Can we stop talking about autonomous cars?

[sdifox]

In terms of shutting down metro areas to nothing but autonomous cars...

Ah yes, alienating people, the surefire way to success in attracting a vibrant business district in a city.

First city that does this goes bankrupt 30 years later. I'd move out if where I lived did this. Legitimately.

You might think, wow less people and less competition downtown, this is great, cheaper parking, cheaper land, all the benefits of the city.

Well so goes the land so goes the taxes so goes the city. That would be the beginning of a death spiral for a large city my dear.

What about all the people that commute in to work right now?

 

[K1052]

Bear in mind that at Detroit auto show last week, Ford and, I believe GM, unveiled their strategy for some 200k automated vehicles within 5? years. These aren't consumer vehicles, but fleets that will be owned by each company, as a service to designated metro areas. It's possible that the non-owner world will come about even sooner than an optimist like myself predicts.

Not hedging against the possibility of an autonomous future is not a risk that those companies can afford to take. The day could conceivably come where both are simply transportation companies that happen to make their own vehicles.

 

[K1052]

In terms of shutting down metro areas to nothing but autonomous cars...

Ah yes, alienating people, the surefire way to success in attracting a vibrant business district in a city.

First city that does this goes bankrupt 30 years later. I'd move out if where I lived did this. Legitimately.

You might think, wow less people and less competition downtown, this is great, cheaper parking, cheaper land, all the benefits of the city.

Well so goes the land so goes the taxes so goes the city. That would be the beginning of a death spiral for a large city my dear.

So suburbanites who drive in manually will dictate the course of cites and the companies that reside inside them? You clearly haven't been paying much attention lately. To access the urban labor pool companies are moving in from the burbs in increasing numbers or at least establishing sizable satellite offices in them so talent under 50 doesn't go elsewhere.

 

[Ichinisan]

You can do this now (Uber, Lyft, taxi), don't need autonomous vehicle for this.

Costs will be far less when full autonomy is reached. Right now, it would cost me more to use Uber all the time.

 

[RLGL]

In response to the OP:
NO lets keep this topic alive for days and days

 

[desura]

Just last night I reached a road where google maps told me to go one way but the sign on the street said "one way" and it was the opposite direction.

This is why autonomous vehicles will never really happen. Our roads are way too inconsistent.

What is more likely is autonomous trucks because truck driving is on predictable highways for the most part. You'd still need a person on board though for the 1% of lapses that the system makes, and beyond that for any emergency repairs like a blown tire. Also to defend the cargo against jacking, I'm serious.

If anyone has ever been in a uber ride, it is pretty obvious that automating the ride makes no sense. You'd get riders abusing the crap out of cars, overloading them, stealing stuff, messing with it.

But yeah, I foresee autonomous being useful in certain instances, like traffic jams or highway driving. But I don't see it ever replacing people.

 

[twinrider1]

Can we stop talking about autonomous cars?

lol, apparently not. Every time I enter The Garage this dang thread is on the front page staring at me.

 

[jlee]

In terms of shutting down metro areas to nothing but autonomous cars...

Ah yes, alienating people, the surefire way to success in attracting a vibrant business district in a city.

First city that does this goes bankrupt 30 years later. I'd move out if where I lived did this. Legitimately.

You might think, wow less people and less competition downtown, this is great, cheaper parking, cheaper land, all the benefits of the city.

Well so goes the land so goes the taxes so goes the city. That would be the beginning of a death spiral for a large city my dear.

I'm guessing you've never been in Manhattan.

 

[Pantoot]

Just last night I reached a road where google maps told me to go one way but the sign on the street said "one way" and it was the opposite direction.

This is why autonomous vehicles will never really happen. Our roads are way too inconsistent.

Yet you figured it out. Not sure why you assume that an autonomous car never could.

I do wonder how they will handle those weird edge cases, like a disabled car in the lane with a driver waving people around, or a dui checkpoint (will the car roll down the window for the officer?), but I don't assume that this means that it will never happen.

