Once I started working with generative music in the 1970s, I was flirting with ideas of making a kind of endless music - not like a record that you'd put on, which would play for a while and finish.

Except in a few cases like Music for Airports, which was a very clear case of noticing a niche [and] saying, "Okay, there's this situation in which people always play music, and nobody has written music for that situation so I'm going to." So, that was a very clear example of spotting a niche and working for it. I have done that occasionally.

We have two different ways of working. One is completely unstructured where somebody just starts playing and somebody joins in and then the other person joins in, and something starts to happen. That's occasionally what happens. What more often happens is that we settle on some sort of - a few sort of structural ideas, like, "Okay, when I put my finger up, we're all going to move to the extremes of our instruments. So, that means you can only play either very high or very low or both. And we're going to stay there until I take my finger down.

I have a definite talent for convincing people to try something new. I am a good salesman. When I'm on form, I can sell anything.

I think that's very significant that we're so attached to the idea now of - it was something I advocated for years, that you can make music in studios, music doesn't have to be made as a real-time experience. But now you see the results of that in people who are completely crippled unless they know that they have the possibility of "cut and paste" and "undo." And "undo" and "undo" and "undo" and "undo" and "undo" again.

Complexity and intelligence grow from simplicity, not from greater complexity.

The dominant theory coming out of Hollywood is that peoples' attention spans are getting shorter and shorter and they need more stimulation.