Design is Becoming a New Team Sport
Something interesting is happening in the design world. Tools — things we utilize to perform specific tasks — are increasingly developing into much more than just things. As they begin to function more like team colleagues or co-workers that collaborate with us to solve problems, tools are becoming reliable assistants.
Jeff Kowalski, CTO at Autodesk, Inc., pursues innovative ideas on the horizon rather than trends of the moment. He envisions a positive outcome for humans and Machine Learning (ML) — a transition that is already unfolding.
“We’re at another major inflection point now,” Kowalski explains, “focused on how one of our most powerful digital tools is expanding into the physical realm.”
As the relationships between humans and machines evolve, the transition promises to turn design into a team sport where people and tools play together, side by side. This new way of working is transforming design in exciting ways, expanding the possibilities of what we can make and how we can shape the world around us.
Let’s meet some of the new team players.
Player #1: Generative Design
One of the most influential tools moving the relationship between humans and machines forward in this new collaborative direction is generative design. This technology allows a designer to feed various criteria and constraints into computers in the cloud, which rapidly generate hundreds of design options that satisfy those conditions.
Do you need to manufacture a chair made only of plastic that can support up to 300 pounds? Within seconds, a generative design processor will present you with a multitude of options able to match your exact requirements.
Would you like to design a quadcopter? While we humans might have a very fixed idea in mind of what this lightweight drone should look like, generative design tools have no such preconceptions. Tools can explore a full range of possible solutions that meet your criteria and then come up with design options you may have never imagined — like a drone frame that resembles the skeleton of a flying squirrel.
The design team at Airbus employed a similar type of biomimicry when they used generative design to create a better partition for the galley section of their A320 planes. The resulting ‘bionic partition’ is 45 percent lighter than conventional partitions and equally as strong. Airbus estimates that this will save a half a million metric tons of CO2 per year.
Player #2: Robotics
While generative design provides the tools for coming up with designs, robotics offers exciting new ways to make the designs.
Robots have been mainstays of manufacturing facilities and other industrial settings for decades, but, increasingly, we are finding ways to enhance their utility by combining robotics with other technologies like generative design and additive manufacturing.
One particularly vivid example of enhanced robotics is being demonstrated in the Netherlands, where Autodesk, headquartered in San Rafael, California, has been working with Dutch designer Joris Laarman and his team at MX3D. Six-axis pivoting industrial robots, using an algorithmic feedback system, have begun autonomously manufacturing a steel pedestrian bridge by ‘printing in mid-air’ — incrementally building the metal structure that will later be installed over a canal in central Amsterdam. This robotic marvel with load-bearing capabilities is already under construction with an expected completion in 2017.
“By printing with six-axis industrial robots, we are no longer limited to a square box in which everything happens,” says Tim Geurtjens, CTO at MX3D. “Printing a functional, life-size bridge is, of course, the ideal way to showcase the endless possibilities of this technique.”
Laarman adds: “This bridge will show how 3D printing finally enters the world of large-scale, functional objects and sustainable materials while allowing unprecedented freedom from form. The symbolism is a beautiful metaphor that connects futuristic technologies with the historical city to showcase the best of both worlds.”
Equally as impressive is MX3D’s suspension technique in which the robots use 3D printing to support their own weight, which is integrated into the construction process and promises to transform the building of future structures.
“Everyone who says 3D printing is just pressing a button doesn’t really know how it works,” Laarman points out. “It’s very hands-on and very elaborate. We were all a bit bored with all the tiny, keychain-sized things people were making, so we really tried to push it to a higher level by using real materials like wood and metals.”
Player #3: The Internet of Things
Once you make things, why not give them a nervous system? The Internet of Things (IoT) — that ubiquitous network of connected sensors — has started playing this role by gathering data about a product’s surroundings and reporting it back to us.
As embedded sensors in products get more sophisticated, we can gain a greater understanding of how a product functions in the real world — and how we can improve its design.
For example, Autodesk worked with the Bandito Brothers, a media production team in Los Angeles, California, to outfit a racecar with dozens of sensors that could collect billions of data samples of how the car performs in race conditions. Taking this data and feeding it into a generative design tool has allowed the team to build a custom chassis that maximizes performance based on the conditions as they were captured.
Player #4: Artificial Intelligence
If generative design, robotics, and the IoT are the tools changing the way humans and machines work together on design tasks, then Artificial Intelligence (AI) is the ‘rocket fuel’ that accelerates their impact.
AI gives our tools a learning capability so they can continuously get better at doing their jobs. This means that generative design tools will begin to learn what types of designs we like and do not like, and take note of our preferences; robots no longer need explicit instructions in their application programs to resolve an appropriate outcome. The IoT can use AI not only to perceive but also react intelligently to the real world.
All of this added intelligence gives the tools the flexibility to be more creative in their problem solving. As a result, computers are improving upon human capabilities like trial and error, intuition, and taking creative leaps. Given the array of design challenges our world currently faces, this added source of creativity is a welcome and positive development.
In this new era of ML and advanced design tools, the relationship between humans and machines is moving exponentially in exciting and inspiring ways.
Designers and engineers should no longer view tools as machines that need detailed instructions in order to operate. Instead, they have the opportunity to embrace these tools as true collaborators capable of helping to solve big problems in ways humans alone could not nor would not resolve without them.
By viewing design as a team sport that includes both humans and machines working together for beneficial advancement, everybody (and everything) wins.
Human beings have a long record of shaping the world. Moving forward, computers will join by shaping the things that shape the world. This unprecedented blend of humanity and technology is exciting to experience, and we will be seeing many similar results in the not-so-distant future.
“Fortunately for us, computers are starting to develop human-style capabilities to augment our own,” Kowalski concludes. “I think it’s going to fundamentally change our relationship with tools and the design process. Humans will no longer be operators with their tools. We will be more like mentors to our tools, coaching them and providing them with guidance and experiences.”
Link: Machine Learning & Design
Sixty years ago, a programmer taught a machine to beat humans at tic-tac-toe. Since then, we have witnessed IBM’s supercomputers beating the world chess champion and contestants of the American game show Jeopardy! More recently, Google DeepMind’s computer game ‘AlphaGo’ defeated the best human at Go, the world’s most complex game. In less than a single human lifetime, the computer has gone from learning a child’s game to mastering the game recognized as the pinnacle of strategic thought.
Another example is the Atari video game Breakout. By looking at only the score and controller input, DeepMind’s AI learned how to play the game better than any human by playing millions of games throughout the night and automatically spread its newfound knowledge quickly to other computers. In human terms, your knowledge of Breakout doesn’t necessarily help your friend become better at the game. In contrast, when one computer masters Breakout, the other machines improve as well because they are all connected.
As ML evolves, generative design will accelerate by automating designers’ reactions and incorporating their unspoken preferences into the design process. ML will also give robots the ability to complete tasks without having to depend on designers for explicit instructions.