Creative Destruction Part 05: Automated Design
As part of the structure of the Creative Destruction Series, we have conducted and curated several conversations with different great thinkers and doers in the field. This includes researchers, designers, technologists, artists, owners, end-users and a few left-fielders.
For this part, I interviewed Gordon Kurtenbach. Gordon Kurtenbach is the Head of Autodesk Research where he leads a large research initiative on topics including human-computer interaction, graphics and simulation, environment and ergonomics, high-performance computing, and CAD for biotechnology.
According to AUTODESK, “Dreamcatcher is a generative design system that enables designers to craft a definition of their design problem through goals and constraints. This information is used to synthesize alternative design solutions that meet the objectives. Designers are able to explore trade-offs between many alternative approaches and select design solutions for manufacture.”
EK: What new directions do you see something like Dreamcatcher opening up for designers? What are the advantages and pitfalls?
GK: Dreamcatcher is perceived as distinct from design as we do it now. For example, if you’re designing something with a computer-aided design product and need an angle bracket, you literally have to specify the actual charting and shape of that angle bracket and think about the loads placed on it. I think the computer should do that automatically.
We can use cloud computing—a vast amount of computational resources—in both generating possible solutions and evaluating them. We're taking a time problem and turning it into a generate-and-search problem.
EK: What if the machine generates a bunch of predictable designs and you have to choose from this disappointing set?
GK: That’s part of the reason for Project Dreamcatcher, to see what happens, especially when the machine suggests designs that are sort of surprising. I’m not sure if it will extract the emotion out of designs or reduce design as value; I think it will probably relieve people from the mundane tasks but also spur creative juices in different ways.
EK: So are the functional requirements really just engineering requirements?
GK: Searching for designs that fit styling elements is technically possible. But how an actual designer’s style gets reflected and how people judge it, what personal values they bring to something is still an open question.
EK: Are you working with designers now?
GK: We have a very user-centric design process; it’s not only tactical research to experiment with the algorithms. We're trying to answer two questions. Technically, can we do this? And what do people use this for?
EK: Do you have vetted groups evaluating some of the tactics coming out of research?
GK: Dreamcatcher is still at that early stage where we just can’t just ask people to try it. But it’s actually a good candidate for a living lab because Dreamcatcher allows somebody who's not a trained designer have the machine design for them. And that can be really powerful.
EK: How long before you have a prototype that’s usable enough for somebody to evaluate?
GK: It's a couple years down the road. Last year, we decided to try and automatically generate a swing arm for a motorcycle. We knew all the forces on the swing arm over the course of its lifetime and the volume of space the swing arm was allowed to occupy. We input it into Dreamcatcher and it automatically generated a swing arm for us. It looked more like a human hip bone with really bizarre texture and curves but when we showed it to people who designed motorcycles, they said, “Oh, that works. But can you make it look like this or that?” So early days, but as Dreamcatcher gains more and more strength in absorbing functional requirements, we’ll be able to do more and more realistic things.
EK: On another topic, we’re seeing a lot more convergence of the physical and virtual worlds, with interesting interfaces between the virtual world of image, information, data and the built environment. Are you doing any work in that domain?
GK: We've been looking at different ways to use augmented reality for a long time. The hardware has great potential but we're waiting for it to reach a certain level of quality. Even once you get there, what is it really good for? What could be accomplished in a different way without having to use a headset? Do I want to wear a headset so that when I walk around a building I can look into walls and see the piping and wires or do I just want that information on my mobile phone? So, we're experimenting with a sober eye on where the real value is. But not too sober because—on the entertainment side—some of the experiences could be pretty compelling.
EK: I think there's other application for this convergence: smart materials in buildings, smart buildings, and digital wayfinding. Are you guys are doing any work in that?
GK: We're super interested in the advancements in material science because materials are becoming programmable. You see that in 3D printing, where you can actually have the computer control the materials. To do that with data obviously that would be a huge advantage. Couple this notion of a digitally programmable material with something like Dreamcatcher, and you have a really interesting combination in not only the design space but the fabrication space too.
Once the computer knows about both the design and the fabrication, life is going to get a lot easier. I think the process of designing will get a lot more powerful because all that stuff will be figured out by the computer and you’ll be able to focus on other things. Does this thing do what I want it to do? Does it copy what I want it to copy?
EK: I know that visionaries tend to be optimists. What is the dark side of this?
GK: I suppose there's the usual dark side about machines that can easily create things and allow people to rip off designs and reproduce them.
Also, are we going to end up with a bunch boring computer-generated designs that all look alike? Whenever automation happens in any area, there’s always the reality that sometimes crap comes out of what you put in.
EK: But you think great thinkers and great inventors and great artists could use the system to do something that’s breakthrough?
GK: Yeah, like with digital audio. Now we can all make crappy music but great people still make great music. I feel that's part of human nature and the society in which we live. I don’t see that being geometrically affected by technologies like Dreamcatcher.
EK: What’s in the future?
GK: What’s going on in the world of life sciences and biology and nanotechnology and how all those intermix, I think that’s just fascinating. Designing for things that change over time is pretty interesting. I think in the future, a product is going to be a lot more sophisticated and be able to adapt to the user's needs and market requirements once it's out there in the field.
You can think of this science-fictiony world where—if you need a bridge as you go—you actually design a bridge growing feet alongside a riverbank. Those are the things we’re pursuing at Autodesk. It’s just that they take a little while.
Creative Destruction Series: Introduction
Creative Destruction Series Part 01: Palpitations on the Slopes of Technology
Creative Destruction Series Part 02: Designing for Plurals, the Evolving Audience
Creative Destruction Series Part 03: Relocating Humanity
Creative Destruction Series Part 04: A Curious Stepchild of Inbound Marketing