I was reading last week’s issue of the Economist (I don’t always agree with them, but their analysis is quite enlightening) including their technology quarterly, and a really interesting thought struck me. The issue was rife with fascinating advances, but an overarching pattern emerged.
Let me set some context: in elearning we talk about blended learning, and I regularly say it’s not about learning, it’s about doing. Really, the view of elearning I’m trying to propagate is more like blended doing, where technology partners with us to make a more effective problem solver. This comes from Don Norman‘s point that from the problem’s point of view, a human augmented by technology is a more fearsome force than just the human alone.
Look, our brains are good at pattern matching (see above) and big picture, but bad at remembering rote things and details. Technology is just the opposite! That’s what performance support is all about. Of course, sometimes our brains need major skill shift changes, and there’s a role for courses, but some times we need information, and sometimes we need people, and sometimes we need support for massive computation.
Jim Schuyler has a mechanism he uses to authenticate comments on his blog. With Captcha, it’s the familiar challenge/response where you type in the familiar image of letters, with a twist. There’re two words, and while the first is known to the system, the second comes from some OCR text with a word it’s not sure of. So not only are you showing you’re human, you’re assisting the digital archiving of some important text.
The Economist mentioned this type of blending that gets people to do a difficult task for computers in what otherwise is a computing intensive task. They mentioned the ESP game where people playing also get some work done, in this case tagging images on the web.
They also talked about evolutionary algorithms (I first learned about them through John Holland’s work on genetic algorithms at UMichigan) to design things. You match a design problem to a set of parameters that then try to evolve to solve the problem, using mutation and selection to populate the solution space. Holland et al were looking to match how the brain works (getting solutions similar to those with neural nets, but with a less directly mappable approach), but others are just attacking design challenges.
Where a human would get massively bored searching through every permutation, this approach turns it into a computation problem and the computer merrily works away. There’s no guarantee of solution, but it really gets into situations where brute force can work and elegance might not. They mention some great outcomes, including better wireless antennas, optic cable designs, and more.
The point being to consider a different point of view, not of the performer, but of the task, and what’s the best solution to achieving the goal? A clue: if it requires rote-memorization on the part of the human, particularly of a set procedure, it probably should be automated. Let computers do what they do well, and let us do what we do well!
So, think of the tasks your performers need to accomplish, and don’t be afraid to think out of the box when you look for solutions. I’ll suggest that what will make a difference going forward isn’t focused on knowledge, but on problem-solving and innovation.