I'm a software engineer with a passion for architecture and I have started a company dedicated to applying generative design to the architectural design process. The problem is that I am not a domain expert in architecture and I don't have any insight into the daily pain points of the design process.
Thus far, we have partnered with several AEC companies to develop and generative floor-plan tool that aims to automate the intake of residential home-buyer needs and wants and produces a set of floor plans that optimize for their inputs. I think this is a great start for us, but the purpose of this post is to collect input from architects all around the world about any blatant pain points architects are facing or any necessary features that would need to be in a new tool in order to make it valuable.
For example, I am aware that some architectural firms will design buildings in teams and some will assign a project to only one architect. Would a valuable tool need to offer some kind of collaborative services in order to bring value to you? Would you like to be able to view, annotate, and co-create documents or 3D models in one tool?
If you had the resources to build out the perfect tool for you and your team what would that tool do?
We are technologically capable and willing to explore even the most radical ideas!
A "generative floor-plan tool that aims to automate the intake of residential home-buyer needs and wants and produces a set of floor plans that optimize for their inputs" you say?
Why yes, I'd LOVE to give you more data for your architect replacement tech bro gadget, where do I sign up? Shall I also burn my degree, lobotomize myself, and shred all of my money so as to help you in your quest to commoditize and monetize my profession?
A great and reasonable response! However, I have no intention of automating away architects. I really respect architects and I believe that architects should always be in control of the overall design process. Instead of replacing architects, I want to enhance and expand your creative ability and exploration by reducing time spent on the slow-moving, less design-focused, parts of the process so that you can spend more time and energy on designing than managing stakeholders relationships. I believe that generative technology can become an architect's best friend if we are able to work together to design the most appropriate and effective solutions. If you use autoCAD or any autoDesk tool then you are already working in cooperation with generative technology. At the end of the day, we all want to provide our clients with the best building design for whatever they want to use it for and we should do whatever we can to ensure we deliver that.
the only optimisation I can think of is the fee for architects, when you increase that all else will fall into place almost miraculously.
Sep 8, 21 1:29 pm ·
·
jessenG
By optimizing the fee, are you speaking from the client's point of view (reduce the cost of having a building designed) or from the architect's point of view (increase the revenue collected for the design of a building)?
Yay, another "design" generating button pushing machine for the future suburban-garbage developer house. Totally "customized" for you for a low-low fee of $5.99!
Hard pass. This is not how design works and a major pain point is when tech-bros think they can rationalise design into a convenient little box. There are already tools for this. We don't need another.
I'm sorry you feel that way about design technology. I really don't mean to claim that the design process can be simply translated into a so-called 'little box', as that would be wildly naive.
Which design tools exist that you already use?
Sep 8, 21 2:02 pm ·
·
Non Sequitur
You certainly did claim this in your original post above. Not everything can be automated and scripted and you're most certainly not the first one to suggest this. Good luck, but don't invest too much tech-bro cash and time in this.
Sep 8, 21 2:04 pm ·
·
jessenG
Thank you for your input! I'm just trying to learn what I can from the experts in the community and even this sentiment is great to be aware of.
Sep 8, 21 2:11 pm ·
·
Non Sequitur
You're welcome. I had a colleague of mine try this while in grad school... some 15years ago, and I made the same comment during a peer review session. I even added that they should first try it on the mundane (washroom, loading, service) spaces of their "project" first to see if it was possible... They literally had no idea but they proposed plenty of shiny blob-like graphs and used clever words. In other words, even the simplest of spaces have loads of complicated variables. I can't imagine the complexities required to emulate quality design using what you describe.
"design technology" doesn't need to be more sophisticated than a pencil and a paper - computers came to help with the iterative process and to reduce error, but to have a software design a space you have to have it thinking first. AI.
just another over-promise and eventual under-delivery from silicon valley.
but to answer your questions, no.
Sep 8, 21 1:54 pm ·
·
jessenG
I agree that there has been a lot of optimistic promises made, but I don't think that gives reason to drop the topic entirely. At one point Autodesk was promising big things and it failed to deliver those promises. But over time and with trust from users, it was able to achieve something that nearly every architect uses today.
"develop and generative floor-plan tool that aims to automate the intake of residential home-buyer needs and wants and produces a set of floor plans that optimize for their inputs"
I'm only posting this because I think it's funny the OP is earnestly responding to each comment.
I was doing research into these types of tools with a professor of mine back in the day. It's an interesting concept and we had a lot of fun researching it. I'm not sure what ever happened with their research after I graduated. Who knows, I might be listed in the acknowledgements of one of their papers.
That's really exciting that you got to participate in some research on this topic. If you're cited in a paper then I'm sure I'll come across it at some point, I'm reading as much as I can.
OP, you don't understand. There are no "pain" points. There are no mistakes early in the process. There are undiscovered relationships which your program would eliminate.
jessenG should write software that does something he knows about. Like writing software. Then he could write software that would write software for him.
Sep 8, 21 10:11 pm ·
·
Block this user
Are you sure you want to block this user and hide all related comments throughout the site?
