That's with one microphone and presumably a single speaker. I wonder how well one could do with several microphones around the room and several speakers, with their relative positions known accurately and some computing power to generate the best possible reverse sound distribution.
I reckon the attenuation you can get will be something like the cosine of the phase difference between the speakers. Basically you'd need to them to be much closer than a single wavelength, otherwise they'll amplify the sound, rather than attenuating it.
I reckon you could probably do some extremely local noise cancelling (which would only work for a few people at most), but then the speakers don't only need to balance out the external noise, but also each other, which could require extreme volumes if the distributions of speakers is a bad fit.
And it's a much harder problem, because antiphase sound from multiple sources will interfere with itself and probably also interact with room resonances, potentially making the problem worse in some locations.
