In architecture, it’s imaginable that without major technical advances, a program can be made available to the public that automates architecture.
Here’s how it may work- a simple video game like interface would allow a user to generate a floor plan, probably giving cues and tips along the way, and autocorrecting for spatial dimensions, proximity relationships, and code. The system could conceivably have the entire code book baked in. Links to materials, products, etc would be chosen and recommend based on some algorithms that anticipate your tastes, that match between products, etc. Probably, this would be the main source of revenue for the creators of the program, like part design software part online shop.
Once the user has the house designed and specified, for a fee, the system would generate construction documents that are code compliant to local standards.
This is probably possible. Probably not good. And it’s probably going to happen at some point in the near future.
SlammingMiruvor
Feb 26, 21 2:09 pm
Why isn't it good? If a project is so mundane and thoughtless that it can be performed by TurboTax's TurboArchitecture v1.0 then so be it. If anything, it opens the door to people who wouldn't be able to afford personalized Architectural services and opens them to the idea of engaging an Architect in the future.
Otherwise, I'm sure my personal career decisions influences this more than anything, but the vast amount of work I've done are renovation efforts. Those efforts, including the client relationships, creative problem solving, and customized solutions are incredibly hard to automate. I'll never say my job cannot be automated, but I'm certainly not threatened directly by your hypothetical. Selfish? Yes, but I stand by what I said about opening avenues for people who normally wouldn't consider hiring an architect. It's inevitable people running through the software would want a dialogue instead, and would turn to a living breathing Architect.
x-jla
Feb 26, 21 2:20 pm
I agree with that mostly
x-jla
Feb 26, 21 2:22 pm
And turbotax is a good analogy
x-jla
Feb 26, 21 2:23 pm
I’m imagining a little number at the top that adjusts estimated cost too as you go...
randomised
Feb 26, 21 4:05 pm
"It is urgent that, from today, the disciplines of the built environment quickly rethink entirely what the rapid digitisation of our world has meant in order for architecture and design to be a catalyst for real, positive change for the everyday person. Instead of falling for the more homogenous, striated, segregated and inequitable world that neoliberal perspectives on technology promote, and that is inherent in what we recognise as ‘digital architecture’, the architecture discipline should be working towards forms and processes of fully automated production that promote equity, sustainability, democracy, heterogeneity and inclusivity. From tackling climate change to the global housing crisis, new radical forms of digital architecture are urgently needed."
Still beating that dead horse. You value design too much. Even if you can custom configure them. The drawing preparation part is cheap. Most of architects effort is in coordination(scope, consultants, contractors, building officials, etc) and unique creative solutions(designs, details, etc). The cost of construction is the main concern. Unless the methodology of construction has been improved to drive the cost down. You can generate endless designs without the actual fund to build. Prefab currently does not solve it, the same way of construction just moved to the factory. Maybe automated robotic construction could, who knows.
Especially in residential, If someone needs a nice modern house construction set with their choice of finish and layout, I could easily do the set for you for 5k-10k. Just pull out my library and modify based on your taste. However, If you can afford 1M+ to build the new house, I am sure you can afford an architect's fee to pamper you on your custom house.
Meh most people just buy used 2x4 vinyl wood stick because they are cheaper. And depends on the market, can even flip that McMansion for massive profit.
x-jla
Feb 26, 21 6:04 pm
Yeah...but these tools will allow the architect to be totally circumvented...especially if contractors use them to provide design build...because then the coordination stuff is taken care of. I agree that there will always be a need for the real custom designs...and that probably cannot be automated away...but the bulk of architecture likely can be automated away. My curiosity is what this does to the industry. It would seem that it would push it even more into a service for the elite...which I’m not saying is good or bad...just saying
bowling_ball
Feb 26, 21 6:22 pm
What a load of nonsense. Clients have opinions. There's coordination between consultants. There are local variations. Hell, there's even variation from site to site. It's not happening beyond maybe auto populated assembly schedules
x-jla
Feb 26, 21 7:36 pm
Consultants could be automated too. The entire building could be...From HVAC to engineering...anything outside of artistic/creative/custom things could be automated, even construction.
