The future is coming, but it’s not quite here yet.

I remember when I was a kid, I was perusing a big book about potential future technologies. There was all sorts of cool stuff, like synthetic skin, personal flying vehicles, and of course, self-driving cars. I showed I picture of the self-driving car to my dad and said “isn’t this cool?” He shrugged and said, “I guess, but personally, I prefer to drive my own car.” That was a little disappointing to hear as a kid, but as an adult, I kind of get where he’s coming from. While the idea of self-driving cars is definitely cool and futuristic, there are quite a few problems preventing their practical implementation.

For one thing, self-driving cars wouldn’t really work unless every car was self-driving. If every car used the same AI system to navigate, that’d be fine, because they move in predictable patterns. An AI, however, can’t predict spontaneous human decisions on the road; their logarithms would only work with the assumption that every human driver is following the rules of the road to the letter, and that sure ain’t the world we live in. With this flaw in mind, a mass rollout of autonomous cars would be a nightmare.

There’s also the matter of autonomous-compliant infrastructure to worry about. Roads aren’t perfect; they crack, chip, and wear away. Some roads are so old they don’t even appear on maps. A human driver can navigate a dilapidated road fairly easily because they can keep a good grip on the wheel and manage the gas. An AI would fly right into a pot hole because it’s not on the map or because the sensors don’t catch it in time.

I want to live in a world of self-driving cars. It would be awesome, of that there’d be no doubt. But unless the kinks can be worked out, and in an unbelievably fast and precise timeframe, we simply aren’t ready for them. Let’s try to get flying cars figured out first, then we can come back to that.