I'm not going to get excited until this can drive in Boston/Cambridge. If it can navigate that nightmare (and not kill any cyclists) I'll be impressed with what self-driving cars can do. Bonus points if they can work in snow when sightlines are obscured.
They have been navigating the SF Tenderloin for over a year. Entire streets are a crosswalk. Most biologic drivers would barely function there. But I do wonder how lidar/visual integration works in snow.
I don't know. Boston and New York may have a culture of taking whatever right of way you think you can get away with. The West Coast historically at least was more known for enforcing jaywalking though given the current state of SF I'm not sure that's still true. Boston/Cambridge also has a huge influx of students every year, many of which are pretty clueless about navigating an urban environment safely.
Depends on the location. I don't like driving in SF especially but there are some areas of Boston that are pretty crazy if you're not used to them and don't intuitively know when to get over into some lane and aren't used to driving pretty aggressively at times.
It's worth keeping in mind that autonomous vehicles find different things challenging than humans do. Weird road shapes and knowing how to route into some lane is pretty straightforward with map data. It's the "normal" parts of driving like predicting other road users that are more difficult.
Totally fair. Things I might find challenging as a human unfamiliar with a given location (though GPS tends to help a lot) are probably different from an autonomous vehicle dealing with a lot of borderline crazy behavior by other cars, cyclists, and pedestrians darting out into traffic. A person also can just process so many inputs at once so you get into situations that are challenging in part because so much is going on.