 

[Ichinisan]

Just last night I reached a road where google maps told me to go one way but the sign on the street said "one way" and it was the opposite direction.

This is why autonomous vehicles will never really happen. Our roads are way too inconsistent.

You don't think more and more roads will be augmented specifically to assist autonomous vehicles? Pretty short-sighted there.

 

[agent00f]

So suburbanites who drive in manually will dictate the course of cites and the companies that reside inside them? You clearly haven't been paying much attention lately. To access the urban labor pool companies are moving in from the burbs in increasing numbers or at least establishing sizable satellite offices in them so talent under 50 doesn't go elsewhere.

It's just wishful thinking. Most people who reside in decent enough cities hardly need to drive much in the first place.

 

[Chaotic42]

It's just wishful thinking. Most people who reside in decent enough cities hardly need to drive much in the first place.

So in the US that's what? DC, New York, and maybe San Francisco?

 

[agent00f]

So in the US that's what? DC, New York, and maybe San Francisco?

I think it's pretty clear that the mindset which accepts the stupidity of automotive commuting perpetuates the US transportation/transit problem. Honestly a giant wave of cars driving into and then out of a dense central hub is dumb as shit, the level of stupid that only exists through faith in the invisible hand in lieu of intelligent planning.

 

[Chaotic42]

I think it's pretty clear that the mindset which accepts the stupidity of automotive commuting perpetuates the US transportation/transit problem. Honestly a giant wave of cars driving into and then out of a dense central hub is dumb as shit, the level of stupid that only exists through faith in the invisible hand in lieu of intelligent planning.

Cities are barely holding it together as it is. Redesigning their infrastructure would be a nightmarish task unless the government really flexed its muscles and just told people and businesses that their land is being taken for rail development. Even then, there's no money. You'd need to keep the commutes short because if they aren't, people will continue to drive. I'm not willing to ride the bus for two hours each way to avoid driving.

 

[agent00f]

Cities are barely holding it together as it is. Redesigning their infrastructure would be a nightmarish task unless the government really flexed its muscles and just told people and businesses that their land is being taken for rail development. Even then, there's no money. You'd need to keep the commutes short because if they aren't, people will continue to drive. I'm not willing to ride the bus for two hours each way to avoid driving.

The bus ride is 2hr precisely because people drive.

 

[jlee]

The bus ride is 2hr precisely because people drive.

Yep. Imagine if 50,000 cars were replaced by 1,000 buses.

 

[darkswordsman17]

Just last night I reached a road where google maps told me to go one way but the sign on the street said "one way" and it was the opposite direction.

This is why autonomous vehicles will never really happen. Our roads are way too inconsistent.

What is more likely is autonomous trucks because truck driving is on predictable highways for the most part. You'd still need a person on board though for the 1% of lapses that the system makes, and beyond that for any emergency repairs like a blown tire. Also to defend the cargo against jacking, I'm serious.

If anyone has ever been in a uber ride, it is pretty obvious that automating the ride makes no sense. You'd get riders abusing the crap out of cars, overloading them, stealing stuff, messing with it.

But yeah, I foresee autonomous being useful in certain instances, like traffic jams or highway driving. But I don't see it ever replacing people.

Yeah, because its completely impossible for cars to learn that, or for us to make them smarter. Or make smarter signs. FWIW, I'm fairly certain they can read street signs and would be likely if nothing else say that you need to take over or would alert you to the issue and have you tell it what to do (and likely simultaneously send an update about the map information being incorrect).

That should be expected in the transition, that you'll still need a qualified driver in the vehicle when its on the actual road (but say in a parking garage for storage it could take over completely).

It'd make more sense to lock down the cargo more than to risk a human life as security. Plus they could drastically change things. They could keep a truck moving at a slower (but not too slow pace, like 50mph) and run it most of the trip (not have to worry about taking breaks for the squishy human inside), which would actually help safety and reduce a lot of the opportunity for thefts (which happen when stopping).

