Read Brad Templeton’s latest post on self-driving cars, which has a number of updates. They’re coming fast, but how fast we drastically reduce transportation deaths, give people back a huge amount of time, reduce stress, and greatly reduce space and other resources dedicated to transportation, and how secure new systems are, is undetermined. Of course there are many reasons to be skeptical — the transition will probably be much slower and more problematic than needed, but in a few decades will still seem a major triumph. But I don’t want the hidden trillions of dollars, hours, lives, carbon emissions, malfunctions, etc. that could be saved sooner to be wasted.

Regarding security, malfunctions, etc., we need to demand use of proven secure protocols and source open to inspection, i.e, not play security through obscurity. Regarding space, planning for urbanity remade (largely, recovered) through autonomous vehicles needs to be the top urban planning priority.

The benefits will be so great that we should also think about how to speed adoption — the only disheartening news in Templeton’s post concerns a survey in which only 20% of car buyers would pay an additional $3,000 for a fully (if I understand correctly) self-driving car. How little respondents value their own time and lives, let alone others’! It’s time to start agitating for road owners to ban human drivers. Most road owners are governments, but not all — consider as an issue of public policy or consumer demand as you wish.

Won’t banning human drivers disadvantage poor people who can’t afford a self-driving car? Possibly very briefly, but on net I expect self-driving cars to have an egalitarian effect — they’ll make owning a vehicle at all unnecessary (a rental can be summoned on demand), reduce housing costs (of which parking is a big part), and allow the recovery of areas walled off and drowned out by highways.

Let’s ban human drivers from at least some roads by 2020. I suggest starting with San Pablo Avenue in Oakland, Emeryville, and Berkeley — because I live close to it! Admittedly a downtown area or certain lanes of a highway might be an easier start.

^*In theory it is usually preferable to increase prices rather than ban altogether. In this case, obvious mechanisms would include drastically increasing driver license fees and tolls for vehicles with human drivers. In practice, a ban may be more feasible.