This month featured the launch of an amazing new service that, potentially, could change the way you interact with your phone, with data, and with the world at large. (No, not Buzz. Julie will talk about that tomorrow.)
Siri is an iPhone app that works on the following model: You tell it what you want, and it makes it so. It can locate nearby businesses, make reservations, buy tickets, and more, all with a context-aware voice recognition interface.
Here’s the pitch (it’s worth sticking around until the end):
Now, it is a brand-new release, so it’s still learning. (I mean to say, new sources of information are being added to its service.) And, like any such service, it’s only as good as the data it draws on. When I asked for a nearby grocery store, it couldn’t find the Stop&Shop right down the street, although it *did* find the florist division within that store. Also, the conversation-based model for search isn’t great at search limiters: If you follow up “”where is the nearest grocery store” with “not a convenience store,” Siri blows right past the “not” and starts throwing up 7-11 locations all over.
But still! What’s terrific about Siri is how simply it parses your request, and quickly pulls together relevant, often actionable, information from quite disparate sources. It’s not that Siri is really finding anything out that you can’t find on your own; rather, it takes a multi-step search process and turns it into one, simple, voice-driven request. (And I agree with Michael Sippey, that Apple “should just buy the company and turn this into the default Voice Control app.”
To understand how this works, and why it’s both awesome *and* worthy of a ProfHacker post (as opposed to say this, which is just awesome), it’s worth thinking back to Julie’s introduction to Application Programming Interfaces (APIs) from the fall (parts one, two, three, and four). APIs are protocols that let you access and manipulate data, or even add new features to a web service. APIs are what let you mash together data sets, use a Twitter client, and more.
Don’t get confused by the awesome voice recognition engine that figures out your speech and what you want with pretty good accuracy. No, that’s not the really cool thing, although Microsoft and other companies have been working on natural language search for many years now and have been failing to come up with anything as useful as Siri.
No, the real secret sauce and huge impact on the future of the web is in the back end of this thing. A few months back the engineers at Siri gave me a secret look at how they stitch the APIs into the system. They’ve built a GUI that helps them hook up the APIs from, say, a new source like Foursquare, into the language recognition engine.
(EDITED TO ADD: Correct link to Scoble’s page. My bad.)
He’s also got a great interview with the Siri folks explaining what’s going on:
Siri is a free download from the App Store. (Versions for other phones coming soon.)
(In part, Siri reminds of how much I liked the late, lamented Sandy, which isn’t really the same sort of service but which was also billed as a virtual assistant.)