On the molecular scale, you find it's reasonable to have a machine that does a million steps per second, a mechanical system that works at computer speeds.

Today we have big, crude instruments guided by intelligent surgeons, and we have little, stupid molecules of drugs that get dumped into the body, diffuse around and interfere with things as best they can. At present, medicine is unable to heal anything.

But if we can manage it so people don't have things forced on them that they don't want, I think there's every reason to believe things can settle out in a situation that is recognizably better than the one we're stuck in today.

The basic parts, the start-up molecules, can be supplied in abundance and don't have to be made by some elaborate process. That immediately makes things simpler.

Protein engineering is a technology of molecular machines - of molecular machines that are part of replicators - and so it comes from an area that already raises some of the issues that nanotechnology will raise.

If you take all the factories in the world today, they could make all the parts necessary to build more factories like themselves. So, in a sense, we have a self-replicating industrial system today, but it would take a tremendous effort to copy what we already have.

I had been impressed by the fact that biological systems were based on molecular machines and that we were learning to design and build these sorts of things.