Thursday, September 20, 2018

Forward And Back

On one hand it's good that people are starting to realize that I am gonna be proved fucking right, on the other hand I worry that upon finally accepting that they aren't going to work 100%, they're going to forget that if they don't work 100%...they don't really work.

Toyota doesn’t necessarily buy the hype about self-driving vehicles quickly taking control of roads in the U.S. and beyond. Leonard himself isn’t sold. “Taking me from Cambridge to Logan Airport with no driver in any Boston weather or traffic condition—that might not be in my lifetime,” he says. On its website, the research institute describes its goal as to “someday develop a vehicle that is incapable of causing a crash.” It doesn’t specify whether this uncrashable car would be driverless.

Of course enhanced safety features, if they *enhance safety* in practice, not just in concept, are good, but there is the basic problem that if any of these features encourage people to not quite pay as much attention as they should, then they're probably self-defeating.

Toyota is seeking a middle ground with a system it calls Guardian, which would harness the machine-intelligence and sensor capabilities that make full self-driving theoretically possible and bundle them in vehicles designed for human drivers. These cars and trucks would be able to see much farther ahead and behind, across multiple lanes of traffic, than any human would, and would be more adept at anticipating the behavior of other cars and pedestrians. Actual people would continue to steer and brake, but when Guardian detected potential danger, it would assume control and swerve, slow, stop, or otherwise act to avoid the problem.