I wish the billions (trillions?) that were being spent on self driving cars was being spent on public transit.

ยท Web ยท 9 ยท 31 ยท 38

Here's a clever study:
It's hard to study how traffic will change with autonomous vehicles since we don't really have them yet.

So scientists gave people free chauffeurs insteadโ€”the original "self" driving car.

Result: People put 83% more miles on their car.

That thing where self-driving cars were supposed to reduce congestion? Doesn't look like its going to happen.

Link to the original paper (it's really easy to read):

Show thread

@readsteven But there is no way to make only selected individuals benefit from public transport, so it gives no advantage, as it improves the lives of everyone equally (even worse, arguably it improves the lives of the poor more than the rich).

@readsteven at least public transit will benefit from that tech eventually.


Unless your transit plans cover the spaces between cities (let's all have a good belly-laugh at the idea of spending money on services for rural populations!) you're still stuck with a lot of people on individual transport.

Seriously, one size doesn't fit all. Money on transit? Good stuff. Doesn't preclude money on self-driving tech, especially when the result is safer for pedestrians, many of whom are poor.

@jankoekepan @readsteven l agree with you, however so far self driving vehicles and collision avoidance tech don't have a great track record with pedestrians.

@teleclimber @readsteven Great, maybe not, but not bad compared to human drivers. Watch this space for substantial improvements.

Please understand: I think self-driving tech isn't a particularly good approach to our traffic, energy or climate situation, but given how crappy transit is in even the finest circumstances, I find it very hard to fit either one into all situations.

@jankoekepan @readsteven yeah I don't understand where the debate on whether self driving would help congestion came from it's obvious it won't.

To me the huge advantage in self driving is safety. Humans are bad at driving and many die because of it. Even if we don't achieve full self driving soon the technology can help human drivers and cover some of their weaknesses. So I'm all for it.

@teleclimber @readsteven

There is some reason to believe that self-driving may alleviate congestion in a number of ways, but it's heavily implementation-dependent. On highways, by removing human factors you can get smooth traffic movement with higher traffic density. In cities you can serve more people with fewer vehicles, and manage denser parking arrangements as well.

The problem is that many of these claims presuppose a majority self-driving, for significant effects.

@jankoekepan You're ascribing a lot of nuance in a 19 word toot. I never said one size fit all, only that I think the money spent on autonomous cars would be better spent on public transit.

I've lived everywhere from rural America to Tokyo and have first-hand experience with transit needs of those places.

@readsteven There wasn't nuance in the toot; that's kind of the point. It wasn't a plea for a rebalancing based on perceived needs, it wasn't an analysis of costs, needs and possible benefits as opposed to risks.

It was a straightforward call for money to be moved from one bucket to another, at the clear cost to one avenue of research and development, in favour of a solution which is unhelpfully irrelevant to many people who might benefit by the other idea.

That's why I said what I did.

@readsteven The only immediate positive about all the self-driving startups: Almost all of them test their cars in the northeast corner of San Francisco where my office is.

Me and my coworkers have joked that all the self-driving cars are being trained on the walking patterns of us going to lunch ever day - which will make us the safest pedestrians on Earth in the future. ๐Ÿ˜œ

*Starts cosplaying as Thomas for increased pedestrian safety*

Around here I see a few Aptive cars every time I go for a walk, but those are the only openly autonomous cars (licensed as such with a safety driver) I see any more.

The little fully* automated shuttle is off the streets now. Not sure what that means. (Probably the funding ran out.)

* There was no driver on board, but there was a remote safety driver, often in a car trailing behind.

@readsteven @tsturm Theres a test program here being done by ride sharing company Via. Kinda tempted to see how they react to me on my bike (they all have safety drivers)

@readsteven @tsturm hmm now im wondering if their LIDAR will interfere with the LIDAR module my senior design team is using for our bike approaching vehicle warning system

@cinebox @readsteven Fighting LIDARs - we have truly arrived in the future. :)

@tsturm @readsteven There was an article a few years ago on Ars or Wired or something about how thats already becoming a problem for self driving test programs. I wonder how it can be avoided

@tsturm @readsteven maybe self driving vehicles can turn off the laser when pointed at each other and use an active cooperative sensing method instead?

@cinebox @readsteven ...or if it is a car from a competitor, they turn up the laser to full blast, aiming at the sensor packages.

@readsteven @cinebox Conundrum: Anti-surveillance gear will make you invisible to self-driving cars.

Do you want to be tracked by the surveillance state or safe from autonomous vehicles?

@cinebox @tsturm
I thought LIDAR did some clever things with polarization (?) so it could listen only to light it sent? (I think it's trickier for ultrasonic sensors, though my current car has those and haven't had visible interference. Yet. That I've noticed.)

@readsteven @cinebox @tsturm one way is to send out a specific coded message and listen for it to come back. That's at least one way I know of that we do radio astronomy.

@cinebox your project sounds awesome. I'm looking forward to having one on my bike to keep me safe from all the crazy people taking naps behind the wheel!

@cinebox @readsteven In SF I see many of the Cruise and Zoox cars with drivers. The Cruise cars are so dense around our office, I can often (nearly daily) spot two or three on the same block.

There's also at least two other companies that don't mark their cars, but have lots of sensors bolted on. I see those also pretty often, but clearly not funded to the level of Zoox and Cruise.

@tsturm @cinebox What's interesting about the Aptiv cars is that the sensors are really pretty subtle. If the cars weren't wrapped in a company logo and have the red license place, I'd never notice.

I've told the Lyft app I'd be willing to ride in one, but haven't gotten one yet. (No surprise given how rarely I ride share.)

@readsteven @cinebox The Cruise cars have a slick blob of sensors on the roof, but without the branding, most people wouldn't take a second look.

@readsteven if we are on wishing, I'd like to see roads and parking lots begging repurposed too. I've been fascinated with the concept of superblocks (I think in Barcelona) and the Columbian mayor that spent tax payer money on buses and pedestrians instead of road ways for car owners.

The only thing that would make self driving cars work is city owned and maintained. Individuals wouldn't get personal cars. But you could still request one to go to the market or whatever.

@readsteven @megfault I would argue that autonomous cars will behave more like public transport in some ways. You might not own one, you most definitely will not park it in front of a building, it will probably be used more efficiently (per vehicle). Not exactly the same, but already a lot better than classical "my car, my parking spot, 1.2 passengers on average, 23hrs parking per day"

In a sense, it is. The future of public transport is unlikely to be buses and trains, it's much more likely to be automated ride services. This is already happening in Japan with there self-driving minibuses from rural towns to cities and should make its way to the west.

Sign in to participate in the conversation

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!