What would there be to steal if they aren't personal vehicles? The assumption is that doing autonomous vehicle ride share would deliberately be handled by non-individuals. I mean I suppose it would make sense to rent your car out when you don't need it (like if you're at work all day, or while you're sleeping), but then why would you own a car in the first place? You'd likely be able to save money by just paying for the rides you need. The thing is, they abuse the vehicle and it can lock them inside and/or take them to police.

And again, arguments that people act like don't also apply to how things already operate. They already do those things to people's cars now. Autonomous vehicles in no way really changes those things (some ways they could make them better, but could be worse, I don't see it making a big difference either way). That's not anything that would stop autonomous cars from being used.

Yet you figured it out. Not sure why you assume that an autonomous car never could.

I do wonder how they will handle those weird edge cases, like a disabled car in the lane with a driver waving people around, or a dui checkpoint (will the car roll down the window for the officer?), but I don't assume that this means that it will never happen.

I'm pretty sure they can figure that out already, its not like that's exactly a crazy scenario, let alone why he thinks it couldn't be pretty quickly managed (I think they're already talking about having control centers where autonomous vehicles can "call" in and get help if it can't figure out what to do and then someone there can look at the cameras and data and tell it what to do. I mean we have people controlling drones firing missiles half the world away, but I guess people can't fathom we could do something similar with cars.

Probably depends on traffic and if it can move over. If it can it'll likely switch lanes and no problem. If it can't it'll probably come to a stop and wait til it can or send an alert that it needs help figuring out what to do.

You don't think more and more roads will be augmented specifically to assist autonomous vehicles? Pretty short-sighted there.

Exactly. They can add cheap sensors, make special markers that are easy for the vehicle to identify. They could even do that in a way that wouldn't be visible to us, so it wouldn't have to also make things worse for human drivers by adding all sorts of other things for our brains to process.

Cities are barely holding it together as it is. Redesigning their infrastructure would be a nightmarish task unless the government really flexed its muscles and just told people and businesses that their land is being taken for rail development. Even then, there's no money. You'd need to keep the commutes short because if they aren't, people will continue to drive. I'm not willing to ride the bus for two hours each way to avoid driving.

Yeah without basically social upheaval, we aren't going to be converting cities to trains. They cost too much, require too much work, and take too long to get a return on the investment to make them worthwhile most places. And even then I think we pretty much have to tax people off the road or subsidize public transport to get people to use them. At least in the US.

I expect more cities will start to ban typical cars in the downtown/dense areas. Replacing them with smaller, slower autonmous vehicles would make a lot more sense and be a lot cheaper than putting in rail.

 

[TeeJay1952]

It is the same greedy monkey of human nature at play. We drive to California once in our youths or hear of someone else's trip and we want to be prepared or have that option available our whole life.
Cars sit in Driveways 23 of 24 hours in a day. We devote so much of our resources for options and improbabilities that we have very little left for when reality strikes.

 

[jlee]

Yeah without basically social upheaval, we aren't going to be converting cities to trains. They cost too much, require too much work, and take too long to get a return on the investment to make them worthwhile most places. And even then I think we pretty much have to tax people off the road or subsidize public transport to get people to use them. At least in the US.

I expect more cities will start to ban typical cars in the downtown/dense areas. Replacing them with smaller, slower autonmous vehicles would make a lot more sense and be a lot cheaper than putting in rail.

Relevant: https://www.nytimes.com/2016/10/14/nyregion/new-jersey-transit-crisis.html

 

[K1052]

Relevant: https://www.nytimes.com/2016/10/14/nyregion/new-jersey-transit-crisis.html

Christie has f'd NJT riders at every possible turn. His lasting legacy will be a trans-hudson commute crisis. Canceling ARC, even with it's flaws, was a horrible mistake. The state has lost so much political leverage that it can't get NY is ignoring them even at the PA.

