"For example, you try to make a jet engine. There are lots and lots of different variables,the operating temperature, the materials, all the different dimensions, the shape. You can't solve that kind of problem all in one go, it's too hard. So what do you do? Well, one thing you can do is try to solve it step-by-step. So you have some kind of prototype and you tweak it, you test it, you improve it. You tweak it, you test it, you improve it. Now, this idea of marginal gains will eventually get you a good jet engine. And it's been quite widely implemented in the world. So you'll hear about it, for example, in high performance cycling, web designers will talk about trying to optimize their web pages, they're looking for these step-by-step gains.
"That's a good way to solve a complicated problem. But you know what would make it a better way? A dash of mess. You add randomness, early on in the process, you make crazy moves, you try stupid things that shouldn't work, and that will tend to make the problem-solving work better. And the reason for that is the trouble with the step-by-step process, the marginal gains, is they can walk you gradually down a dead end. And if you start with the randomness, that becomes less likely, and your problem-solving becomes more robust."
And: "Just because you don't like it, doesn't mean it isn't helping you."
- Tim Hartford, economist, from his TED Talk, "How Frustration Can Make Us More Creative"