Writing

Feed Software, technology, sysadmin war stories, and more.

Tuesday, November 20, 2012

My future iPhone prediction: local "Siri"

I want to make a prediction about a future iPhone. This one is inspired by the fact that it seems to be so ridiculously slow and finicky when it comes to things like Siri. The original 'voice control' was fast and easy, and seemed to get it all done locally. Siri may have added a bunch of new features but it also took away a lot of reliability and added tons of latency.

I understand there are other services out there which do the whole voice transcription far better and on the same hardware, too. If this doesn't light a fire under those guys, nothing will. I suspect if they manage to ramp up, they'll do it by putting yet more CPU power in their phones, and then using that to really crunch through voice processing locally.

This will have a bunch of effects. First, it will only work on the Shiny New iPhone 9, so anyone with the 8 or 7 or whatever will be left in the cold (as usual). Also, since Siri as it exists now works as a server-side process, they will just retire those servers and then things will break for existing phones.

This is not much of a stretch. For precedent, just go look at a Mac which is still running Snow Leopard and see how it copes with MobileMe being gone. There are bits and pieces where it just gets confused about connections which can't be established or syncs which fail. This is something which used to work but works no longer because the server-side rug has been permanently pulled out from under it ... and burned.

That's my prediction. Hopefully, I turn out to be wrong on this one, but given their recent behavior, I can't bring myself to expect anything better.