Editor’s note: Rudina Seseri is a partner at Fairhaven Capital and serves as an Entrepreneur-In-Residence at the Harvard Business School. Wan Li Zhu is a principal at Fairhaven Capital. He is an MIT-trained computer scientist and contributed to artificial intelligence projects at the Media Lab.
The robotics revolution has been in the making for decades, but market expectations have historically outpaced technology readiness. While industrial and military sectors have adopted a number of high-priced robotics solutions, the consumer sector has lagged due to lack of technological maturity and high costs.
In recent years, we have seen accelerated levels of innovation in both software and hardware that are now driving new possibilities for consumer readiness and adoption of personal robotics.
Proliferation of devices, compute, and bandwidth. Instead of requiring expensive custom onboard computing, robots are now able to leverage commodity hardware, smartphones and cloud computing for processing and storage. This has implications on both cost and availability of data for machine perception and learning. Under this new paradigm, a robot is often a set of commoditized sensors (and actuators) that leverage the cloud for intelligence. Nest is an example of this.
Progress in natural language processing, speech, vision and machine learning. New hardware and algorithms, tuned by corpora of training data, have made machines more perceptive and improved interface with humans. This is a self-reinforcing loop: As machines can better understand the real world, they learn at a faster rate. These new interfaces open the market for consumer applications where users can interact with machines in near-natural language and gestures. Siri, Kinect, and Wolfram Alpha are examples of this.
API-fication of online services. The ability for machines to tap into multiple online data sources and services allows them to quickly stitch together value, reducing time and cost to go to market. Location-based data, financial and weather, and increasingly healthcare data are all examples.
In the past few years, there have been major investments by large players that have helped to validate and reinvigorate the robotics market. IBM’s Watson platform, and Google’s driverless car and string of recent acquisitions in AI/machine learning come to mind. As most breakthrough technology innovations require large entities to lay down the costly foundations, we are now seeing that happen in AI. For example, Watson has reduced the barrier to creating innovative AI applications that process large amounts of unstructured data to arrive at accurate answers.
What do the next five to 10 years of consumer robotics look like? Will cars drive themselves? Will household robots assist with our daily chores? And will robots ultimately interact and transact on our behalf and even with other robots? We predict the following developments will take shape.
Redefined Robotic Form Factor
The humanoid robot popularized by media will not be the dominant form of consumer robotics. Artificially intelligent devices will take on a multitude of forms where the form factor will more closely match its functions and use case. Many more will be in the form of embedded intelligence within everyday systems we are already familiar with. One thing we can predict: goodbye flat, rectangular devices; hello, form factor diversity.
Technology Cost, Productivity Gain
Robotics will initially augment and eventually replace high-cost human labor. The market acceptance of this progress will be driven by reliability and safety and there will be interim solutions. For example, driverless cars are preceded by cars with heads-up displays and automated sensory-enabled breaks already in the market.
As devices learn and begin to anticipate consumer needs, they make more recommendations and even make transactions on the consumer’s behalf. In the Internet of Things world, device providers will be more willing to lower prices of the hardware and appliances if they get a cut of revenue from services that are delivered on those platforms. Consumers will be able to download “apps” (either paid or free), which then connect to the cloud to deliver functionality via a subscription model, for example.
A New Platform and Ecosystem
Finally, the computer industry structure will loosely map to the robotics landscape. There will likely be a dominant open OS platform adopted and supported by large vendors (like Linux/Android) and a few manufacturer specific closed OS’s (like Windows), a number of “app” stores (like iTunes), and a wide variety of apps and API-based services offered by small and large companies. A new wave of education providers, engineers and service professionals, security providers, financial/insurance products, and legal frameworks will also emerge.