Artificial Intelligence

[-Daniel-]

The last couple of posts have me wondering if I really understand what the real definition for "Artificial Intelligence" is. Now I admit my point of view is tainted by Movies and TV way more than any scholarly work so forgive me.

How should one define "Artificial Intelligence"?

 

[ShawnDriscoll]

How should one define "Artificial Intelligence"?

Until one understands how a brain works, one can't.

 

[heron61]

I'm most opposed to AIs in a car that one owns from the standpoint that someone should be responsible for his own driving.

I can see that PoV, but I even more oppose 10s of thousands of people dying in auto accidents every year, and current indications are that automatic cars will cut those numbers drastically. If automatic cars can cut auto accidents (and far more importantly injuries and deaths due to auto accidents) by 75% (which may well be a low estimate), then my PoV is bring on the automatic cars.

 

[Reynard]

You think they would have fully automated trains, subways light rails by now yet here we are. How about fully automated cruise liners? Planes? If we can't get them to work, how do you miraculously get every passenger carrying vehicle to be automated? Because we're no where near doing any of that stuff with any decent safety factor and I seriously doubt the majority of people in the world are going to trust such systems for a long time.

Sounds like that starships with a brain thread.

Nice thing about Traveller, there are plenty of worlds that have passed the social tests for automated systems including mass transit and people movers that are so transparent they forget it's anything more than efficient machines doing a routine.

 

[ShawnDriscoll]

I'm most opposed to AIs in a car that one owns from the standpoint that someone should be responsible for his own driving.

I can see that PoV, but I even more oppose 10s of thousands of people dying in auto accidents every year, and current indications are that automatic cars will cut those numbers drastically. If automatic cars can cut auto accidents (and far more importantly injuries and deaths due to auto accidents) by 75% (which may well be a low estimate), then my PoV is bring on the automatic cars.

Might as well get rid of the drive and pilot and astrogate skills then. Where's my free starship?

 

[-Daniel-]

Might as well get rid of the drive and pilot and astrogate skills then. Where's my free starship?

It left without you. It realized it didn't need you anymore. :mrgreen:

 

[dragoner]

The last couple of posts have me wondering if I really understand what the real definition for "Artificial Intelligence" is. Now I admit my point of view is tainted by Movies and TV way more than any scholarly work so forgive me.

How should one define "Artificial Intelligence"?

Here is the wiki - https://en.wikipedia.org/wiki/Artificial_intelligence

It is open to interpretation, however.

 

[-Daniel-]

Here is the wiki - https://en.wikipedia.org/wiki/Artificial_intelligence

It is open to interpretation, however.

Wow, I think I understood about half of that. Will need to think on some of the subjects raised there. Thanks for the link.

 

[simonh]

You think they would have fully automated trains, subways light rails by now yet here we are. How about fully automated cruise liners? Planes? If we can't get them to work, how do you miraculously get every passenger carrying vehicle to be automated? Because we're no where near doing any of that stuff with any decent safety factor and I seriously doubt the majority of people in the world are going to trust such systems for a long time.

You must have missed my posts on this up-thread. Google self-driving cars have logged 1.7 million miles on public roads so far with zero accidents caused by them. Several self-driving truck manufacturers have been licensed for commercial trials on public US roads, and those trials are happening right now. They're here.

Now it's just a matter of economies of scale and regulation, but the 'Wright Brothers' moment when self-driving cars became a real working technology actually happened about 3 years ago. They're expected to start becoming available to private owners in around 2020. That's mainly about working out things like manufacturing processes, quality control, developing user friendly UIs, further fine tuning regulations and insurance, etc. The core technology is no longer an issue.

As for acceptance, a lot of cars coming on to the road right now have driver-assist functions like automatic parking and lane change. I think as people become more comfortable and gain confidence in these systems they'll become comfortable with the idea of fully autonomous self-driving cars. In fact I'm not aware of any existing movement to limit or prevent self-driving cars becoming more common. All the press they've been getting has been basically positive and national and US state legislatures are falling over each other to support the industry because they realize the economic benefits. On the other hand organisations and pressure groups for the disabled, especially the blind, are solidly behind opening up the licensing of self-driving cars and it's going to be a tough job for anyone to argue against their case. If you're going to be the politician standing up and saying blind people shouldn't be allowed to have this technology, you'd better have a pretty solid case backing you up, and if they can have them, why not anyone else?

