> Are you optimistic we can get speech synthesis to run on mobile devices in the near future?

You mean high quality right? I mean speech synth has been around for decades that can run on cheap hardware and is understandable. Speech recognition has also been around for a long time, but there's a huge difference in usability between "pretty good recognition" and "pretty good synthesis". One is useful, the other not so much.