Apple’s Perilous and Stumbling Project to Build an Autonomous Car

The Apple Car Crystal Ball

Once upon a time, Apple thought it could go from zero to 100 in the car business with no prior experience. We’re now seeing how hard it turned out to be.

I’ve just been looking at an interesting article at The New York Times by Daisuke Wakabayashi. “Apple Scales Back Its Ambitions for a Self-Driving Car.” I couldn’t help but pass along my own thoughts.

Apple concept car from earlier times.
One of the early, dreamy Apple car concepts.

Faith in Apple?

From an Apple observer’s viewpoint, I think I and many others had an inordinate faith in Apple when we first heard rumors about this project. (It came to be called project Titan.) The pervasive reasoning, as I read more and more, was that a company with so much wealth and techical expertise could easily equal the work of Tesla. And then add its own imaginative special sauce to dominate the market.

What we’ve experienced, according to the New York Times is that very large projects do not succeed simply by virtue of expertise in related areas combined with great wealth. (Aerospace managers already know that.)

First, I think we surmised that Apple would be able to learn from the mistakes of others. A project like this needs a seasoned automotive executive who has the full authority to execute a vision. Outside Apple’s executive team, there isn’t much authority. And there was little experience to steer by.

Next, I think we assumed that Apple would avoid the pitfalls of the waterfall systems approach and opt for an incremental, iterative spiral model. First, learn to build a car. Then build an electric car. Then build an autonomous electric car. If that had been the approach, I think we would have heard rumors about a giant car factory and the purchase of those giant manufacturing robots. But we never did, and I thought that was a key indicator.

Finally, we assumed that Apple would use best practices in software engineering. The NYT article reveals that there was a squabble about whether to use Swift or C++ for an an advanced, secure AI project. C++ is on the AI list, but it’s not #1 from what I’ve read. Java and Python are said to be better candidates, but they’re NIH. Not Invented Here.

Once again, this appears to be symptomatic of Apple declining to keep its toes fully in the waters of the science community and drift away into iPhone mania. But Apple can and should do both.

Telltale Signs

Sprinkled throughout the NYT article are the telltale signs. It was almost beyond imagining that Apple could avoid the mistakes that an inexperienced company would make in going from zero car expertise to going toe-to-toe with the major car manufacturers, many of which have a century of experience. And beyond imagining it was.

Now, what will be the result of the work done so far? Likely, a fallback to a few autonomous transport vehicles. From the NYT article:

Apple’s testing vehicles will carry employees between its various Silicon Valley offices. The new effort is called PAIL, short for Palo Alto to Infinite Loop, the address of the company’s main office in Cupertino, Calif., and a few miles down the road from Palo Alto, Calif.

This isn’t the first, large project to fail. Heaven knows, the history of American technology and the military-industrial complex is littered with monster projects whose dreams were dashed. Money squandered.

What is a little bit surprising is that Apple, a company steeped in modern success, could have made so many obvious management mistakes, enumerated in detail by the NYT sources. It seemed that we just assumed Apple wouldn’t make them.

Apple truly believed. We believed. But experienced Big Project managers might have foretold that belief isn’t enough.

3 thoughts on “Apple’s Perilous and Stumbling Project to Build an Autonomous Car

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.