x-jla
Feb 26, 21 7:41 pm
Site plans could be automated too. I don’t see why this would not be doable. The question is, who’s using it. If it’s something that prospective home owners will use, probably not that disruptive. If it’s something a GC can use to expand scope into design-build without hiring a designer...then it’s probably more disruptive because the GC can take it from start to finish. And, the GC can handle all the coordination
x-jla
Feb 26, 21 7:43 pm
The program itself can even potentially stamp the plans.
bowling_ball
Feb 26, 21 7:46 pm
x- , I know we disagree about politics and that's okay. But you're so far out of your element here, it's laughable. Let's get specific if you want - HOW could consultants be automated? Give us an example.
x-jla
Feb 26, 21 8:18 pm
Well let’s keep it simple for arguments sake. A client designs a 2 bedroom house, chooses a standardized wood framing, etc. The AI can apply simple rules to figure out the framing and engineering. the AI figures can figure the min code compliant number or receptacles, lights, etc...And then it can generate a wiring/electrical plan. This isn’t that hard to do. It is because it requires a huge data base of information, codes, etc....but essentially it’s just about creating rules and relationships between the various components....giving the user some freedoms...and the AI autocorrects in real time and coordinates these components, codes, etc in real time. Kind of like a way more complicated version of TurboTax.
x-jla
Feb 26, 21 8:21 pm
What’s so complicated that cannot be automated is a better question?
x-jla
Feb 26, 21 8:28 pm
Another example is engineering. Here’s a really simple example. A user designs a wall. The user inputs their zip. The program accesses data about city, local codes, wind, etc. as the user manipulates the length and height of the block wall, the system adjusts the engineering and calcs in real time. Engineering done. anything with an objective answer, like engineering, can be automated. Why not.
x-jla
Feb 26, 21 8:30 pm
Highly variable things like site and topography may require uploading surveys or whatever. But once that’s in, I don’t see why most other things cannot be AI
curtkram
Feb 26, 21 9:39 pm
AI can send a drone to map topography. probably be faster and better than what civil engineers do
.
bowling_ball
Feb 27, 21 12:42 am
I'm listening. But I'm also thinking that nobody really needs an algorithm to figure out how many receptacles a room needs - they just need a $15/hr apprentice and a tape measure. Because design isn't just about meeting code, it's about creating comfortable, durable, delightful spaces as well. And that's why we'll never be replaced.
bowling_ball
Feb 27, 21 12:43 am
Also I'd like to pat myself on the back for this post because it's been a long week and I'm high AF and I think I explained myself well for a change. Have a great night.
zonker
Feb 26, 21 6:12 pm
In Modular, properly done that is, it is possible to build for 15% less - it's getting there, with automated shop drawings. I've seen it - Rad Urban, Factory OS
bowling_ball
Feb 26, 21 7:31 pm
Modular is not the same as automated.
Non Sequitur
Feb 26, 21 6:46 pm
jla, if there was a market for this level of generic automation, then auto desk would have developed it years ago.
George Valdes
Mar 20, 21 12:34 am
Autodesk only acquires companies they don’t really build anything novel in house. Also, see companies like Hypar, Testfit.io, and Skipp.co
zonker
Feb 26, 21 7:06 pm
It's on my Revit tool bar, Autodesk is working on it
SneakyPete
Feb 26, 21 7:23 pm
might wanna edit your last name outta that
zonker
Feb 26, 21 7:35 pm
How do I edit it?
curtkram
Feb 26, 21 8:08 pm
photoshop
bowling_ball
Feb 26, 21 7:36 pm
The benefits of automation are in construction, not design.
Let's take a $10M project - a 10% savings in construction is $1M; a 10% savings in design might be $50k. Big difference.
