Monday, May 30, 2016

The failure of logic

Some time ago, I read a book about experiments on brains, You Are Not So Smart. As it turns out, many of the ideas humans commonly have about their own thinking capabilities are pretty far off the mark because our brains in some way work with illusions.

For instance, human eyes only have a small field in the center where the resolution is good, yet when we open our eyes, a wide field of vision appears. Because the images our brains see is something it has composed itself.

One of the recurring themes was logic versus intuition.

The point was that people mostly think with intuition.

Intuition is powerful. It lets you reason from the cumulative experience you've had. It works effortlessly and often only takes moments to do its job.

Logic is extremely limited. Like a simple computer processor, you can only work with a limited set of concepts at the time and make simple deductions.

People seem to think they make rational, logical decisions. But when you actually test them in experiments where intuition and logic disagree and logic has the upper hand, then most people don't wait and think through the deductions. So they arrive at the wrong conclusions and aren't actually as smart as they think they are.

But it has occurred to me that in some instances, you can see the opposite problem. People trying to use logic to solve problems that require more than simple deductions.

For instance, how do you use logic to decide what you're going to do in life? What education, where to work.

In the software world, an eternal question is what software to bet your future on. It affects end-users as well as developers. In the office next to IOLA, there once was a company that was betting on a particular technology offered by Microsoft. It was even embedded in their name. From the outside, they appeared prosperous. Then one day Microsoft decided to shut down their framework.

How do you decide which programming language framework will serve you well for many years?

I think you can use logic to elucidate some points. Likewise, you can try things out to gain more knowledge. But overall it is a huge trade-off with many factors.

Some people realize that and try to split the hard decision up. So you might look at it from different angles and make up a matrix comparing all the choices.

Helpful as it may be in understanding the problem, it can lead down a path where people spend their time arguing about unimportant matrix details. And someone may get the brilliant idea that all that's needed is a scheme for assigning scores to the various parts and then combining them with weights into a final rank.

This is, I think, the ultimative failure of the faith in logic - thinking that simple-minded strategies will beat the raw power of a trained human brain. How do you assign the weights so that the end result is meaningful? Nobody knows.

If you need a complex ranking, feed the details to a trained intuition, one that has seen what works and what doesn't, and let it decide.