What is a Car?

Some years ago, I was on a modern bus (an Airporter bus from San Francisco airport to Marin County). As the bus cruised at about 30 miles per hour on a San Francisco city boulevard, the engine suddenly cut out, the lights went out. The driver steered the suddenly powerless bus to the side of the road, and brought the bus to a halt. He pushed open the door and walked around, to the back of the bus. Returning to his seat, he restarted the bus, the lights came on. What had happened? The bus had crashed, he explained, not into another vehicle, but its operating system had failed. He’d gone to the back of the bus to, quite literally, reboot it. (Windows CE, I believe.)

A week ago, on literally the same road, I passed a Google driverless car. This is how far the merging of information and automotive technologies have come. How far, that is, to this point.

We have reached the point where the definition of a car is changed.

You buy a new car now, and (presumably unless it’s a low-end vehicle) most of the owner’s manual deals with the IT and user-interface issues. The most complex new issues in learning your new vehicle will include how to interface your specific phone with the car’s IT systems.

We’ve long passed the point where computers just aid and optimize the performance of cars as they were before the addition of the computers. Increasingly, a car *IS* a computer, albeit a specialized computer with wheels and seats. A computer that will know, better than its human occupants that there’s a backup ahead of the traffic lights, a half-kilometer ahead and over the hill, and will adjust its speed or lane choice accordingly.

As the definition of a car changes, the obvious question is: "What, then, for the automakers?" The core technical competences of the automakers have for a century or so been an evolving skill set centered in the design and manufacture of the internal combustion engine and powertrain, aided by market-side skills in visual design and brand management. Can these firms survive into a destiny in which these skills are subordinated?

The fundamental business idea here is simple: that electronics IN the car now will change the nature OF the car in a way that has not previously happened, and that these changes may be so profound as to change the nature of the car-making business.

Is this really true? It is, of course, evident that electronics have been taking an increasing role in autos:

  • First, of course, via entertainment electronics (car radios) - automakers did such a bad job of these that, for years, car buyers would take their cars to after-market firms to add better radios (and cassette players, and speakers and the attendant wiring). Now only a fringe of high-volume enthusiasts go this route, since car makers long ago adopted good sound system products and practices.
  • Then, car electronics adopted integrated circuits, embedded O/S, and so on, in a place far less visible to the customer - in engine and transmission control. This was a process wedded to the evolution of more efficient cars, and driven largely by the automakers and their electronics vendors.
  • Third, more recently, we've seen GPS / SatNav systems. These started as after-market add-ons from firms like Garmin, Magellan, and TomTom. These firms innovated with detailed and frequently refreshed databases of streets, routes and distances, and with good step-by-step navigation, read aloud to the driver (in selectable voices). The database complexity always suggested that fewer firms would develop GPS / SatNav systems than make cars, and the long lead-times of car electronics make updates challenging. The outcomes seemed to be either partnerships between the GPS database firms and car makers - or drivers using the GPS capabilities of their mobile phones.
  • Most recently, new cars come with electronic systems that try to integrate, often inelegantly, the GPS and in-car sound and phone access capabilities.

None of these steps fundamentally changed the car. That's what, in principle, is possibly different about the nature and value of a driverless car. 

But it doesn't guarantee that the current automakers, with their roots in metal-shaping and in the design-and-manufacture of powertrains, will somehow get lost in the world of cars-as-computers. Audi, for example, showed off its driverless autos at CES 2013 (although they were without a doubt less advanced than Google's). And from another direction, we see Ford and GM collaborating on 9- and 10-speed transmissions: feats of engineering complexity that are considerable, even if less 'cool' than driverless cars.