Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Even after $100B, self-driving cars are going nowhere

bloomberg.com

Post reply on HN
Print view View original

61–70 of 501 posts

Re: Even after $100B, self-driving cars are going nowhere

#61

Post by awillen »

batmanturkey wrote: post #32

WHY is a fully autonomous vehicle even interesting to pursue, given that wheeled vehicles and their infrastructural requirements are such a large expense and social and environmental problem relative to the benefits provided? I mean, out of all the moonshots, why die on this particular hill? I consider an actual hyperloop to be a much more compelling vision for the future: those pneumatic tubes the banks use in subur…

A hyperloop is a much, much harder problem. Not technically, but practically. Just look at California's high speed rail and how that's gone - with that history, the idea that we could introduce hyperloops in US cities is, to put it only mildly punnily, a pipe dream.

As others have pointed out, millions are dying in car accidents now, so a "more compelling vision for the future" is much less important than something that can get done. While we certainly haven't seen the progress many have hoped for/promised on self-driving cars, I'd wager everything I own that we'll see a significant percent of cars on the road in CA operating autonomously before we see that high speed rail finished.

Re: Even after $100B, self-driving cars are going nowhere

#62

Post by sgtfrankieboy »

batmanturkey wrote: post #32

WHY is a fully autonomous vehicle even interesting to pursue, given that wheeled vehicles and their infrastructural requirements are such a large expense and social and environmental problem relative to the benefits provided? I mean, out of all the moonshots, why die on this particular hill? I consider an actual hyperloop to be a much more compelling vision for the future: those pneumatic tubes the banks use in subur…

Hyperloop is crap, just go with highspeed trains. They get the job done in city corridors.

But you can't have a train stop in front of your house... Nor can you build massive tubes everywhere in place of cheap roads.

Walking, Bikes, Busses, Metros, Trains, Cars, Planes. They all have a place in how we transport, right now the distribution between them is just out of whack.

Re: Even after $100B, self-driving cars are going nowhere

#64

Post by jimmytidey »

batmanturkey wrote: post #32

WHY is a fully autonomous vehicle even interesting to pursue, given that wheeled vehicles and their infrastructural requirements are such a large expense and social and environmental problem relative to the benefits provided? I mean, out of all the moonshots, why die on this particular hill? I consider an actual hyperloop to be a much more compelling vision for the future: those pneumatic tubes the banks use in subur…

I think non-human payloads make more sense for virtually every kind of innovation:

-- Radically reduced safety margins, if the vehicle itself contains no humans.

-- You can often move cargo at any time of day or night, and therefore do a much more sophisticated job of collaboratively load-levelling traffic.

-- Doesn't matter how uncomfortable a non-human payload is with in reason. Acceleration, cornering, waiting, daylight, motion sickness are all constraints that are eased for cargo.

-- It would be nice if you can move freight anywhere, but even if you can only serve a very limited network its still useful - say between fixed points in major cities. By contrast, a car that can't go almost everywhere is very severely limited.

Example: in major cities, all the freight moves in lorries, which are extremely dangerous and polluting. Meanwhile, the humans are packed into trains in underground tunnels. Surely freight should be moving in autonomous underground trains and the humans enjoying lorry free surface travel?

Re: Even after $100B, self-driving cars are going nowhere

#65

Post by spaceman_2020 »

case0x00 wrote: post #4

This article is kind of a wrapper for the linked bloomberg article, which is more interesting IMO. Driving of cars is a massive cooperative game with high stakes, and autonomous cars essentially need AGI in order to play to a degree that is safer than a human with other humans. Fully autonomous cars would be sick, but IMO you'd need massive infrastructure changes (realistically restricted to cities/urbanized areas) i…

Practically all the AI-assisted automation systems I'm familiar with, whether its identifying objects in an image or parsing documents, have the same problem. They work for 90-95% of problems but get trumped up by edge cases and a human has to intervene.

That's not a problem when you're transcribing a video, but becomes a matter of life and death when you're driving a car.

Re: Even after $100B, self-driving cars are going nowhere

#66

Post by seshagiric »

Is it a data problem? If you have data modeling every nook and corner of the world and every possible combination of human, objects, roads and traffic/ weather conditions, then the AI can learn. If not, it keeps getting confused, will enter private driveways or get bamboozled by traffic cones. Till then it will be an iterative process where the AV companies will keep learning from incidents like those described in the article.

On the other hand, I do think the biggest impact of AV will be in controlled environments like the mine dump trucks in the article or the relatively easier long-distance highway only trucks (AV trucks haul till city outskirts, human drivers supply into cities).

Re: Even after $100B, self-driving cars are going nowhere

#67

Post by zinekeller »

Keyframe wrote: post #48
batmanturkey wrote: post #32

WHY is a fully autonomous vehicle even interesting to pursue, given that wheeled vehicles and their infrastructural requirements are such a large expense and social and environmental problem relative to the benefits provided? I mean, out of all the moonshots, why die on this particular hill? I consider an actual hyperloop to be a much more compelling vision for the future: those pneumatic tubes the banks use in subur…

- 1.3 million deaths worldwide, annually (94% cause by human error) - existing infrastructure (roads) isn't going anywhere (no need to build anything) - reduced traffic congestion (autonomous vehicles can better react/drive) - improved supply chains (autonomous trucks) (can work 0-24) just to name a few off top of my head

> existing infrastructure (roads) isn't going anywhere (no need to build anything)

I think that this is a sunk-cost fallacy. Sure, you would need to build infrastructure for trains, but after that trains are now reasonably automated, can carry cargo, and do other things with the exception of stopping at the exact destination you wanted to (in other words, it requires predetermined stops which people outside the US shrugs and just walk or bike the last feet). Also, they're proven to work: even China (which previously didn't have trains to its far-flung places) and they've done what you've expect. I think that the sole reason that anyone wants to invest in automated (non-train) driving is because trains are boring while AI is oh-so-shiny.

Re: Even after $100B, self-driving cars are going nowhere

#68

Post by micwag »

hasmanean wrote: post #60

Self driving trains would be far more achievable and practical…more tiny engines and smaller cars means latency and switching delays would go down. I don’t think it even needs much technology.

There are already many self driving trains operating around the world:

https://en.wikipedia.org/wiki/List_of_automated_train_system...

Re: Even after $100B, self-driving cars are going nowhere

#69

Post by MrDresden »

hasmanean wrote: post #60

Self driving trains would be far more achievable and practical…more tiny engines and smaller cars means latency and switching delays would go down. I don’t think it even needs much technology.

> "more tiny engines and smaller cars means latency and switching delays would go down."

Thorougly disagree.

The whole reason why trains are so efficient is due to their ability to haul so much weight in one go without practically any other traffic being in the way. Train track infrastructure could not handle multiple mini trains without more track which, as we know from the excessive road construction for car traffice, does not solve the congestion issue.

Re: Even after $100B, self-driving cars are going nowhere

#70

Post by saimiam »

feketegy wrote: post #28

I've said it before, if you truly want to experience what a real life raod looks like in most of the World, just go to Eastern Europe and create the test models there.

And then there is India. Please come here if you want zero infra AND zero adherence to basic rules of driving AND the language and script changes every few hundred kilometers AND all roads have human and non-human users at all hours of the day.
Post reply on HN