Folk around here talk to Scott reasonably often. In my experience, he hasn't been that interested in the core issues you were talking about. A generic tour of computational complexity theory would seem to go in the same category as other relatively peripheral talks e.g. on quantum computing or neuroimaging technology. You're right that the philosophy and computer science stuff he has been doing recently might naturally lend itself to a more "core" talk.

Any others?