Archinect
This is your first comment on Archinect. Your comment will be visible once approved.
Generative Architecture: Optimizing the Architectural Design Process
I'm a software engineer with a passion for architecture and I have started a company dedicated to applying generative design to the architectural design process. The problem is that I am not a domain expert in architecture and I don't have any insight into the daily pain points of the design process.
Thus far, we have partnered with several AEC companies to develop and generative floor-plan tool that aims to automate the intake of residential home-buyer needs and wants and produces a set of floor plans that optimize for their inputs. I think this is a great start for us, but the purpose of this post is to collect input from architects all around the world about any blatant pain points architects are facing or any necessary features that would need to be in a new tool in order to make it valuable.
For example, I am aware that some architectural firms will design buildings in teams and some will assign a project to only one architect. Would a valuable tool need to offer some kind of collaborative services in order to bring value to you? Would you like to be able to view, annotate, and co-create documents or 3D models in one tool?
If you had the resources to build out the perfect tool for you and your team what would that tool do?
We are technologically capable and willing to explore even the most radical ideas!
A "generative floor-plan tool that aims to automate the intake of residential home-buyer needs and wants and produces a set of floor plans that optimize for their inputs" you say?
Why yes, I'd LOVE to give you more data for your architect replacement tech bro gadget, where do I sign up? Shall I also burn my degree, lobotomize myself, and shred all of my money so as to help you in your quest to commoditize and monetize my profession?
A great and reasonable response! However, I have no intention of automating away architects. I really respect architects and I believe that architects should always be in control of the overall design process. Instead of replacing architects, I want to enhance and expand your creative ability and exploration by reducing time spent on the slow-moving, less design-focused, parts of the process so that you can spend more time and energy on designing than managing stakeholders relationships. I believe that generative technology can become an architect's best friend if we are able to work together to design the most appropriate and effective solutions. If you use autoCAD or any autoDesk tool then you are already working in cooperation with generative technology. At the end of the day, we all want to provide our clients with the best building design for whatever they want to use it for and we should do whatever we can to ensure we deliver that.
the only optimisation I can think of is the fee for architects, when you increase that all else will fall into place almost miraculously.
By optimizing the fee, are you speaking from the client's point of view (reduce the cost of having a building designed) or from the architect's point of view (increase the revenue collected for the design of a building)?
the cost of having a building designed is peanuts
Yay, another "design" generating button pushing machine for the future suburban-garbage developer house. Totally "customized" for you for a low-low fee of $5.99!
Hard pass. This is not how design works and a major pain point is when tech-bros think they can rationalise design into a convenient little box. There are already tools for this. We don't need another.
I'm sorry you feel that way about design technology. I really don't mean to claim that the design process can be simply translated into a so-called 'little box', as that would be wildly naive.
Which design tools exist that you already use?
You certainly did claim this in your original post above. Not everything can be automated and scripted and you're most certainly not the first one to suggest this. Good luck, but don't invest too much tech-bro cash and time in this.
Thank you for your input! I'm just trying to learn what I can from the experts in the community and even this sentiment is great to be aware of.
You're welcome. I had a colleague of mine try this while in grad school... some 15years ago, and I made the same comment during a peer review session. I even added that they should first try it on the mundane (washroom, loading, service) spaces of their "project" first to see if it was possible... They literally had no idea but they proposed plenty of shiny blob-like graphs and used clever words. In other words, even the simplest of spaces have loads of complicated variables. I can't imagine the complexities required to emulate quality design using what you describe.
"design technology" doesn't need to be more sophisticated than a pencil and a paper - computers came to help with the iterative process and to reduce error, but to have a software design a space you have to have it thinking first. AI.
just another over-promise and eventual under-delivery from silicon valley.
but to answer your questions, no.
I agree that there has been a lot of optimistic promises made, but I don't think that gives reason to drop the topic entirely. At one point Autodesk was promising big things and it failed to deliver those promises. But over time and with trust from users, it was able to achieve something that nearly every architect uses today.
I call BS on the OP's claims.
Which claims are you doubtful about? I am happy to c larify!
This part:
"develop and generative floor-plan tool that aims to automate the intake of residential home-buyer needs and wants and produces a set of floor plans that optimize for their inputs"
How are you doing this?
Layouts, programming, bubble diagrams?
I'm only posting this because I think it's funny the OP is earnestly responding to each comment.
I was doing research into these types of tools with a professor of mine back in the day. It's an interesting concept and we had a lot of fun researching it. I'm not sure what ever happened with their research after I graduated. Who knows, I might be listed in the acknowledgements of one of their papers.
Good luck.
That's really exciting that you got to participate in some research on this topic. If you're cited in a paper then I'm sure I'll come across it at some point, I'm reading as much as I can.
Thank you for the comment!
OP, you don't understand. There are no "pain" points. There are no mistakes early in the process. There are undiscovered relationships which your program would eliminate.
jessenG should write software that does something he knows about. Like writing software. Then he could write software that would write software for him.
Block this user
Are you sure you want to block this user and hide all related comments throughout the site?
Archinect
This is your first comment on Archinect. Your comment will be visible once approved.