Our Self-Driving Future Isn't Decades Away

 

“Andrew Miller is too pessimistic about our self-driving future ("Remember When the Information Superhighway Was a Metaphor?," op-ed, Dec. 3). Fully autonomous driving will worsen traffic conditions for a time, he argues, introducing cars that will be too deferential and unable to coordinate with other drivers. As a result, "no single behavioral rule book will govern the road," and everything will slow down.

 

The grandmotherly driving of today's Waymos conjures images of two, excessively polite Midwesterners locked at an intersection, each beckoning the other to go first, back and forth in perpetuity. Yet the technology is a moving target, getting better by the year. As the first-mover in full autonomy, Waymo took the cautious approach to prove that its self-driving cars aren't merely safe but dramatically safer than having a human behind the wheel.

 

Tesla's self-driving cars are also safe but evince greater assertiveness. Contrary to Mr. Miller's worries around human-AI coordination, Teslas even respond to common hand gestures, like being waved through by a construction worker. This is an emergent byproduct of being trained, end-to-end, on human driver camera data, leading the cars to drive more as humans do by default.

 

The personality differences between Teslas and Waymos may cause some frustrating anecdotes in the short-term, but they're unlikely to be a major drag on traffic, much less one that will require "decades" to work through. The brain of a self-driving car is constantly updated and refined; it will learn to coordinate seamlessly with other drivers, human and AI, in short order.

 

Mr. Miller's idea of creating segregated lanes or pursuing similar infrastructural innovation for self-driving cars could be worthwhile if we want, say, a high-speed autonomous autobahn. But for everyday driving, Mr. Miller underrates the rate of improvement in AI, and the extent to which the problems he identifies are amenable to machine learning.

 

The biggest inhibitors to self-driving are mostly political. Take Charles Allen, the District of Columbia councilman who this fall delayed Waymo's rollout in the nation's capital by demanding that the city government study it further. This sort of obstruction will be quantified in avoidable traffic accidents and fatalities.

 

Mr. Miller's creative yet data-free theory that autonomous vehicles will somehow make traffic less efficient does no one any favors.

 

Samuel Hammond

 

Foundation for American Innovation

 

Washington” [1]

 

1. Our Self-Driving Future Isn't Decades Away. Wall Street Journal, Eastern edition; New York, N.Y.. 08 Dec 2025: A16.