want, just draw a straight wall.

It's.

Yeah, because in design that's all we need.

That's all we want.

Right.

We're not creating fabrication level drawings.

We're creating the design drawings.

Yeah, Fabrication is a whole other story.

Yeah, go ahead.

Whole other story.

No, go.

You want to explore that for a sec?

Oh, yeah.

We actually, I was just talking to one of our, our partners this morning and they do scanning in facilities where when they pour the concrete and they've got the concrete pad laid for these really, really large basically like warehouses and they need to come in and create all of the like frames for the shelving that's going to go in this warehouse.

They go in and scan the variation of the levels of the concrete and the difference in the height to the underside of the ceiling beams so that they can cut the columns that are going to go in there.

Exactly.

Those are non structural columns.

Obviously they're for these for these storage shelving and everything that's going to go in there.

So but each column is slightly different.

So they go in and they scan it and they.

That's how tight the tolerances are, is they can tell them exactly how long each one of these steel columns needs to be to come in and fit, fit out this entire warehouse.

So that I thought that's just like super fascinating.

And fabrication level scanning is like a whole other ball game like completely other side of the spectrum from Matterport or your iPhone.

But they are relying on that more and more than just like the old kind of traditional surveying techniques where you're out there with a total station and you still need that in order to set certain.

Basically when you go to multi story buildings it's really difficult to get point clouds because you're standing in this level scanning and then you're standing in this level scanning.

It's really difficult to connect the two with the outside coordinates.

Right.

So that's where they still use some of the older style survey tools to connect that data together in the most accurate way.

Yeah.

And so what you're saying is with AI advancements this is this step, the two step should start smoothing out and costing less and less and less.

That will actually then give access to this technology or just about any size project.

Everything could be as accurate, you know, with you know, with, with waviness or whatever or as you know, I guess as straight as you want it to be.

Yeah, you like ideally with a plug in the LOD you want and it will give it to.

Yeah, exactly that.

So that would be awesome.

How far away is that Megan?

I mean I've already seen with the lidar scanning softwares that will already try to detect okay, those are trees, this is a road.

And it does pretty well.

So I mean it's on the way.

I think it's like any other AI.

It needs people to feed it and work with it.

And I got really excited about your little dream there, Megan.

I imagine like you dump it all in and then you just get like here's a report of these are the things we are 95% sure about.

Here's the 80% accuracy.

And like you just go through and answer questions and work through parts of the model to fix it.

Yeah.

So I think we're very there.

And I also have seen, you know, in very, very big projects where you have more money and you're maybe a fancy company, they purchase these.

Have you seen these, these spots?

They're these robotic dogs that walk around and scan they are creepy.

They're so creepy.

But the idea with them is that they are self driving.

So they will walk around all creepy whenever you want them to, usually in the evenings.

And they will scan parts of a construction site and feed the data back into the whatever software you're using and basically give you an update on where what's happening.

So.

And if something got built wrong, it flags it.

Yeah.

If it doesn't match the model.

Yeah.

It's.

They're, they're on construction sites, I've seen them.

So now those though again, it always comes down to with all this technology you need someone who understands it and someone who's driving it and someone who's organizing it and making sure the data is going where it's going to needs to be because you can collect it all and then it can just sit on a server somewhere and not be used.

Yeah.

And you need that something bigger than a server in most cases.

Oh yeah.

Or multiple servers.

Yeah.

In order to house the data.

Yeah.

Well, very cool.

I say whoever builds this solution is going to do very well for themselves.

Maybe we should.

Maybe we should.

I'd like to think there's already people out there working on this and I looked into it about 10 years ago and they, there were companies claiming that they were able to do this.

But I test it and it was not there.

It would be like unknown beam lots of unknowns.

I think they're working on it and actually it is on my to do list to reach back out to some of those companies and see how far they've come.

Well, all we have to do is raise, I'd say, I don't know, seed, maybe 50 mil.

Not a big deal.

No problem.

Right.

In seed, bring in some AI training experts.

You know what they need to learn.

So you just tell them and I'll just run a podcast about it.

Okay.

Do you know some people?

You can introduce us.

Yeah.

Whoever wants to invest in this and you have a $50 million extra that you willing to lose on this venture.

You'Re not going to lose it because.

This is the future.

This is going to happen.

We just have to do it fast before the others figure it out.

Just.

No, we don't do it fast, we.

Just do it better.

