September 22, 2016
Consider the platform
I first met Watson in 2011. He was a bit cold, rippingly witty and, at the time, one of the most famous pop cultural figures in the country. The month before our meeting, Watson had dazzled the country with a phenomenal week of shows on Jeopardy, taking on the game show’s most prominent nerds with an effortless command of obscure trivia and soft banter with Alex Trebek. But at interest this particular month were the servers that made him tick. Nestled in the quiet exurb of Yorktown Heights, IBM’s Thomas J. Watson IBM Research Facility is a hotbed of thinking about what computer science can deliver the world; more than five years ago, I could not have anticipated the influence Watson’s machine learning algorithm-turned-platform could ultimately cultivate.
At the time, I was serving as part of the team developing IBM’s Smarter Cities Challenge, a global IBM Corporate Citizenship effort to deploy units of senior IBMers to answer thorny urban challenges. Selecting among hundreds of city applicants, we had been heads down for weeks sorting and making difficult decisions between water challenges in Eindhoven and operational dexterity opportunities in Newark. Watson was in his silicon-bound stoicism a fine diversion.
As the nation was given its first taste of what cognitive computing might accomplish, IBM had to sort what it was going to do with the attention, equity and assets it had built. It had spent the better part of a decade developing the most sophisticated cognitive computer in the world, and took it for a very public test drive. During the decade software ate the world, could Watson make sense of it all?
Watson is a transformative, generationally significant, platform.
After trouncing Jeopardy! contestants, Watson’s handlers turned their sights to commercial applications. The team needed to prove the viability of a cognitive computing tool in the most opportunistic and lucrative sectors. It first took on the medical community, taking reams of therapeutic and medical data, symptoms and diagnosis histories, and deploying Watson in limited tests in hospitals as another voice in the room for doctors and patients. If Watson could untangle decades of trivia, could it make diagnoses easier, too?
Team Watson also began to understand the far-reaching implications of a platform and the applications to the broader public. Most interesting to my work were the first-take applications and ambitions for education, taking and ingesting lesson plans, individual classroom outcomes and textbook content to recommend teachers the most effective learning plans for the day. To note, these tests are still in progress at the IBM-sponsored preparatory schools in New York City, but they have demonstrated promising early results.
Watson went quite small as well – its suite of developer tools and lightweight service offering has allowed businesses and governments to build applications and systems engage with its core technologies. Watson now helps make decisions across dozens of industries, from powering the recommendations Apple’s health application delivers users to powering image and data analytics in emergency management for local and state governments. Watson now helps the National Institutes of Heath find cures faster as well, and select marketers and media companies now use Watson to rapidly segment consumers in real time, identifying opportunities to shift media placement and content in real time. Watson both helps sell cola and helps doctors develop interventions to combat the obesity it causes.
“Watson as platform” is a groundbreaking idea.
Transformative platforms have reverberating consequences as big ideas as well. As a kernel of an idea about how more than a century of industrial capability and foresight might produce humanity’s leading learning and thinking algorithm, Watson has had a generational impact in introducing cognitive computing to the world. IBM and Watson did not invent machine learning – they scaled it, made it relevant, public, funny, and portable.
In the shift from hardware giant global services company, IBM sought to remake its image as well. The process has taken multimillion-dollar advertising campaigns and the buildout of one of the country’s most sophisticated design and corporate development workforces, but I believe the payoff has been a force multiplier for what connected companies and institutions can achieve with and beyond their data.
Less than a year ago I checked in with Watson at Austin’s South by Southwest festival. Things are going well, with several thousand employees now working in New York and around the world in IBM’s IBM Watson business and health divisions. Machine learning is big business, and in the cultural and industrial embrace of IBM’s cognitive computing has saved lives and made the business of government easier as well.
My bet – and I believe the wager IBM has made – is that Watson as platform hasn’t yet hit its stride. While the IBM platform as a whole has helped governments seize untapped revenue, plan city events and deploy public bicycle systems during peak usage. The most interesting applications of the Watson approach are to come, where governments and institutions can unlock the value of billions of points of information; IBM faces the challenge of explaining the value cognitive computing can create and making the abstract highly tangible and effective for millions of business owners across the country in the decade to come. My, how Watson has grown.