With the rise of machine learning and its impact on the news ways of making things, Business Today talks to Tom Wujec, Fellow at the American engineering, and animation software company Autodesk, about the next level of design thinking for achieving breakthrough results.
BT: What is generative design and how does it differ from the usual design process?
Wujec: Generative design is design thinking plus technology. In that sense, it's about using the computer as a partner rather than just another tool.
In traditional design process, the designer comes up with an idea, and then she will build the digital model, then the physical model, then test that model, evaluate it and then may be make changes. This process enables her to make 3 or 4 or may be 10 iterations.
But, with the emerging technologies and super hi-speed computers, with next generation algorithms, it's possible for a computer to do much more than designing. In Generative Design, the designer doesn't do the sketches, but describes the goals and constraints of the design in a way that the computer can understand. For instance, if you want to design a chair, of certain weight, certain material etc- the computer will generate a design that fulfills these goals, and then makes another design with a little variation, and then another and so on. So, by the time you finish a cup of coffee, the computer would have generated 2,70,000 designs - more than a designer can make in their lifetime. So the designer here orchestrates and choreographs the design process that leads to breakthrough of design in astonishing ways - managing complexities, managing alternatives and finding designs -- that never ever would occur to a human.
BT: Any instance you can share where this kind of process was applied.
Wujec: Yes, at our new office in Toronto called MaRS building, where MaRS stand for Medical and Related Sciences, which will be used by our people in research and media and entertainment. The lead architect for this building is David Bejamin. As we were working on it, we had to address the problem of getting the right optimal space. There is no one correct design, there are rather many optimal designs. You can change the culture of an office by making some things more important and other things less important. So, we applied Generative Design and Machine Learning to design our office space. We let people fill some really simple questionnaire about their preferences -- each person's desire in terms of space, what they want and don't want - if they prefer window seat, open seating, closed space, distance from pantry and others. All of this was put in the computer which then produced a variety of designs and the space was designed to offer several combinations of work spaces based on the outcomes.
BT: Does this process led to subjective biases incorporated in design?
Wujec: It is not like we offload an excel sheet of goals and constraints to the computer. What it is doing is, when done well, is bringing more of human to the equation and elevating something which is most human, which is trust.
Let me tell you how. So, I am working on a book called The Future of Making Things, which is a follow up to the book -Imagine Design Create. In this book, we are looking at 4 classes of technology that are changing everything - capture, compute, create and compose. Capture is about the sensors, which helps to bring the physical world into the digital world. Compute, uses the digital model for simulation, third is Create, which is about fabrication, which is about Robotics, which is continuing to evolve. 3D printing is one of the six ways in which we can manipulate matter, so that we can sculpt, mold etc. So, capture, compute and create is what human beings are doing, when we overlay computer technology on top of that, we tremendously transform the way in which we design and make things. Fourth technology class is called compose, which is using material sciences so we can better understand the property of materials. Now we are able to capture, compute and create material themselves, either by better understanding specific properties of existing materials or generating new materials that fits the goals that we have.
Take for example concrete. Half of the things in this world are made of concrete - buildings, floors, etc. So, there is this microbe, found on the ocean floors, that makes silicates when it comes in contact with water. This microbe is put in the concrete, so when there is crack and water seeps in, the microbe, repairs the rock. That gives us self healing concrete.
BT: This seems to be a futuristic way of making things. Has this already been used in manufacturing?
Wujec: Yes, certainly. Capture, compute, create, compose is changing the future of making things. I have been spending some time with Nike, where they were working on a shoe for a gold medal athlete to create the perfect running shoe for her. They 'captured' the property of her feet and form, and put it in a scanner. They calculated the dynamics of the pressure curve of her feet and also the motion during turning. They create a digital model of her feet, generated designs, and fabricated the shoe for her.
It's the marriage, the synthesis of the physical and digital world. So the future of making is bringing more of ourselves, and almost having superhuman power, to sense the world, to think or compute, to be able to manipulate matter. That's really the exciting part! There is another company called Feetz, where you send photos of your feet from your cell phone to them, and they will 3D print the sandals perfect for your feet.