Hooray for dumb software

T’other day I said:

In fact, I’d argue that I very much do not want “intelligence” in any software I’m using or even “actionable social knowledge”.

The short reason as to why I don’t want my software to be intelligent or have “actionable social knowledge” is that I don’t believe that social problems require technological solutions. Even in “social software”.

Making software “smart” typically involves embedding it with social protocols. Or rather, what the designers of the software imagined the appropriate social protocols to be. The reality of how people live in the world in so incredibly nuanced that even the best-designed embedded social protocol will necessarily be incomplete and thus annoying. And most embedded social protocols are not best-designed.

Rather than attempt to design good social protocols, or even good-enough ones, good software should allow people the chance to do what they do best: be people. People are social protocol negotiating machines. In order to allow people to be social protocol negotiating machines, good software should give people the information they need to perform social negotiations. (Toni Robertson from UTS has an elegant paper on this subject.)

In the end, it may be harder to build software that allows people to be people. But ultimately such software will be taking playing to the strengths of people and computers, rather than attempting to give computers a little bit of human-ness.