Wired Magazine has a fascinating article about playing chess. They point out that IBM's Deep Blue was able to beat Gary Kasparov, then the world champion by using brute force. Not particularly interesting or relevant to the real world. But Deep Blue's feat spawned a new hobby of using off the shelf software to compete in computer enhanced chess tournaments.
What the players found was that it wasn't the best chess players or the players with the best chess playing software that won these hybrid man-machine tournaments but the players that best integrated their chess skills with their automation. It turns out that one has to know when to take the software's advice and when to reject it. It was the integration between the automation and the expert that made the most powerful match.
This has profound implications for how we design software. Quite a bit of software seeks to replace the business analyst or decision maker with codified business rules that 'always' do the right thing. But all systems have exceptions or areas at the limit where the decision tool is not robust. Therefore the key to competitive advantage in these highly automated decision environments may end up being in the expert human - computer interface. In other words, in having humans with the expertise to know when the system doesn't maximize outcomes and to intervene accordingly.
Training experts to override rather than simply operate the software. Something that knowledge based businesses need to think about while riding their cyborg hovercraft out on the Openwater.
No comments:
Post a Comment