Just over a year ago Google presented WaveNet, a new deep neural network for generating raw audio waveforms that is capable of producing better and more realistic-sounding speech than existing techniques. It’s gotten a LOT better and is now capable of producing natural sounding human voices.
Google is using it for Google Assistant but hard for an old radio guy like me not to imagine this tech replacing radio announcers (are there still radio announcers?)

Assuming I’m not overlooking something, the oldest doc I have is from April of 2010. Google Drive launched in February 2007. This would mean I didn’t use Google for the first few years. Can’t imagine why but that’s possible.