If I have my iPad and iPhone in front of me, both with "Hey Siri" activation turned on, both will register that I asked. But only one of them will respond 'verbally'.
I'm not sure if that's what the OP meant, though.
I'm not sure if that's what the OP meant, though.
The article implies that it will fuse voice data from multiple devices which both only heard part of the query into a single query, which is a lot easier to write down in a patent than to actually make work.