Fully automated vehicles - 'not in our lifetime'?

AnotherJoe

HAMISH_MCTAVISH wrote: »

He said that no matter how good the technology, and no matter how good the programming, during the multiple decades where automated cars share the roads with non-automated vehicles or people - at some point an automated car will be forced to make a decision that has no good outcomes.

The choice - to simplify - is in the worst case scenario of an impending accident where the car has to choose between one of two bad outcomes - does the car make a decision that will likely kill it's driver and passengers... or does it choose to kill other road users instead.

You see this "issue/question" posed in recent philosophical debates about full autonomous cars from time to time and its roots are in something called the trolley problem, but i believe its a complete irrelevance, because there is an unfounded underlying assumption to this question, which is that any system can know what the outcomes of its behaviour will be.

First of all, just on what we know about how the universe works, thats actually impossible and second in a world of other "actors" all taking unknowable actions in response to your however carefully planned moves, you cannot predict their actions, either pragmatically or even theoretically and so again you cannot predict the outcome.

So, this is a non issue as AI cars wont be making any such decisions as they dont and cannot know what the outcomes will be.

Dont hit something and try to avid being hit but will be abouta s far as they will take it (though at present they cant even do that so i laugh when people blithely talk about how near AI is when the designers of one deemed safe enough to venture on the streets didnt even manage to enable the vehicle to beep its horn let alone evade a 3mph truck ! )

GreatApe

HAMISH_MCTAVISH wrote: »

:rotfl:

Not in my lifetime.

Buy life insurance if your so certain you haven't more than 20 years to live

kinger101

Herzlos wrote: »

Why would robot taxis have a higher occupant rate than human taxis? Especially if they are much cheaper.

Or are you assuming there would be some kind of ride sharing app to help match up travellers? Because you could do that with human drivers.

People getting taxis don't, on the whole, want to share it especially if it takes them longer to get somewhere.

Unless you're talking about robin taxis that are more like single user pods which take up a fraction of the road and can potentially enter buildings or train stations.

I could see how it may potentially work. The robo-taxis are more likely to communicate with each other, and fill voids by moving to areas where there is greater demand.

AnotherJoe

GreatApe wrote: »

The competition would be human cars vs robo cars. The mayor is much more likely to favor robo cars which on average transport more people. It means cleaner air and less traffic and fewer accidents and it will be useful for children and the disabled.

Not if they are displacing cars and trains they dont.
And thats the only evidence to date.

There's a lot of theory about how this and that will happen, but what we actually see, in practice, in the real world, is that (using Uber etc as a proxy for full autonomy) they increase traffic miles and congestion. mostly due to displacement of public transport miles.

Add to that, if it displaces personal driving, let me know how a car that drives to me and then to my destination, drives for fewer miles than a car i drive directly to my destination.

HAMISH_MCTAVISH

AnotherJoe wrote: »

There's a lot of theory about how this and that will happen, but what we actually see, in practice, in the real world, is that they increase traffic miles and congestion. .

[startproject]

Technologist: "Driverless cars are great - people will call them up then share rides and it'll reduce journeys"

People: "Stuff that - I'm not waiting for it to come and I'm not sharing with anyone either."

Technologist: "But but but - we just blew through $20,000,000,000 of other people's money developing the tech!"

People: "LOL... Sorry, better luck next time"

[/endproject]

kabayiri

GreatApe wrote: »

Buy life insurance if your so certain you haven't more than 20 years to live

What Hamish means is that the latest and best technology is unlikely to hit the outer reaches of civilisation (oops I mean Scotland).

Aberdonians will be running their gas guzzlers to keep the embers burning in the fading oil industry.

GreatApe

Many of you seem to be assuming perfection is needed which is wrong we just need the machines to kill no more than 1 million people each year and that would be an improvement on Human driven vehicles.

GreatApe

AnotherJoe wrote: »

Not if they are displacing cars and trains they dont.
And thats the only evidence to date.

There's a lot of theory about how this and that will happen, but what we actually see, in practice, in the real world, is that (using Uber etc as a proxy for full autonomy) they increase traffic miles and congestion. mostly due to displacement of public transport miles.

Add to that, if it displaces personal driving, let me know how a car that drives to me and then to my destination, drives for fewer miles than a car i drive directly to my destination.

These fleets will be massive probably in the region of 10 million robot taxis in the UK. What that means is you will literally never be more than a few meters from one.

The distance from dropping off one customer to picking up the next will be meters and seconds not like uber today which might have to travel a mile and ten minutes between customers.

Uber only works after a critical mass. If uber didn't exist and you tried to set one up you have a network problem so few drivers that its worthless for customers. But as the network grows so does the usefulness of uber. Uber today is very usable we've given up the second car and just use uber for the times we need two cars at once. These robot fleets will be 100x or more in scale and will thus be much better than uber is today. Much less waiting much less dead mileage and much cheaper too.

As for people not opting to share that's fine too it just means traffic remains roughly where it is. If half choose not to share but half do share traffic goes down. Places like London could also regulate it also only shared trips are offered. In which case traffic should go down lots.

GreatApe

HAMISH_MCTAVISH wrote: »

[startproject]

Technologist: "Driverless cars are great - people will call them up then share rides and it'll reduce journeys"

People: "Stuff that - I'm not waiting for it to come and I'm not sharing with anyone either."

Technologist: "But but but - we just blew through $20,000,000,000 of other people's money developing the tech!"

People: "LOL... Sorry, better luck next time"

[/endproject]

Uber is already valued at $50 billion

Robot taxi fleets are just uber on steroids. Much quicker arriving much less dead mileage and cheaper if you share the trip much much cheaper.

Even if the old like you choose not to use it the kids will love it. Get driven in a nice nearly new Robot taxi for 20p a mile with no parking worries or costs or hand over £6k in costs for a ten year old banger and insurance etc.

kinger101

GreatApe wrote: »

Many of you seem to be assuming perfection is needed which is wrong we just need the machines to kill no more than 1 million people each year and that would be an improvement on Human driven vehicles.

I'm not sure that's true. If a car driven by a human kills someone, it's either put down to human error, bad luck or a combination of both. Each incident is isolation.

If a driverless vehicle kills someone, there's a whole corporation behind it, and questions will be asked whether there was negligence behind it, whether the corporation knew about any potential flaws that might lead to the accident, and if so, why these were not remedied. There are potentially class action lawsuits. The corporations pockets are much deeper for any civil action.

You're right that perfection isn't needed, but the level of safety required is likely to be nearer those we see for the airline and rail industries.