I'm interested in this avenue, but I think Ron missed an important first step. Rather than speech recognition I'd like to get my phone to do speech synthesis. Image the app: you're driving home on your daily commute when you phone tells you that it has been monitoring the traffic reports on the web and suggests that you take an alternative route. I have a feeling I've seen an API or library for the nokia/symbian python that lets you generate a voice. I think I'll take a shot at a spoken "Hello World" app.