Perfection is the Enemy of Good Where AVs Succeed

News articles about automated transportation technologies consistently imply that the challenge of scaling and expanding destines the technology to be if not useless then at least a bit silly. This attitude is both wrong and unproductive in solving major transportation and climate related problems.

For example, a Scientific American article about delivery robots has limited discussion of the benefits of delivery bots but a lot about problems. It notes that though delivery robots are useful on college campuses, they may not be useful everywhere because of space, other sidewalk users, paving conditions, etc. Ultimately, it concludes that delivery robots are not the future of delivery; generally, in my opinion, the tone is negative on bots’ practicality despite their success in college towns.

A more obvious example is those who refer to the “Holborn problem:” because self-driving cars will not know how to handle a London business district during rush hour, self-driving is “a ludicrous tech nerd fantasy… indulged by clickbait-chasing journalists.”

I disagree with the idea that because an automated vehicle is useful in one context but difficult to apply in another, it is an impractical absurdity, much ado about nothing. Perfection is the enemy of good.

Delivery bots are good. Transporting takeout via car means moving 2000 pounds of vehicle along with a small bag of food; talk about energy inefficient! A little delivery robot requires less power to accomplish the same goal, not to mention effect on air quality, safety, road construction and maintenance etc. Replacing one car doing delivery with a robot doing delivery is an environmental net positive.

In other words, even if delivery bots that deliver on the UC Irvine campus will never do the same in New York City, they’re still great! These successes should be celebrated and taken advantage of, not responded to with a litany of challenges in expanding to new contexts.

This negativity discourages further development and expansion. Suppose a city counselor is approached by a delivery bot company for a pilot. They will web search autonomous delivery robots to learn more, encounter articles listing all the potential problems, and decide it’s too risky, impractical, etc. for their town. There goes an opportunity to get some cars off the road.

Pundits and AV-thinkers should approaching these autonomous technologies with so much fatalism and pessimism. Yes, they’re not perfect, and they can’t go everywhere. But if they’re useful at all, they’re useful even if they’re only in limited areas. Stop letting perfection interfere with good.