Last week Apple announced the iPhone 4S.  Among some nice upgrades, the marquee feature in this year's model is Siri, the voice assistant.

It's easy to dismiss Siri as nothing more than a novelty.  Voice user interfaces have existed for a long time and have always been a let down.  It's more likely than not that Siri will follow this pattern.

But what if it doesn't?  What if Siri is at least good enough for real-world use?  Apple has a good track record in mainstreaming emerging technologies.  The mouse, GUI, portable music players, smartphones, touch UI, and the tablet--all were more or less niche products or features that Apple brought to the masses.  And they did it by making them "just work."  It's possible that Apple has finally progressed voice UIs to the "it just works" stage.

If so, the implications are HUGE.  The use cases shown in the demo are compelling: setting timers, checking and making calendar appointments, reading text messages, looking up the weather, etc...  But that's only the beginning.

Imagine if Apple creates an API for third-party developers to tap into the voice actions?  You can book movie tickets on Fandango, purchase Kindle Fires on Amazon, send the open web page in Safari to Instapaper.  And if the AI gets good enough, perform combo actions like reserve an 8pm table at Slanted Door and automatically set up a calendar event.

Apple will also likely expand beyond the iPhone 4S and make Siri available on its full line of connected products: Mac, iPads, iPhones, and Apple TV.  I wouldn't be surprised if they made Windows PC and Android Siri apps.

And Siri will learn fast.  Everyone's voice commands will funnel through Apple's iCloud servers, powering the intense computations need to make "it just work," becoming more accurate everyday.

Who wins?  We do.  So does Apple of course.  Who loses?  Maybe Android and Windows.  Siri, powered by iCloud will create a serious moat between iOS devices and the competition. But the biggest loser will almost certainly be Google Web Search.  

Web search is just a proxy for asking a question.  If a voice query can be made quick and efficient, Apple can deliver answers without going through Google.  Yelp and CitySearch can power restaurant and local business searches. Amazon and other etailers can power product searches.  The "where was Ricky Martin born?" questions can be answered directly from Wikipedia.  

Google makes almost all of its money from AdWords displayed in search results.  Android exists to establish Google Search on mobile devices and maintain the dominance it has on PC web searches. Not all queries can be answered in one hop, but if 10-20% of your searches go through Siri, Google will feel it.  It will have trouble growing its topline as more web activity moves to mobile devices and increased competition from Siri's voice queries.

But it all depends on how well Siri works.  I'm not getting a 4S, but i'm going to follow closely how this new technology plays out in the real world.