That's true.

Yeah.

Just better.

It just needs.

Well, so.

So here's the thing.

The point is, and in all seriousness, all these kinds of tools and products are all about timing.

So you think about like, okay, think about like you could take any example of any product.

There's maturity to the product, it's going, it's going, it's doing well.

And then there's new product comes out like somewhere else.

It's not a new idea, it just executed better because it utilizes all the new technology, doesn't have all that code base, all that, all that stale, you know, remnants that they have to drag win themselves into the new age.

And so I think the success about companies like this, I mean, we're way off topic, but I think it'd be interesting.

The success of companies and products like this is the timing.

So the question is, is AI at this point where training it with that kind of data amount is feasible?

On compute.

Compute is expensive, tokenized, you will have spent a lot of tokens.

50 million might be a drop in the bucket depending on how dense.

And I think the data is dense because it's kind of the world interpretation of the world.

It's everything.

So that's one piece and the second piece, are we the right people to do it?

Which is the second is easy.

It's.

Yes, yeah, that's easy.

But are we there in technology where within 6 to 10 months, 10, 6 to 12 months, working prototype that actually works pretty well can be made fairly?

I think we are.

Yeah.

I think a year ago I would have said no, but I think, I think we are now.

Yeah.

No, but it's interesting what you're bringing up, Alex, because I mean, if you think back to like the first base launch, the technology amounts to the technology you have in a graphing calculator.

The amount of computing power.

Right.

And that was, you know, rooms full of computers and stacks of handwritten code.

And just now.

Yeah, that is a really great question.

The amount of computing pattern needed to do this, does it exist in an accessible way and does it exist in accessible way for customers, not just us who could get our hands on it, maybe.

Right.

In other words, how much is the compute?

Right.

And I think it's more than actually teaching that model, Megan.

I think it's, I think Liz is right.

Like is what is it going to cost the, the, the product design, like let's call it, let's call this product, you know, the two.

Right.

So what is it going to cost to the company to actually process all that data and then by connection, what is it going to cost the consumer?

Because the two has to make low margin.

The two does not work for free.

The two wants to make little cheddar.

Right.

So if it costs us, you know, $10,000 to process a one story building makes no sense because you have to mark that up.

You don't have the manual labor of doing it anymore.

Yeah, that's expensive.

Yeah, labor.

Well, and this is where I think a lot of this, like before you even go to investors and say, hey, you know, we have this great idea.

Here's, you know, here's our pitch.

Check, please.

Yeah, there needs to be all that very serious research done in the form of, like, hey, are we there on attack?

Like, yeah, are we there on computer?

And I think that's all it is to it, really.

And if we're there, it's not that complicated.

I think execution is complicated, but I think it's not out of realm of possibility given that code writing can be done.

A lot of the AI training is very doable.

I don't know.

We way off.

Interesting.

No, this is on topic.

We see the gap in the market.

For the AI for the two part, and we're trying to solve that.

That's what we all do.

We're problem solvers.

Yeah.

Well, and I think our homework before we do the next episode, like this is I do think people are out there trying to solve this because it has been a big problem, the two part, for a long time.

So I want to find out what.

Where have they gotten to this year?

Because I'm hoping it's somewhere good.

All right, well, this is going to be the first episode I ever hosted that has homework assignment homework built in.

I'm assigning it to myself.

No, but I'd like to actually, you know, now I have to go because I'm kind of speaking with general knowledge of where AI is because I'm very interested in keeping up on what's going on.

But I don't know whether this specific tech is.

So let's go find out the next one.

We'll come back, update our viewers and go from there.

Any other partying, thoughts of wisdom or more homework?

Download one of the scan apps.

They have them for free, and try it.

It's super fun.

Okay.

Yeah, thank you.

That's great.

All right, so those of you who are still watching and listening to us, if you want to invest, go to.

Aurora bim or we have to tell our website manager to add an invest button.

Like tiers starting at 50 million.

Big goals with like a thermometer that.

Fills up as people invest.

Yeah, we're going to be cold for a while.

Nancy can build that.

Yeah, that's excellent.

Yeah, but, you know, getarc it.

We do it for architecture design engineers.

Aurora BIM does BIM consulting and all kinds of other technology help for your architecture design engineering firm.

On behalf of Alex, Megan and Liz, we wish you a fantastic rest of your day.

We're going to go do our homework and we'll see you next week.

Thank you for watching.