I think the least important thing about science fiction for me is its predictive capacity. Its record for being accurately predictive is really, really poor! If you look at the whole history of science fiction, what people have said is going to happen, what writers have said is going to happen, and what actually happened - it's terrible.

I assume that - because you can get degrees in journalism from very reputable universities - I assume that people can be trained to be journalists. I've never been entirely certain that anyone can be trained to be a novelist in the same way.

Addictions [...] started out like magical pets, pocket monsters. They did extraordinary tricks, showed you things you hadn't seen, were fun. But came, through some gradual dire alchemy, to make decisions for you. Eventually, they were making your most crucial life-decisions. And they were [...] less intelligent than goldfish.

The deadliest bullshit is odorless, and transparent.

I knew where the magnets were, behind the gyprock, and the magnets were very powerful. I think they had to be powerful for me, otherwise the reader wouldn't have a reciprocal experience. But I was very careful to bury them deeply, deeply in the plaster and paint over them. I didn't want anybody to directly access them, and that's gradually changed for me.

That's something that tends to happen with new technologies generally: The most interesting applications turn up on a battlefield, or in a gallery.

We're living in a future that's weirder than anybody except possibly.