In A Little Duplex Skepticism, John Gruber says what I’ve been thinking about the Google Duplex demo:

It’s totally credible that Google would be the first to achieve something like Duplex, but the fact that all they did — as far as I’ve seen — was play a recording just seems off. It feels like a con.

I’ve only heard a bit of the “booking a haircut“ recording on a podcast. I thought it sounded a) impressive if real, but b) very possibly fake.

That kind of technology will come, eventually; but are we that close to it today?

(If we are, then whether or not we want it to be used in the kind of way demonstrated, is a whole other question.)

Google, of course, gave no timelines, no suggestion of when such a feature might be available. Given that, it makes you wonder why they even bothered to demo it.