This all reminds me of "AutoPC" from Clarion in the late 90's[1]. A friend of mine had it in his truck, it truly was amazing at the time -- he could make phone calls, play music, render maps on his display, you name it, all hands free on a tiny little SH-3 processor running windows CE (yes, remember that?!) under his seat.
It's sad it really hasn't progressed very far -- Essentially, we've moved on to cloud-hosted audio analytics to handle better speech recognition, but the usage of the thing really hasn't changed.
That sounds a really interesting device, thanks. I too have noticed that voice recognition has been shifted "to the cloud" despite the improvement in local processing power; even 15 year old Nokias had voice recognition of some sort (BLUETOOOOTH you would shout), and old single-core PCs had Dragon Naturally Speaking working on them; I am sure modern devices have more processing power (and storage).
Interesting, I didn't know that AutoPC had actually gone anywhere. I was at a WinCE Developer Conference in London where it was presented, and the voice recognition completely failed to work on-stage in a large, hushed auditorium. Given that it was supposed to work in a car, with road noise, it didn't impress anyone (that I could see).
It's sad it really hasn't progressed very far -- Essentially, we've moved on to cloud-hosted audio analytics to handle better speech recognition, but the usage of the thing really hasn't changed.
[1]: http://news.cnet.com/Clarion-to-build-AutoPC-using-MS-softwa...