Uber vs. California - Robot cars not ready for roads?


Just give Amazon an address, maybe they’ll deliver him to us.


I survived! It was very easy to use and pretty disappointing at the same time. It’s essentially a convenience store with less stock. If you want a quick drink or a snack, it’s fine I suppose. I don’t think I’ve ever been too frustrated with long waits to check out at 7-11 though.


Just wait until someone gets in line ahead of you with a bag full of pennies…


I find this prediction incredibly likely.


Not to mention awesome.


Naked bodies all over the street, can’t wait for those kind of wrecks to happen. Oh yikes, they might replace the tour buses so busy cities can be jammed full of autonomous cars.


Wait. I don’t understand why we don’t have cars and vans functioning as roving brothels now. Why couldn’t a pimp/accomplice drive while you “party” in the back of the vehicle now?


To take this conjecture more seriously than it deserves, I think the motivators are supposed to be cost and perception of privacy.


There is an entire series of “documentary” videos that are available exploring this theme. Or so I’ve heard.


How do you know I’m we’re they’re not doing it already?


Well I guess I can go ahead and toss out that small business loan application I was working on.


I heard on the local news a few weeks back that we’re getting some variety of roboshuttle here in Grand Rapids. Be interesting to see how they handle snow.


Why do I get the feeling they’re just replacing one boondoggle with another.



Kind of a slideshow / video presentation, not a straight up written article.


Whoever convinced whoever at WaPo that this would be a good way to present information, needs to be beaten. Okay I’m watching a video, wait now I’m reading a paragraph of text, wait now I’m clicking through a slideshow… cripes.


Yeah it’t not my favorite either. It also doesn’t ever give you an indication of how long it is. I started thinking it would be an article I can read in a few minutes, and then it was this weird combo thing… but it does give some interesting information and video despite the delivery method.


Multimedia! Bring back Encarta!


Current state-of-the-art self driving cars-- still inconveniently slower than a human driver, unless you turn off some safety checks, in which case they’ll kill people if there isn’t an on-the-ball safety driver.

My impression is that they won’t be economically viable (except for fixed-route shuttles) until there is the sort of AI breakthrough that none of the companies is even working on yet, where you create a model of other drivers expected behavior (and pedestrian behavior) based on matching observed behavior and observed properties of the people and their cars with an array of learned driver types. This is the sort of thing people do quickly and unconsciously that self-driving car companies haven’t begun to tackle at all.

Teska eems to think the problem is just marrying lane assist with GPS. Waymo seem to be more realistic about how hard the problem is, but I think still underestimate the level of abstraction that will finally be required for “real” autonomous Level 4.


There are VERY good people working on this, world wide. Its is, dare I say, one of the most important uses of AI and research for transportation, currently. They WILL cross that barrier. Look not at Tesla, look at Google.

Look, there is no future that wont include self driving cars. Human drivers are too expensive, too error prone, and too much of a liability.

You realize there are autopilot technologies in air and sea traffic getting a ton of development too, right?

I should mention, there is an article I saw recently about Tesla’s accident rate, and despite their, “lane assist and GPS,” they are at a quarter of the accident rate per million miles of non-assisted vehicles.