The peche de cher of the moment.

A long conversation today revolved around complexity and composable systems; i.e. you can solve some complex problems by breaking them down into smaller, more tractable sub-problems. Some tasks have large dimensionality, so it's not clear that we can make a composable model from a subset of the dimensions.

Still, it's useful to assume that you can make such a model, and then to try to develop one. The process of exploration may bring additional understanding; in the worst case, you can write up your woes on the 'net and allow someone to avoid duplicating what didn't work. In the best case, you discover patterns and find that with some basic rules, you can build comprehensible complexity.

Who really cares, though? Science advances and there's not a whit we can do about it. Well, we immediately care when technology directly affects our ability to earn a living.

The gravest threats to our livelihoods may not come from direct automation of our tasks, it may come from the decomposition of our work-tasks into cheap usable building-blocks. We may find our jobs being done by relatively unskilled people with basic composable tools who know roughly what they want. We've already seen this happen to travel agents. Tax accountants, general practice doctors, and our own jobs may be next.

Sometimes I feel like that proverbial fish that first crawled around on land because it was too dang dangerous to be in the water.


http://online.wsj.com/article/SB124648494429082661.html - this seems germane to your post. There is a lot of it I don't understand, but the concept of breaking big problems into subproblems I get. --Andy