My “Girlfriend”
October 14th, 2011. Mark that date. I believe history will treat it as the day we began using speech to interact with our computers. It is the day Apple released the iPhone 4S with Siri, the automated assistant built exclusively into the 4S. Before I elaborate, a brief disclaimer. I am an unapologetic geek and a huge Apple fan. Having said that, I honestly believe I’m being objective when I predict Siri will be viewed as a transformative technological innovation.
Apple describes Siri as an “intelligent assistant”. After using it for just three weeks, I believe their description is accurate. We’ve been able to speak commands in varying degrees to our computers, cars and smartphones for some time now. So, what’s new and/or different about Siri? A lot. First, Siri doesn’t restrict what you ask or tell it. For example, my car is a 2006 model and uses voice input to dial phone numbers, set the A/C temperature or change the radio station, but I have to remember the exact commands and speak them clearly (and in just the right sequence) for the system to work. Not much intelligence there. Siri, by contrast really seems to understand the English language. For instance, I can ask Siri to “Tell Katherine dinner is at 6” and it is smart enough to know I want to send Katherine a text message. Siri responds, “I have created a new text message for Katherine” then displays “Dinner is at 6″ and asks, “Would you like me to send it?” I simply say yes and the text message is sent. Not extremely groundbreaking, I agree, but if instead of saying “yes” I said “Change that to 5:30“, Siri is smart enough to change the text message to read, “Dinner is at 5:30″! That’s my second point. The artificial intelligence built into Siri understands commands contextually. Siri knew I wanted to change dinnertime. It didn’t try to send a text message to “5-3-0″ or create a new message, “Change that to 5:30″. It responded like a human would. This is far beyond any voice recognition system I have ever used and light years ahead of what the Germans programmed into my sedan.
Siri also remembers relationships. I told Siri, “Ann Kelley is my wife” and after a confirmation, Siri remembers it forever or until I change it (I really hope I never have a reason to do that!). The same goes for all my daughters. So, when I ask Siri, “Tell the family my flight arrives at 8“, it creates an MMS to my wife and three daughters, “My flight arrives at 8″ and sends the message to all four. The same works with other relationships or nicknames like “boss”, “secretary”, “butt-head” and “landlord” or just about anything you can think of. Along the same vein, with a free iPhone app called “Find My Friends” I can ask Siri “Where is Katherine?” and up pops a map with the location of Katherine’s iPhone. I naturally assume that’s where Katherine is because if she gets more than a few feet away from her iPhone she has nervous tremors.
Siri is also very valuable when you travel. I was in Corpus Christi last week for business and didn’t know my way around too well. When I got hungry one evening I asked Siri, “What’s a good Mexican restaurant nearby?” and a list of Tex-Mex joints popped up on the screen complete with distance to, ratings and price ranges. Since I asked for a “good” restaurant the results were sorted by rating. Out of curiosity I rephrased the question to “Where is the closest Mexican restaurant?” and the same list appeared, but this time it was sorted from nearest to farthest. After dinner I asked Siri “How do I get back to the Hampton Inn?” and a map popped up on the screen with pushpins showing my current location and the hotel plus two different highlighted routes to get me there. While I was driving back to the hotel I received a text message. I asked Siri, “Read the text message” and Siri replied telling me it was from Robert and read the message, “When would you like to meet tomorrow?” I told Siri, “Reply 8:30 AM” and Siri sent “8:30 AM” in response and my hands never left the steering wheel. Next I asked Siri to “Create an appointment with Robert for 8:30 tomorrow morning.” and after a confirmation the new appointment was added to my calendar. When I was ready for bed back at the hotel I asked Siri to “Wake me up at 6:30“. Siri responded “I have set an alarm for 6:30 AM, Joe“, and did so. (Siri will refer to you using any name you wish. I tried “Your Highness” first, but that just sounded weird.)
