Sorry. Your car will never drive you around [video] | Hacker News Reader

Sorry. Your car will never drive you around [video]

youtube.com

31 pointshnuser000029 comments

Despite the video saying that we'll never have full self driving cars (I pretty much agree,) let's assume that it is possible. I think we'd end up with such stringent controls on the technologies, way more than industries like the Boeing and other aircraft manufacturers have to deal with and that FSD would never scale up to present day automobile use.

The whole conclusion—summarized in the A Better Idea? chapter—is him pretty much just slowly rediscovering a metro train system with modern signaling.

Of course. He has figured out that separate lanes, or even complete separate infrastructure (subway, tram) is easier for self driving implementation, however it did not occurred to him, that if he wants door to door self driving, then developers needs to implement fully autonomous vehicle anyway thus separate lanes makes no sense. If he is ok with walking into a point of collection, then it is just public transportation.

Maybe Tesla cars will never drive you around, but Waymo is already doing it. That basically sets an end date for when we will expect to have consumer grade self-driving cars to “wherever Waymo’s patents expire.”

waymo is not your car tho is it

Waymo vehicles have a significant amount of extra equipment. They are commercial, justifying the added expense. After all, we can all get a helicopter if we want to, but for some reason, they aren't common.

I would absolutely pay the extra cost to not have to drive my car.

Yes, but if I posted a whole treatise on why helicopters can't work to YouTube, I'd be laughed off the Internet; and rightly so.

Waymo'e existence as a working taxi service proves self driving cars are very much possible, even if the market is limited to sections of Phoenix, San Francisco, and soon parts of LA and San Antonio, so a story claiming never, for something that already exists today, seems shortsighted.

Another video that could exist is titled "why you'll never own a Ferrari". We then get into semantics as why you'll never own one is different than they don't exist, but I've seen Ferraris so I know they exist, and a video saying I'll never own one is a depressing look at why I'm not offensively rich, rather Ferraris not actually existing.

Point is, self driving car technology is here and providing value to a tiny group of people. Google is scaling up as fast as they can. Whether you'll get one is an implementation detail.

We don't all have helicopters because they are extremely dangerous and loud and require a helipad and trained pilot. Small helicopters can be purchased as cheaply as a used car and I suspect even cheaper if they were being mass produced due to much higher demand.

"Never" is a very long time.

Maybe it won't happen as soon as many predicted but I'd say it is likely to happen at some point.

Maybe the word doing the heavy lifting is more the "you" than the "never", in the same way we will never see the Sun engulf the Earth because you and I are going to have snuffed it in a few decades.

Either way, a self-driving car is only worth having if it lives up to the premise/promise/hype, is reliably safe et cetera, otherwise it's the enshittification of personal transportation.

The many HN readers in their 30s have roughly 50 years ahead of them.

50 years ago was 1974, the year the Altair 8800 was launched [1]. Let's assume that the best self-driving car of today is an Altair 8800 and try to imagine 2074. So who knows.

[1] https://www.computerhope.com/jargon/a/altair.htm

The takeaway isn't the demographics of the audience, it's that we should probably focus less on personal transportation in the form of cars, focus more on mass public transport infrastructure, and encouraging people to work remotely whenever possible. There's precious little point in RTO to sit at a computer in a different room and still take Microsoft Teams meetings instead of going to the boardroom. It just represents wasted energy.

We need to use less energy, not more. Perhaps self-driving cars are a waste of brainpower in the long-term? If people don't actually want to drive in 2024, perhaps don't get a car, take public transport, hire a chauffeur, and/or work remotely. Self-driving largely seems like a half-baked solution in search of a problem. Self-driving should really be a lot like self-hosting—do it yourself.

That's just my two penneth though, and I can come up with exceptions, edge cases et cetera all day long too. There's no binary right or wrong with it, and there's trade-offs to everything.

Was only thinking earlier that if we should be using less juice, why are we adding AI gimmicks to consumer gadgets? I'd like to think the new hotness with the average Joe will end up being saving energy to spend less money, and having products that last longer. Sure, that's not what the HN demographics want to be reading, our current economic model doesn't exactly go hand-in-hand with this idea. There's a lot of very big questions ahead of us in the next fifty years. Perhaps UBI makes more sense instead of bad technology, forced obsolescence, gimmicks, wasted energy et al if it's helping the race and planet.

