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

Oh, PLEASE let this happen. It’s one of the few times I really, really hope a company gets held to the stockholder irons regarding stupid shit that management does. Full disclosure on costs for legal representation of management? Yes, please. Uber has repeatedly shown itself to be a cesspool of pretty high magnitude. (This is me assuming that any settlements would be significant enough to Uber’s bottom line that they would get disclosed, and I think they aren’t big enough to hide them from a 10-K.)

The truth is that it’s a grift. Everyone promises what they know they can’t deliver because they want the moneys now and don’t care about what happens later. Musk keeps telling everyone that any day now their Teslas will autonomically drive to Mars and bring them back Frosties while they sleep, and people go oh, cool and buy that shit. It’s fraud, but no one will say it’s fraud.

I almost got into a wreck today making an unprotected left hand turn (my fault). I can kinda see why Waymo would take the long way around a block in order to avoid them.

Here’s a pretty great article about the status of self driving vehicles:

There was a lot of hype about lots of “futuristic” stuff for the last century at least.

Watch this video.

So what you’re saying is that in our driverless future I’m going to have to walk outside to get my food?

Never!

SMMT (UK motor trade group) roadmap for autonomous vehicles:

Bear in mind L1 basically means cruise control.

For a simple levels breakdown:

Bear in mind that there isn’t a single L5 prototype / testbed ‘car’ in the world. Not in a lab, not on a test track, not anywhere. Not for $100m or $10b. It doesn’t exist, which means there isn’t any production version coming anytime soon. The idea that it’s a year or two away is all hype and stock price boosting.

Yeah, I work for a company that does stuff in that space (I do not work on those project should personally) and L3 is the aspirational.

Like how to navigate a parking garage and park properly is a major recent breakthrough. And that ain’t using stock solutions.

Navigating the non-Euclidean geometry of parking garages is quite an accomplishment, though!

The L in L3 stands for Lovecraftian.

That is glorious.

It bugs me that people just accept this ‘fact’ that waymo are the ‘market leader’ in self driving because they can shuffle people at 5 miles an hour around some geofenced community about a mile across. Its laughable.
geofencing is not a general solution for self driving. Tesla are holding an ‘autonomous driving investor day’ on the 22nd of this month, and I suspect that will change peoples minds about the state of the art for this tech and where they stand regarding everyone else.

I suspect it’s a very practical one, though!

I’m looking forward to what Tesla has to show.

I look at the way Waymo has been faking it in Arizona and I think that the state of the art hasn’t really advanced in the 30 months since I made this post.

We may have reached a limit on what can be accomplished with detailed maps and deep neural networks. Economically-useful autonomy may require AI breakthroughs that aren’t even in the labs yet; something that understands the outputs of its sensors as representing more than just obstacles, but as potential agents whose behavior is best predicted with an understanding of social cues and psychology.

I wonder if people would be more open to an increase in the Federal gas tax if it was targeted specifically at improving roadways for autonomous cars.

Probably not yet, but in the future?