Can we move on, please?
natematt
Feb 27, 21 4:50 am
I was going to say this. Design is the cheap part of most construction. Automation will increase in the profession, and ideally should be used to make it easier and cheaper. Fully replacing architects in the mid future seems fairly unlikely though, and probably not as cost effective as is being presented. The real cost savings are in automation of construction, which in pieces is much easier and more impactful to do.
However, the construction industry is also much bigger, and even more resistant to change than architecture, if they don't see it as a way to make more money, they aren't going to want to do it. I already like to point out that given 100 dollars to do a thing, the architect would put 80 in the material and 20 in the labor... the contractor would flip that around.
randomised
Feb 27, 21 5:36 am
Architects have already been replaced, or were never involved in the first place. So much of what’s being built is done without architects and has been since forever.
In architecture, it’s imaginable that without major technical advances, a program can be made available to the public that automates architecture.
Here’s how it may work- a simple video game like interface would allow a user to generate a floor plan, probably giving cues and tips along the way, and autocorrecting for spatial dimensions, proximity relationships, and code. The system could conceivably have the entire code book baked in. Links to materials, products, etc would be chosen and recommend based on some algorithms that anticipate your tastes, that match between products, etc. Probably, this would be the main source of revenue for the creators of the program, like part design software part online shop.
Once the user has the house designed and specified, for a fee, the system would generate construction documents that are code compliant to local standards.
This is probably possible. Probably not good. And it’s probably going to happen at some point in the near future.
Why isn't it good? If a project is so mundane and thoughtless that it can be performed by TurboTax's TurboArchitecture v1.0 then so be it. If anything, it opens the door to people who wouldn't be able to afford personalized Architectural services and opens them to the idea of engaging an Architect in the future.
Otherwise, I'm sure my personal career decisions influences this more than anything, but the vast amount of work I've done are renovation efforts. Those efforts, including the client relationships, creative problem solving, and customized solutions are incredibly hard to automate. I'll never say my job cannot be automated, but I'm certainly not threatened directly by your hypothetical. Selfish? Yes, but I stand by what I said about opening avenues for people who normally wouldn't consider hiring an architect. It's inevitable people running through the software would want a dialogue instead, and would turn to a living breathing Architect.
I agree with that mostly
And turbotax is a good analogy
I’m imagining a little number at the top that adjusts estimated cost too as you go...
"It is urgent that, from today, the disciplines of the built environment quickly rethink entirely what the rapid digitisation of our world has meant in order for architecture and design to be a catalyst for real, positive change for the everyday person. Instead of falling for the more homogenous, striated, segregated and inequitable world that neoliberal perspectives on technology promote, and that is inherent in what we recognise as ‘digital architecture’, the architecture discipline should be working towards forms and processes of fully automated production that promote equity, sustainability, democracy, heterogeneity and inclusivity. From tackling climate change to the global housing crisis, new radical forms of digital architecture are urgently needed."
–Mollie Claypool
taken from: Fully Automated Architecture
the program is called Dynamo.
Exactly, I'm learning Dynamo using my prior life where I was a programmer and a video game designer to drive architecture.
Project Frog has already used Dynamo to perform the design of school buildings, do the engineering and stamp it. -
https://www.projectfrog.com/
Still beating that dead horse. You value design too much. Even if you can custom configure them. The drawing preparation part is cheap. Most of architects effort is in coordination(scope, consultants, contractors, building officials, etc) and unique creative solutions(designs, details, etc). The cost of construction is the main concern. Unless the methodology of construction has been improved to drive the cost down. You can generate endless designs without the actual fund to build. Prefab currently does not solve it, the same way of construction just moved to the factory. Maybe automated robotic construction could, who knows.
Especially in residential, If someone needs a nice modern house construction set with their choice of finish and layout, I could easily do the set for you for 5k-10k. Just pull out my library and modify based on your taste. However, If you can afford 1M+ to build the new house, I am sure you can afford an architect's fee to pamper you on your custom house.
Meh most people just buy used 2x4 vinyl wood stick because they are cheaper. And depends on the market, can even flip that McMansion for massive profit.