Another thing that makes Siri so amazing is there’s no training involved. I didn’t have to memorize a list of things I could ask (like my car system made me do) and I didn’t have to train Siri to recognize my voice like Microsoft’s voice recognition software. I just picked up my iPhone and started using it. Does it make mistakes? Sure. Nothing’s perfect, but it is astounding just how accurate it is. Unscientifically, I’d estimate Siri gets it right about 75% of the time and the more I use it, the better it gets. I think that’s because I’m getting better at communicating with it and it is learning what I expect. I’m also learning that a good network connection is absolutely essential. Since Siri’s brains reside on Apple’s servers in North Carolina, a good 3G or Wi-Fi signal is necessary to send your requests. The only thing happening on the iPhone itself is the recording of your voice. That recording is then sent to Apple’s servers where those really powerful computers interpret your recording and formulate a response. That’s why Siri is so effective. With a good network connection Siri is also very fast. Even the most complex request takes only a second or two. Siri even understands Texan quite well (something I’ve found to be problematic with other voice recognition software). My wife jokes that I sound like I’m talking to a girlfriend when I converse with Siri (who’s automated female voice is just short of human). That’s the way it is, like a very natural conversation.
Currently, Siri is officially “beta” software. That’s geek speak for “not fully baked”. It also means that in addition to admitting bugs, Apple is just getting started with this technology. Some tech pundits predict Siri will be implemented on all Apple devices in the near future. I tend to agree. Given how well Siri works today, it’s bound to be completely astonishing a couple of years from now. Don’t misunderstand, we’re still a long way from HAL 9000, but Siri is the first voice recognition application I’ve ever been tempted to say “Thank you” to.

As the netbook form factor exploded onto the computing scene last year, Linux gearheads everywhere rejoiced in the notion that they finally had a toe-hold into the mainstream computing world. Hardware manufacturers almost universally offered price-leading netbooks (many <$300) with some flavor of Linux pre-installed. Unsuspecting bargain hunters sucked them up but quickly realized none of the software they wanted (or already owned) would run on their new, tiny laptop and the familiar buttons on the screen didn’t do the same thing as the ones on their other computers. Consequently, Linux-based netbook manufacturers saw a 35% return rate. The open source community got their collective bubbles burst when netbook owners shunned their beloved kernel in favor of the six-year-old Microsoft retread, Windows XP. Redmond happily obliged by offering the home version of their old reliable OS to netbook makers for thirty bucks, effectively jabbing the entire Linux community right in the nose. Sure Linux is free, but $30 is really close to free and certainly worth it (for most users) to avoid learning a new operating system.
Anybody remember Friendster? How about AOL? Twitter? Bingo! I’m sure that last one rings a bell, even to the non-techies of the world. You can’t tune into a news channel these days without hearing the word tweet every few minutes. I know when my 56 year-old minister asks me about Twitter that its’ cool factor is nearing an end and the technorati will soon abandon it for the next big thing. Being of considerable years myself (53), and having spent the past thirty years in the computer field, I’ve observed and participated in this technical upheaval with more than a passing interest. As I’ve watched the crowds moving from AOL to Friendster to MySpace to Facebook I’m reminded of a similar phenomenon of the non-technical social networks, nightclubs. Although I haven’t been a “clubber” for many years, I have kids in their early twenties and their recounts of social behavior don’t differ much from my own recollection of post-pubescent revelry.
A friend of mine sent me an e-mail asking how my WWDC experience was. I replied: “It’s like drinking from a fire hose!” That response may have been a bit melodramatic, but I still stand by it. The conference presents so much information so concisely its difficult (if not impossible) for a mere mortal to absorb it all. If you’ve ever seen a video of an Apple keynote you have an idea of how polished their presentations tend to be. Even though the subject matter was different, every session I attended at this year’s WWDC was put together that well. Instead of talking about new Macs or iPhones like in a keynote, the session presenters talked about new code constructs, OS X Leopard features or any of 100+ different subjects at varying levels of detail and complexity. The labs were done just as well. The first floor of Moscone West was over half lab space, filled with enough new Apple hardware to make any fanboy think he was at the North Pole with Santa Steve and the four labs I attended were staffed with knowledgeable, helpful engineers. As is customary with Apple, nothing gets done half way.
Recent Comments