I don’t think fully capable self-driving cars will exist in 3 to 5 years. It will be more like 25 years or more. The engineers are going to hit a wall in the next few years when they try to solve tough technical challenges like passing slower vehicles on 2 lanes roads in the country. That’s a very tough software development challenge and there’s zero room for error because any mistake can result in a fatal head-on collision. We might see fully autonomous vehicles in controlled environments in cities, where wifi beacons can tell the vehicles where they are at all times. Something like that might work, but there’s an astonishing number of situations that can occur out in the suburbs and in the country—things like frontage roads running right next to 2-lane highways with vehicles driving in the opposite direction on those frontage roads. That’s just one example; there’s so many situations and different kinds of road markings and weather conditions.
The litigation issue is complex. Autonomous vehicles could end up being safer drivers than humans overall, but when a human causes an accident, the victims can only sue that person and then his insurance company pays for damages. Those individual drivers generally don’t have very deep pockets so the damage payouts are generally quite limited. But auto manufacturers have deep pockets and the trial lawyers will go after them for billions of dollars if those autonomous vehicles cause numerous accidents, even if the accident rate per 1000 miles is much lower than the rate for human drivers. It’s a complex subject, but I think VC people are getting way ahead of themselves about fully autonomous fleets of vehicles.