Autonomous Stupidity

:date: 2024-07-02 13:20 :tags:

It's been a long time since I've covered the topic of autonomous vehicles and it's questionable whether I should now. But there have been several "news" items that came across my screen that I thought I should comment on.

First up, the world's video website recommended I watch the inaugural "race" of the Abu Dhabi Autonomous Racing League. Ug. Do I have to? Yes. I had to.

And just like I could have told you, the driving oscillated between the literally superhuman and the literally brainless with an average somewhere around "janky". Watch a quick minute of how it went. Time marked link to the actual driving bits of the video.

According to their web site (a2rl.io) there was $2.2e6 in prize money at stake. Wow. When I did autonomous car racing, I was pretty happy to win 150 dollars - flimsy Australian ones!

That's right - recall that I did autonomous car racing way back in 2012. But we didn't do it so stupidly. First, I got into it 12 years ago but it had been going on for a few years before that. So, not exactly cutting edge. But the critical thing was we did not use real cars. Looking at the web site a bit, I can see they must have some event with model cars and that's probably the right place to start. (Paging Duckietown!)

To jump right into F1 style cars? What's the point? You'd need to have money to burn... ohhh. Riiiight. Well Emiratis, you have fun with your petrodollars and your stacked Tour De France Team and your ... everything else.

It is a shame though. A lot of possible potential and progress wasted  —  kind of the story of autonomous vehicles in general. In the video here the guy makes the correct point, "This is not a good advertisement for the technology. ... It just makes it all look like a joke. And there are people who worked really hard on this and they deserve better..."

Yup. If you want to watch slightly less janky autonomous car racing where no southeast Asians are exploited as slaves, you can watch the videos of my old project.

Next in stupid autonomous car news, Waymo recently opened up its San Francisco autonomous car taxis to anyone with a phone. I actually don't know if this is progress or not. What is stupid about this is that San Francisco  —  a city I have traversed by bicycle dozens of times  —  is hard for autonomous cars.

One theory could be that they started with "easy" level of difficulty over in Chandler, AZ and now have graduated to "expert" and are therefore ready to take on all road systems. Sounds good, right? Only the reality is probably not like that. I suspect they've just pissed away tons of development resources to make their system tractable in SF.

When you consider that it was four years ago that Waymo announced they were opening their service to the public in Chandler, AZ, you have to wonder about the progression. Why did they push so hard to get this running in SF? As far as I can tell it's the same reason as in the UAE: rich idiots doing vanity projects on home turf. Just remember, that at a million people a year dead from idiot human drivers, this kind of vanity probably came with costs. Or, if not, then the whole thing's a scam and the tech is nowhere near ready. Take your pick.

And finally let's go to Japan where The Japan News notes, "An expert panel of the Land, Infrastructure, Transport and Tourism Ministry has proposed the development of automated logistics links that transport goods on median strips or through underground tunnels along expressways.

Wow, that's kind of funny. First I wrote about exactly this kind of thing way back in 2020, calling it Real Progress In Autonomous Vehicles. And honestly I have mixed feelings about a proposal to develop a special lane to move cargo. That special lane  —  that's really the key. As I've said for 15 years, without idiot drivers in the mix, we've been technologically ready to replace driving since the 1990s.

But this Japanese proposal is for cargo. They show a special lane with (what I call) "idiot driver insulation". And that's all good. Very sensible. But with jammed passenger cars to either side, could they not envision another type of "cargo" that could really use a system like this? Is there any reason that people couldn't also be moved like this? Whatever. The article (and this one) also fret about the scary cost of building tunnels for this kind of thing. Ug. This is totally triggering my base rate fallacy allergy. (Ahem, how much does it cost to build a tunnel for a subway? How much does it cost to just build a surface road?)

At the end of the day, as I said in 2020, the good thing about this is giving some thought to keeping idiot drivers and sensible transport separate. Autonomous vehicle development is not a technology issue, it is a policy issue. It does show us that Japan or UAE or some random place is just as likely to stumble upon this fact as the USA. I'm still not holding my breath that there will be any really useful developments before 2030.