This essay updates a November 2013 article. We’ve expanded it in light of the success of our first Solid conference in May 2014, where we tested many of these ideas, and the announcement of our next Solid conference in June 2015. In addition to this update, you can stay in the loop on the latest developments in the space through our weekly newsletter.
Real and virtual are crashing together. On one side is hardware that acts like software: IP-addressable, programmable with high-level procedural languages and APIs, able to be stitched into loosely coupled systems — the mashups of a new era. On the other is software that’s newly capable of dealing with the complex subtleties of the physical world — ingesting huge amounts of data, learning from it, and making decisions in real time.
The result is an entirely new medium that’s just beginning to emerge. We can see it in Ars Electronica Futurelab’s Spaxels, which are LED-equipped quadcopters that make up a drone swarm to render a three-dimensional pixel field; in Baxter, which layers emotive software onto an industrial robot so that anyone can operate it safely and efficiently; in OpenXC, which gives even hobbyist-level programmers access to the software in their cars; and in SmartThings, which ties web services to light switches.
The new medium is something broader than terms like “Internet of Things,” “Industrial Internet,” or “connected devices” suggest. It’s an entirely new discipline that’s being built by software developers, roboticists, manufacturers, hardware engineers, artists, and designers.
Ten years ago, building something as simple as a networked thermometer required some understanding of electrical engineering. Now, it’s a Saturday afternoon project for a beginner. On the other end of the scale, companies like Local Motors and Lit Motors have demonstrated that lightweight start-ups with a handful of engineers can develop entirely new car platforms. It’s a shift we’ve already seen in programming, where procedural languages have become more powerful and communities have arisen to offer free help with programming problems. As the blending of hardware and software continues, the physical world will become democratized: the ranks of people who can address physical challenges from a lot of different backgrounds will swell.
The outcome of all of this combining and broadening, I hope, will be a world that’s safer, cleaner, more efficient, and more accessible. It might also be a world that’s more intrusive, less private, and more vulnerable to ill-intentioned interference. That’s why it’s crucial that we develop a strong community around the new discipline.
Solid, which Joi Ito and I presented for the first time in May 2014, brought members of the new discipline together to discuss this new medium at the blurred line between real and virtual. We talked about design beyond the computer screen, software that understands and controls the physical world, new hardware tools that will become the building blocks of the connected world, and frameworks for prototyping and manufacturing that make it possible for anyone to create physical devices. And we gawked at dozens of exhibits from companies large and small that demonstrated a fluid relationship between the concrete and abstract worlds.
The business implications of the new discipline are just beginning to play out. Software companies are eyeing hardware as a way to extend their offerings into the physical world — think, for instance, of Google’s acquisition of Nest and its work on a driverless car — and companies that build physical machines are viewing software as a crucial component of their products. The physical world as a service, a business model that resembles software as a service and that’s sometimes referred to as “generation, not generators,” promises to upend the way we buy and use machines, with huge implications for accessibility and efficiency. These types of service frameworks, along with new prototyping tools and open-source models, are also making hardware design and manufacturing vastly easier.
We’ve identified a few interrelated concepts that led us to Solid:
APIs for the physical world
Abstraction, modularity, and loosely coupled services — the characteristics that make the web accessible and robust — are coming to the physical world. Open source libraries for sensors and microcontrollers are bringing easy-to-use and easy-to-integrate software interfaces to everything from weather stations to cars. Networked machines are defining a new physical graph, much like the web’s information graph. These models are starting to completely reorder our physical environment. It’s becoming easier to trade off functionalities between hardware and software; expect the proportion of intelligence residing in software to increase over time.
APIs for the physical world can be anything from drag-and-drop services like IFTTT, which abstracts the complexity of your furnace to the same level as a Tweet, to sophisticated industrial systems like sensors on airliners that make turbulence data available in the same context as satellite images.
Manufacturing made frictionless
Amazon’s EC2 made it possible to start writing and selling software with practically no capital investment. New manufacturing-as-a-service frameworks bring the same approach to building things, making factory work fast and capital-light. Development costs are plunging, and it’s becoming easier to serve niches with specialized hardware that’s designed for a single purpose. The pace of innovation in hardware is increasing as the field becomes easier for entrepreneurs to work in and financing becomes available through new platforms like Kickstarter. Companies are emerging now that will become the Amazon Web Services of manufacturing.
The result will be a massive disruption to economies of scale. Designing, manufacturing, and distributing a physical product has until very recently been so expensive and required so much expertise that only big companies could build household appliances, cars, electronics, and industrial equipment. That’s changing as fixed costs drop and the price of flexible, on-demand manufacturing approaches the cost of mass manufacturing.
Software intelligence in the physical world
Machine learning and data-driven optimization have revolutionized the way that companies work with the web, but the kind of sophisticated knowledge that Amazon and Netflix have accumulated has been elusive in the offline world. Hardware lets software reach beyond the computer screen to bring that kind of intelligence to the concrete world, gathering data through networked sensors and exerting real-time control in order to optimize complicated systems. Many of the machines around us could become more efficient simply through intelligent control: a furnace can save oil when software, knowing that homeowners are away, turns down the thermostat; a car can save gas when Google Maps, polling its users’ smartphones, discovers a traffic jam and suggests an alternative route — the promise of software intelligence that works above the level of a single machine. The Internet stack now reaches all the way down to the phone in your pocket, the watch on your wrist, and the thermostat on your wall.
Every company is a software company
Software is becoming an essential component of physical devices. Consider a Nest thermostat, for instance: a large portion — perhaps the majority — of its $250 value is in its software intelligence. Software is becoming an increasingly important value component in more sophisticated devices like cars and jet engines as well. As a result, companies that might traditionally think of themselves as far removed from the technology industry have rapidly become tech companies, adding software to their traditional expertise in product engineering, manufacturing, and distribution.
Even companies that consume big machines must adopt a software-oriented mindset; just as manufacturers of jet engines are becoming software companies, the consumers of jet engines — airlines — must take on that mindset as well in order to get as much value as possible out of their assets and remain competitive.
As a result of this shift, a software start-up with promising technology might just as easily be bought by a big industrial company as by a Silicon Valley software firm. This has important organizational, cultural, and competency impact.
Design after the screen: Interactions and connected devices
Our interaction with software no longer needs to be mediated through a keyboard and screen. In the connected world, computers gather data through multiple inputs outside of human awareness and intuit our preferences. The software interface is now a dispersed collection of conventional computers, mobile phones, and embedded sensors, and it acts back onto the world through networked microcontrollers. Computing happens everywhere, and it’s aware of physical-world context.
Software replaces physical complexity
A home security system is no longer a closed network of motion sensors and door alarms; it’s software connected to generic sensors that decides when something is amiss. In 2009, Alon Halevy, Peter Norvig, and Fernando Pereira wrote that having a lot of data can be more valuable than having the most elegant model. In the connected world, having a lot of sensors attached to some clever software will start to win out over single-purpose systems.
We’ll explore each of these areas in the months ahead through our newsletter, events, analysis, research, and webcasts. Feedback is essential, so please share your thoughts through Twitter (@JonBruner) and email (firstname.lastname@example.org).