 

[Ichinisan]

Maybe this deserves its own thread?

I just read about Tesla Network. This might start testing in some cities this year!

https://electrek.co/2016/10/28/tesla-model-3-self-drivin-ride-tesla-network-cost-per-mile/

Why not recruit cars of current owners to build a massive fleet of autonomous vehicles? That's an awesome idea!

Lots of the cars are capable of full autonomy. Tesla already covers the drive system and battery for life on most of the vehicles they've sold. Owners would make some money by allowing their cars to operate autonomously while they eat/sleep/work. Tesla would offer a cheaper and better experience than anything Uber/Lyft is offering.

 

[agent00f]

Maybe this deserves its own thread?

I just read about Tesla Network. This might start testing in some cities this year!

https://electrek.co/2016/10/28/tesla-model-3-self-drivin-ride-tesla-network-cost-per-mile/

Why not recruit cars of current owners to build a massive fleet of autonomous vehicles? That's an awesome idea!

Lots of the cars are capable of full autonomy. Tesla already covers the drive system and battery for life on most of the vehicles they've sold. Owners would make some money by allowing their cars to operate autonomously while they eat/sleep/work. Tesla would offer a cheaper and better experience than anything Uber/Lyft is offering.

There are ~zero cars capable of full autonomy on the road today.

 

[zinfamous]

You don't think more and more roads will be augmented specifically to assist autonomous vehicles? Pretty short-sighted there.

even so, a random change like that--road not going where the map says, an accident, detour, whatever--is instantly updated the first time one car approaches that. The map is changed and all cars on the road at that point and later now have the updated info and adjust accordingly. Again and again, these "impossible" issues are really quite simple and obvious. (yes, the necessary compute power isn't as simple, but the solution is).

Google recently inked that deal with AMD to diversify their deep-learning potential (well, you know in that industry, "diversify" means adding both vendors. 😀).
https://cloud-computing-today.com/2016/11/28/1075598/

Anyway, AMD is now on board with VEGA in these high-end cards, and their compute power is generally superior to what nVidia has always offered. So, there is tons of movement going on in the silicon world that is specifically aimed at the autonomous world.

Metric boatloads of money has already been invested across many sectors in making this happen. The fleets are already happening. It is extremely naive to argue "this will never happen because __x__ is impossible." But, you know, generation after generation, we still endure "Well, technology will never be able to....!" Those are the people that tend to lose out in the new paradigm because they couldn't see the gigantic steam engine baring down on their stone and brick house.

 

[agent00f]

even so, a random change like that--road not going where the map says, an accident, detour, whatever--is instantly updated the first time one car approaches that. The map is changed and all cars on the road at that point and later now have the updated info and adjust accordingly. Again and again, these "impossible" issues are really quite simple and obvious. (yes, the necessary compute power isn't as simple, but the solution is).

Google recently inked that deal with AMD to diversify their deep-learning potential (well, you know in that industry, "diversify" means adding both vendors. 😀).
https://cloud-computing-today.com/2016/11/28/1075598/

Anyway, AMD is now on board with VEGA in these high-end cards, and their compute power is generally superior to what nVidia has always offered. So, there is tons of movement going on in the silicon world that is specifically aimed at the autonomous world.

Metric boatloads of money has already been invested across many sectors in making this happen. The fleets are already happening. It is extremely naive to argue "this will never happen because __x__ is impossible." But, you know, generation after generation, we still endure "Well, technology will never be able to....!" Those are the people that tend to lose out in the new paradigm because they couldn't see the gigantic steam engine baring down on their stone and brick house.

"Deep learning" aka neural networks are data-based algs that nobody can really understand or debug in any meaningful way. They just "seem to work" until they don't. Fine for search classification, less so for driving.

 

[TeeJay1952]

My prediction is that the Auto Companies will become ride services as opposed to just manufactures. Ford/Uber, GM/Google Car, Chrysler/Lyft. Amazon/Do you mind if we drop a package?