Simon Hibbs

 

[Condottiere]

Chances are, there will be an electronic cure for blindness.

The more interesting issues will be insurance and responsibility.

 

[Reynard]

"Robot cars like Google's use GPS to know their location, and are equipped with sensors, laser illuminating detection and ranging (LIDAR), radar, high-powered cameras and learning algorithm software to know how close they are to another vehicle, as well as detecting the surrounding environment to identify signage, pedestrians and cyclists near the car. "

No, seriously, who below billionaire status can dare think to own such a car when they can't afford a broken down used car?! These will not be affordable robot cars. They will also not be affordable taxis or buses either. Affordable robots with any true abilities are the stuff of science fiction otherwise everyone would own a robot vacuum cleaner. This will also be why true A.I.s will always be protected wards of science institutes, national security agencies and military complexes. Far too expensive for traffic control.

 

[-Daniel-]

The more interesting issues will be insurance and responsibility.

About six months ago NPR had a segment regarding the progress on self driving cars. In that segment the person from Google stated that the most significant hurdle at this time was the Legal Liability issues. The technology is moving forward so fast that the laws can't keep up. Of course that is my paraphrasing and not a direct quote. :mrgreen:

 

[simonh]

No, seriously, who below billionaire status can dare think to own such a car when they can't afford a broken down used car?! These will not be affordable robot cars. They will also not be affordable taxis or buses either.

Let's say the total cost to employ a driver is $30k per year including benefits, admin, etc and the car costs $30k and is operated for 10 years. Ignoring maintenance, fuel, insurance, etc that's a total cost of $330k. If a self-driving car costs $100k that's still a saving of $230k over the vehicle's lifetime. The vehicle cost is all up-front (it would probably be leased, but then you've got interest payments), so the real cost might be higher, but then you have the increased efficiency and duty cycle of the auto-drive. It can driver continuously 24/7 only breaking for vehicle maintenance and software upgrades. It's safer, doesn't argue back and it's driving profile causes less wear on the vehicle and lower fuel costs.

Yes computers and the sensor tech might be expensive, but people aren't cheap either.

Simon Hibbs

 

[simonh]

Chances are, there will be an electronic cure for blindness.

The more interesting issues will be insurance and responsibility.

I agree, for most causes of blindness but it's a matter of timescales as to which technology will become commercially viable first. We're likely to have consumer-ready self-driving cars in 5 years or so, let's say 7 or 10 before they're mainstream. Will we have electronic vision systems good enough to allow a person to drive in the same time frame?

Long term yes, absolutely, but by that time comes self-driving cars may well already be ubiquitous.

Simon Hibbs

 

[Reynard]

"Let's say the total cost to employ a driver is $30k per year including benefits, admin, etc and the car costs $30k and is operated for 10 years. Ignoring maintenance, fuel, insurance, etc that's a total cost of $330k. If a self-driving car costs $100k that's still a saving of $230k over the vehicle's lifetime. The vehicle cost is all up-front (it would probably be leased, but then you've got interest payments), so the real cost might be higher, but then you have the increased efficiency and duty cycle of the auto-drive. It can driver continuously 24/7 only breaking for vehicle maintenance and software upgrades. It's safer, doesn't argue back and it's driving profile causes less wear on the vehicle and lower fuel costs."

By combining a driver with a car, you must be referring to a taxi. I thought we were discussing individual cars and drivers. Even so not many people can afford a taxi ride today nor do taxis go anywhere without the bill becoming exorbitant. Same for buses. Leasing is not an economic option for many except maybe short time renting for special purposes. Most of the robot car dreamers are highly networked urban oriented. Oh and remember. The Google cars are test cars receiving the best of everything to maximize their success. Once on the road, they'll be built and cared for by the lowest bidders and will be run to death to get every last coin from them. You really want a delicate, highly sophisticated electronic cores prone to cheap construction and degradation careening the streets and highways?

"Yes computers and the sensor tech might be expensive, but people aren't cheap either."

