July 19, 2016 | Written by: Noah Syken
Categorized: Data | New Thinking
Share this post:
From our fingerprints to our irises to the very DNA that directs our growth and development, we are singular, solitary, unprecedented individuals. Why, then, does the world insist on training us all the same?
We teach our school children using standardized curricula. Doctors ply their trade based on “best practices.” And we train our athletes to swing a bat, throw a football, and shoot a jump shot in the same repeatable, predictable way.
We do not do this because we believe all people are the same. Nor do we believe there is only one best way to learn in school, or practice medicine, or swing a baseball bat. We do it because we have to. To train every person in a unique and individual way would introduce too many variables and require too many resources. Uniformity of instruction affords us certain efficiencies of scale. To consider any other way would be impractical to the point of impossible.
Or would it?
When I learned how to swing a golf club, I was taught the same things everyone else is taught: lock your pinky and your pointer, keep your left arm straight, head still, shoulder turn, open your hips, finish high. We are taught this way because on average, and over time, this technique has produced results. In other words, it scales.
But reflecting on the announcement of IBM’s new partnership with Sesame Street, in which Watson — the world’s first cognitive computing system — will be crafting individualized curricula for early childhood learning, I got to thinking: what if my optimal golf swing is nothing like the one I’d been taught?
Here’s what I mean: Watson is customizing educational programs to a child’s skill level and learning style. More than that, it’s not limited to what has been tried before. It has no preconceived notions about the best way to teach. It has no biases. It can try out entirely new combinations, and adapt its content to each person that’s using it.
What if Watson could do the same for my golf swing? What if Watson could tell that for a man of my height, weight, age and body type, with slightly impaired vision in my left eye, an old shoulder injury and a tendency to pronate, my swing would benefit from a flatter backswing and an abbreviated follow through?
Sounds crazy, right? Well ask Jim Furyk if there’s only one right way to swing a golf club.
I’ve seen Watson do this kind of thing before. Last year at SXSW, Watson conceived of recipes that master chefs never would have dreamed: Peruvian potato poutine; Belgian bacon pudding; Vietnamese apple kabob. All unprecedented (as far as we know). All delicious.
But these kinds of suggestions would be nearly impossible for classically trained chefs to imagine. Our biases are too deeply ingrained, passed on from generation to generation. Our thinking is too rigid. And traditional programmable software is no help here, because it requires programmers — aka “people” — to imagine the possible outcomes, the if/then scenarios. Cognitive systems are not limited by our thinking. They learn on their own. Big difference.
I’m not saying that baseball players are going to start swinging the bat backhanded. Or that basketball players will favor the skyhook over the jump shot any time soon. But I am saying that training — both for skill development and physical fitness — can and should be much more customized than it is today.
Because success in sports — or business — is not one-size-fits-all. And I am looking forward to a day very soon in which cognitive computing supports a new generation of athletic achievement, one with fewer injuries and longer drives.