Yeah...but these tools will allow the architect to be totally circumvented...especially if contractors use them to provide design build...because then the coordination stuff is taken care of. I agree that there will always be a need for the real custom designs...and that probably cannot be automated away...but the bulk of architecture likely can be automated away. My curiosity is what this does to the industry. It would seem that it would push it even more into a service for the elite...which I’m not saying is good or bad...just saying
What a load of nonsense. Clients have opinions. There's coordination between consultants. There are local variations. Hell, there's even variation from site to site. It's not happening beyond maybe auto populated assembly schedules
Consultants could be automated too. The entire building could be...From HVAC to engineering...anything outside of artistic/creative/custom things could be automated, even construction.
Site plans could be automated too. I don’t see why this would not be doable. The question is, who’s using it. If it’s something that prospective home owners will use, probably not that disruptive. If it’s something a GC can use to expand scope into design-build without hiring a designer...then it’s probably more disruptive because the GC can take it from start to finish. And, the GC can handle all the coordination
The program itself can even potentially stamp the plans.
x- , I know we disagree about politics and that's okay. But you're so far out of your element here, it's laughable. Let's get specific if you want - HOW could consultants be automated? Give us an example.
Well let’s keep it simple for arguments sake. A client designs a 2 bedroom house, chooses a standardized wood framing, etc. The AI can apply simple rules to figure out the framing and engineering. the AI figures can figure the min code compliant number or receptacles, lights, etc...And then it can generate a wiring/electrical plan. This isn’t that hard to do. It is because it requires a huge data base of information, codes, etc....but essentially it’s just about creating rules and relationships between the various components....giving the user some freedoms...and the AI autocorrects in real time and coordinates these components, codes, etc in real time. Kind of like a way more complicated version of TurboTax.
What’s so complicated that cannot be automated is a better question?
Another example is engineering. Here’s a really simple example. A user designs a wall. The user inputs their zip. The program accesses data about city, local codes, wind, etc. as the user manipulates the length and height of the block wall, the system adjusts the engineering and calcs in real time. Engineering done. anything with an objective answer, like engineering, can be automated. Why not.
Highly variable things like site and topography may require uploading surveys or whatever. But once that’s in, I don’t see why most other things cannot be AI
AI can send a drone to map topography. probably be faster and better than what civil engineers do .
I'm listening. But I'm also thinking that nobody really needs an algorithm to figure out how many receptacles a room needs - they just need a $15/hr apprentice and a tape measure. Because design isn't just about meeting code, it's about creating comfortable, durable, delightful spaces as well. And that's why we'll never be replaced.
Also I'd like to pat myself on the back for this post because it's been a long week and I'm high AF and I think I explained myself well for a change. Have a great night.
In Modular, properly done that is, it is possible to build for 15% less - it's getting there, with automated shop drawings. I've seen it - Rad Urban, Factory OS
Modular is not the same as automated.
jla, if there was a market for this level of generic automation, then auto desk would have developed it years ago.
Autodesk only acquires companies they don’t really build anything novel in house. Also, see companies like Hypar, Testfit.io, and Skipp.co
It's on my Revit tool bar, Autodesk is working on it
might wanna edit your last name outta that
How do I edit it?
photoshop
The benefits of automation are in construction, not design.
Let's take a $10M project - a 10% savings in construction is $1M; a 10% savings in design might be $50k. Big difference.
Can we move on, please?
I was going to say this. Design is the cheap part of most construction. Automation will increase in the profession, and ideally should be used to make it easier and cheaper. Fully replacing architects in the mid future seems fairly unlikely though, and probably not as cost effective as is being presented. The real cost savings are in automation of construction, which in pieces is much easier and more impactful to do.
However, the construction industry is also much bigger, and even more resistant to change than architecture, if they don't see it as a way to make more money, they aren't going to want to do it. I already like to point out that given 100 dollars to do a thing, the architect would put 80 in the material and 20 in the labor... the contractor would flip that around.
Architects have already been replaced, or were never involved in the first place. So much of what’s being built is done without architects and has been since forever.
AI will soon be smarter than we are.