(Updated June 11th, 2012 after Apple's WWDC with news about Siri-Maps integration and the Eyes Free Siri feature.) Of technologies sporting the largest gaps between vision and reality, I'd rank talking car computers up there with videophones, laser pistols and flying cars.
Who wouldn't want to be zooming down I-5 while telling your personal KITT to book a reservation for two at the nearest four-star restaurant serving gazpacho? Or drilling into real-time sales pipeline data on a desktop-monitor-sized LCD touchscreen?
So I was shocked by Sam Grobart's article in the Sunday New York Times, "Touch, Tap, Speak: Taking 5 Connected Cars for a Spin," as it showed me how much state-of-the-art connected cars remain glorified GPS/stereo systems, and how much they lag their smaller mobile brethren, smartphones and tablets.
1) Voice control remains, for the most part, "comically bad" and rarely used, according to Grobart. That's because making the software smart enough to figure out what to do when you say "Call home" without having to navigate verbally through a series of menus and sub-menus (i.e. saying "phone," "call contact," "home,", "line 1," and "yes") is apparently pretty darn hard for most car computers. It's not just because Siri is so awesome and car computers suck. Road noise, blaring stereos and the distance of the microphones from drivers' mouths make accurate voice recognition very hard and costly in terms of processing power.
2) Alternate user interfaces are unsatisfactory in their own way. My visit last fall to the Tesla factory where I drooled at the Model S's 17-inch computing-based infotainment system got me excited about touchscreens in cars. There are several issues that keep touchscreens from cars from being as useful as they are on tablets like the iPad. Multi-touch capacitive touchscreens that allow you to swipe and pinch don't work with glove-encased fingers. A bigger problem is that touchscreens require users to look at them while they swipe. That's not practical while driving.
As a result, car makers have tried to augment or replace plain old touchscreens with touchscreens that vibrate and/or buzz to provide yes/no feedback (Cadillac's XTS sedan), joysticks that do the same (Lexus's GS 350), or a series of rotary knobs and buttons (the approach of German carmakers like Audi, BMW and Mercedes-Benz). And nearly all automakers with in-car computers offer buttons on their steering wheels. While handy in the midst of driving, it strikes me as being as crude as computers in 60s sci-fi shows that had no screens or keyboards, just switchboards with various one-control buttons and levers.
3) Think Android is fragmented? Think again. I don't know any developers at ISVs serving the connected car market, but I'm guessing they are a depressed bunch. There is one popular multi-car platform called QNX CAR that is reportedly in use by 200 vehicle models, giving it 50% of the market. The problem is that many apps need to connect to the real-time sensors and controls of the car. How one does that varies a lot by car, and probably requires a ton of fussy, low-level programming that would sap any developer's patience. Also, car makers like to customize the screen UI to fit their brand or the car model. And don't forget the idiosyncrasies of each car's physical UI, i.e. the vibrating buttons, knobs and joysticks. The net effect is that each carmaker effectively has its own platform.
Besides having too many slices, the pie itself is tiny. According to ABI Research, only 8.2 million connected cars will ship this year worldwide, growing to 39.5 million in 2016. Contrast that to the 1.8 billion smartphones that will ship this year, growing to 2.3 billion in 2016, according to IDC. Not a profitable scenario for would-be ISVs.
4) Innovation is excruciatingly slow. The lifecycle of a car is far longer than your typical smartphone or tablet (2 years). Os Grobart puts it:
"Automakers have a problem — the companies that make those smartphones and tablets are faster and nimbler, making significant updates almost every year. A car model may be in dealerships five years or more. If tech is Ferrari, the auto industry is Peterbilt."
What Siri and iOS 6 Offer
The slow pace of innovation is one big reason I really like what General Motors is doing. Rather than try to create another marginal, proprietary platform, GM wants to let you leverage your smartphone and/or tablet while driving. In its Chevy cars, it is building 7-inch in-dash touchscreens that give you control of your connected mobile device.
This is not only less expensive for GM, but it also lets it ride on the much faster innovation happening in the mobile space.
Take the iOS 6 update announced Monday at Apple's World-Wide Developers Conference, which will integrate Siri's voice control into an upgraded Maps and GPS app (the latter courtesy of TomTom). Drivers will be able to say "Take me to the nearest sushi place with 4 stars on Yelp" and have Siri find and automatically create the route to take you there.
You could argue most of these carmakers already get it. BMW, Mercedes, GM, Land Rover, Jaguar, Audi, Toyota, Chrysler, and Honda have all agreed to integrate a new feature called 'Eyes Free Siri' into their vehicles. This allows drivers to press a button on their steering wheels or dashboards to prime Siri to accept commands, such as search for destinations, accept dictation, write Tweets, etc.
Again, the uncertainty is how smart Siri will be able to understand what you want in these expanded contexts, and how well it will simply be able to recognize your voice in a noisy car cabin. Mounting a Bluetooth microphone above the windshield in front of the driver's seat would be the solution to the latter.