I recall improvisational drummer and composer Michael Evans telling me a story of someone who had the opportunity to meet Cage and give him a record, and John Cage just smiled and said, "You know I have nothing to play this on?"

The danger of these collaborations across disciplines is in having too strict of a division of labor - in my case, of getting stuck doing the music. When I make an album, I write music, I write lyrics, I come up with the visual design, etc. I get to do all of that stuff.

There's a book of interviews with John Cage by Joan Retallack called Musicage that was finished the summer that he died, in 1992. And in one of the last interviews, he was very excited to talk about nanotechnology. There's real technophilia from him, a kind of utopian embrace of the idea that nanotechnology will free people up to do what they really want to do.

Records have always been the most extraordinary form of time travel for me, and that's why it matters to know when something was circulated, and if it had an audience of five or 50,000.

If the records that I make have one thing in common, it's that there is little recapitulation, and the idea is that it should end in a place very different from where it began, and that you've heard musicians undergo a change or be irreversibly transformed.

As there are more online archives of improvised music, it becomes more like the daily practice of playing it. It lessens the idea of there being masterpieces of improvised music through benchmark recordings.

What a strange thing - that musicians grant permission to places like Ubuweb, and then because it's free, it'll probably be listened to more often than something that is still wrestling with this idea of making a profit.