A lot of time mistakes are very interesting - you look for the behaviour that's not the one you expect.

When I began to think about the head of the family, the storyteller, the rise of television which became the new storyteller, the break-up of the American family as an idea and then Avalon came.

I got a chance to work with Mel Brooks on two of his films: Silent Movie and High Anxiety.

I got involved with an acting school and studied for a couple years. They used to have improv exercises that you would work on and you would do improvs.

Different reactions while film test screening doesn't mean even the audience thinks ambiguity is a bad thing. But if you're asking them right away to start checking things off, they don't know what to do. I think at their best, it applies to when the audience knows what it is. Then, when they say, "Oh, well, I thought it was too boring in blah-blah-blah part," then you better pay attention to it. It's like going for the hamburger. Better be the good hamburger I went for.

If you test Iron Man and that audience doesn't respond well, you can be damn sure that there is something wrong with the movie that you have to address. Because they're expecting a certain amount of action, right? They want a hero. There are certain things that have to be compatible with the way the audience is thinking about it. If you take some other film, like No Country for Old Men, you can end up with all kinds of crazy reactions.

All I try to do is create an atmosphere that seems comfortable enough, that it removes tension and everyone feels free. If they feel free then behaviour happens, small moments happen and that's what ultimately works the best for me.