Autonomy is real now

Amara’s law frames the piece: a technology is overestimated in the short run and underestimated in the long. Autonomous vehicles are his best example, and he says the 2017 enthusiasm, his own industry’s included, stayed premature for years. What changed is evidence rather than promises. Waymo rolled out city by city because it needed to train models on real streets and earn public acceptance, and the safety data now exists: tens of millions of rides show roughly a tenfold advantage over human drivers, including those using the latest assistance features. Expansion to three more cities and a factory lot holding nearly three times the deployed fleet suggest the company thinks it has crossed over. His bet follows: because training appears to transfer between cities, first-mover advantage may compound into a position he calls possibly unassailable.

The bet depends on economics as much as on driving, and here the argument turns. Removing the driver removes the largest cost in a taxi business, but the equipment that replaces him runs to six figures a vehicle. His estimate is that the cost per mile already beats taxis and rideshare anyway, and he expects parts to cheapen with volume. The energy conclusion falls out of that capital cost rather than any environmental preference: a vehicle this expensive has to run almost constantly to pay for itself, and high utilization is where an electric drivetrain’s lower maintenance and fuel costs dominate. The sensors and processors also need powering, compounding the advantage. That, he suspects, and not corporate virtue alone, is why the fleet is already electric.