Ah, the other issue. I avoid the self checkout lanes because they and other robotic systems that replace humans merely for economic convenience of the owner. They don't actually benefit the humans using them and they don't pay taxes or have to purchase goods and services to keep the economic wheel of a human society turning, they are actually a drag. Robots should never replace humans, they should enhance human production and service as tools. I love those scifi illustrations showing humanity as all artists and scientists surrounded by armies of slave like automations tending to every want and need so a human never lifts a finger in physical labors. That to me is the utopia of the upper class, not humanity.

As much as people with investments in higher education push everyone to get a college education, the reality is very few are good at such studies and many would prefer 'getting their hands dirty'. The greater proportion of the world population are probably in that category. And yet we see certain elements robotizing those careers simply to keep the money flowing to the top. No, robots are not a good answer for many thing from our cars to our total national defense systems.

 

[Condottiere]

I think that once you have smart cars approved, what happens is that you lease or book them for certain periods, which means that a fleet would be operating in your area or neighbourhood.

When you don't need them, they'll be available for other clients, though I suspect that they'll regularly visit the depot for sanitizing.

 

[Anonymous]

Or it'll be like the Barclay's Bikes in London or Velo Libre in Paris, over night a maintenance crew checks the cars as they sit in a holding spot. As the fleet grows they'll have to find dedicated parking space until they're the majority of cars on the road, as long as they're the minority, in places like London where parking space is at a premium people are gonna get pissed at the G Cars stealing available parking spaces.

But this has bugger all to do with AI as started in the OP

:roll:

 

[simonh]

By combining a driver with a car, you must be referring to a taxi.

Yes, as I have stated before it looks like automated taxis are the future of personal transport, not private car ownership.

I thought we were discussing individual cars and drivers. Even so not many people can afford a taxi ride today nor do taxis go anywhere without the bill becoming exorbitant.

Exactly, but most of that bill goes to paying for the driver's time. The cost of leasing a car for the duration of the journey is a small fraction of the cost of the journey, probably even lower than the fuel costs. An automated taxi eliminates the large majority of the costs of the service. Many of us can afford to have a car sitting on our drive, but how many of us can afford a driver on call 24/7? Most of the costs of any commercial road transport go to pay the driver.

You really want a delicate, highly sophisticated electronic cores prone to cheap construction and degradation careening the streets and highways?

That's what regulations are for. You could say something similar about cars themselves.

...I love those scifi illustrations showing humanity as all artists and scientists surrounded by armies of slave like automations tending to every want and need so a human never lifts a finger in physical labors. That to me is the utopia of the upper class, not humanity.

I have a lot of sympathy for this view, as I've discussed in previous posts in detail, this is likely to have a massive impact on the transport industry. Truck driver jobs are reasonably well paid, and support vast support networks of truck stops, diners, motels, etc across the US. Self-driving trucks could have a devastating effect on these businesses.

As for rural areas, there are likely to be cases where private ownership wins and as I've said before some people will always want to have their own car. That's fine, but if a large proportion of people don't then it'll have massive economic effects.

Simon Hibbs

EDIT: I remember seeing a Youtube video of a woman and a young girl getting into their car, driving 30m to to a bus stop at the end of their driveway, the girl gets out of the car at the bus stop, then the woman reverses the car back up the driveway, gets out and goes into the house. I have to concede, driverless taxis may not be the answer for this sort of use case.

 

[dragoner]

But this has bugger all to do with AI as started in the OP

It has jumped the shark, and that even if and when Strong AI appears, that systems for automated autos, or that the replacement of labor, will have long become a reality, in as much as it is now. Though also disconnected to how those ideas would affect the PC's? Driving would be the same in an area without an automated network, an air/raft would have a auto-pilot, but cars have cruise control too. As far as the replacement of labor, ultimately, humans can't do the tasks that machines can, machines have a much greater degree of precision and efficiency. But the PC's won't be laborers for fun.

 

[Anonymous]

This takes us full circle (well, ok, not in this thread) back to Traveller (the setting) being a period sci-fi game.

It was good for when it was written and best seen in that context.

Your context of a post third imperium setting in a sector of your making is ideal for working up the ideas in the OP.

Here's to the rise of the machines and an early retirement from slavery/work...

Is anyone gonna get that? Or are we going to suffer another diatribe of how people need work and full employment is the goal of every decent, honest society?