This is all mostly orthogonal with the feasibility or even desirability of fully self-driving vehicles.

For commercial vehicles (taxis, buses, delivery, trucking, etc) they are a real solution to cost and productivity, which could potentially have ripple effects on private car ownership, parking space, etc. For private vehicles they can free the driver who can then do something else like you would in a taxi or public transport.

For instance, if cars drive themselves then the operating costs of a private car and a taxi are the same so we could imagine that taxi fares would drop. Now if indeed people work from home more and don't need to drive everyday then perhaps not owning a car but insyead to summon a self-driving taxi when needed would make sense.

There should be a "robot road" which is autonomous driven cars only. There are so many reasons to support this. Robot cars and humans do not belong on the same roads yet.

Then improve efficiency by adding rails to reduce friction, and then connect multiple cars together. Finally, remove the battery and use overhead wires or a third rail to power this “robot car.”

Most autonomous cars can also drive non-autonomously with human control. In your example, those cars could then disconnect and go on their merry way, like cars on a ferry. The last mile (or several miles) has always been the shortcoming of mass transit.

More like rental cars at an airport. There are Park n' rides, taxis, busses, bikes, scooters, friends with a car, etc. fully capable of delivering the last mile.

The last mile is only a shortcoming in underfunded transit systems of sprawling suburbs.

Yes and when those shortcomings exist, a robot lane added to existing route sounds like a fine idea.

See: HSR in the California Central Valley.

The I5 is too narrow already, it needs widening and the HSR won't fix that. Adding car rental and expanding the local transit solutions at each station along the route will balloon its cost to extremes. It's already unacceptable.

Ah yes, highway projects are cheap, transit projects are expensive. Classic.

If trains don’t take cars off the road, we have to explain why Chicago doesn’t have a 26 lane highway but Houston does despite being a similarly sized metro area.

Where do all the cars go when over 20 million passenger trips take place on Chicago public transit every MONTH? Houston’s numbers are at least 4x less from what I can tell.

California needs to withdraw from the cat drug, not up the dosage.

Even in California. If you compare the Transbay tube to the Bay Bridge, you really see how the Bay Area simply could not function as currently without BART. The tube services around 27,000 people at peak hour per direction, while the bridge only services just over 14,000 people (in 9,000 cars).

Without the tube you’d need two additional 10 lane bridges to service peak demand. And that just wouldn’t be possible.

https://youtu.be/HbZXtSnabWc?si=-mWIkQg3gTV-ZSMe&t=227

The last mile was only a shortcoming when we started building for cars instead of people.

The transition to full AV-only roads will be slower and quicker than most people think. It will take a long time still for AV technology to be perfected, but once it is proven we will see various bans on human drivers coming in rather quickly. Texas has already been talking about dedicated corridors for AV trucks.

Agreed. You could also maybe have special tracks in these roads that the cars could attach to so they would not need tires. Additionally I bet you could link them together so the robots could carry hundreds of people at once. I bet you could even run them on a regular schedule so people can use them to commute on the regular.

That'll never work. People would have to move, using their feet, from where they are, to the common stationary meeting points that this special vehicle met at, which makes them far less convenient than taking a car which picks you up and drops you off right where you need to be. In places where things aren't very dense, this "stationary common meeting point" is liable to be far away from where the person needs to be. Furthermore, this "schedule" you talk about. In order for this proposed system to work, the "schedule" would really have to be more of a frequency than a set schedule. This means that there will need to be a large number of these special cars in order for them to run frequently. This will cost lots of money, far in excess of what any one rider could pay. So then it needs to be subsided by the government, who are reluctant to spend money on things, so it needs to be well supported as an important expense by the whole population. A population that already owns regular cars. So it might work in dense locales with a consencious population, aka Europe, but I don't forsee such a system gaining popularity in the US outside of dense urban environments, and even then.

Besides, self driving cars, which are already here and work thanks to Waymo, don't need special roads or rails to run on.