Suggestions for using AI

[npts2020]

On 8/16/2023 at 7:37 AM, swansont said:

There are people who take mass transit, which you would eliminate, who do so because they can’t afford to own a car. You can’t just assume they can borrow a car (which they aren’t currently doing) and taxis are more expensive than mass transit. And not everybody with a car buys new - lots of them get used cars. What do they do?

How much more per passenger mile do you estimate an automated system would be vs a comparable rail system? Taxis are expensive primarily because you have a driver to pay.

 

On 8/16/2023 at 8:12 AM, Sensei said:

Trains can go much faster than any non-sport car.. See trains in Japan, France or South Korea etc.

Why couldn't an automated system be designed to go just as fast?

On 8/16/2023 at 8:28 AM, swansont said:

I don’t think the issue of more roads was tied to elimination of mass transit. You would definitely have more congestion if you replaced mass transit with individual cars, just based on how much space multiple cars take up as compared to a bus. Or the added cars replacing commuter trains.

And we’ll still have to pay for roads and the cars.

Are we replacing the current system or just putting automated vehicles on it?

 

On 8/16/2023 at 12:20 PM, iNow said:

Who owns and manages this fleet of vehicles large enough to provide rides to every human being?

Since a modern system would likely have to be owned and operated by a government entity of some kind (similar to the postal service), the US will not likely be a leader in this kind of thing. This would not have to preclude people from owning their own vehicle but I can't imagine many people wanting to go through the expense to do so.

[iNow]

5 minutes ago, npts2020 said:

Why couldn't an automated system be designed to go just as fast?

Is this a serious inquiry? Why can’t an auto out vehicle on city streets be designed to go as fast as a maglev train?

6 minutes ago, npts2020 said:

Are we replacing the current system or just putting automated vehicles on it?

Replacement is ideal, but unrealistic. 

7 minutes ago, npts2020 said:

a modern system would likely have to be owned and operated by a government entity

Why? The scale is huge, but Walmart or Lockheed Martin and other similar private entities could surely pull it off. 

[npts2020]

2 minutes ago, iNow said:

Is this a serious inquiry? Why can’t an auto out vehicle on city streets be designed to go as fast as a maglev train?

You are comparing apples to oranges, it is precisely because "it is on city streets". If you had a similar design as a maglev, why couldn't it go just as fast?

Edited August 19, 2023 by npts2020

[iNow]

Turns. People. Other similar obstacles and intrusions. 

[npts2020]

2 minutes ago, iNow said:

Turns. People. Other similar obstacles and intrusions. 

Again you are comparing apples to oranges. What maglev system in the world has any of those things where the train goes as fast as a car?

Edited August 19, 2023 by npts2020

[iNow]

You tell me. You’re the one who asked why autonomous vehicles can’t be designed to go as fast as Japanese trains. 

See also: Relationship between velocity, perceived speed, vestibular response, and passenger comfort: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7559987/

[npts2020]

13 minutes ago, iNow said:

You tell me. You’re the one who asked why autonomous vehicles can’t be designed to go as fast as Japanese trains.

And I asked, with similar system design, why cars couldn't go as fast, your reply is because some people might feel motion sickness? I have never been unfortunate enough to have motion sickness but I hear tell it happens in all manner of transport; boats, planes, human driven cars and even trains.

Edited August 19, 2023 by npts2020

[iNow]

Ok

[swansont]

14 hours ago, npts2020 said:

How much more per passenger mile do you estimate an automated system would be vs a comparable rail system? Taxis are expensive primarily because you have a driver to pay.

Driver take-home is typically 1/3 of the fare. 

https://work.chron.com/much-fare-taxi-drivers-keep-22871.html

Even if you got taxi rates down to $1 per mile (it’s over $2 per mile where I live, plus the initial fee) a 20 mile commute is $20, which is a lot more than the subway fare

 

14 hours ago, npts2020 said:

Why couldn't an automated system be designed to go just as fast?

Tires vs rails, for one. Cars tend to skid on roads in ways that trains don’t. 

 

14 hours ago, npts2020 said:

Are we replacing the current system or just putting automated vehicles on it?

You tell me. You’re the one who said automated cars would mean mass transit would be unnecessary. 

 

 

[npts2020]

9 hours ago, swansont said:

Driver take-home is typically 1/3 of the fare. 

https://work.chron.com/much-fare-taxi-drivers-keep-22871.html

Even if you got taxi rates down to $1 per mile (it’s over $2 per mile where I live, plus the initial fee) a 20 mile commute is $20, which is a lot more than the subway fare

 

Tires vs rails, for one. Cars tend to skid on roads in ways that trains don’t. 

 

You tell me. You’re the one who said automated cars would mean mass transit would be unnecessary. 

 

 

OTOH a cab will pick you up at your front door and drop you off wherever it us you are going at any time of day. Even if one lives next to a bus stop or train station and is going near another stop, there are frequently times and places where mass transit doesn't run. It would be interesting to see what the other 2/3 of the fare is and how much of it is the same for mass transit.

 

Well, humans drive cars around in circles (actually ovals if you want to be technical) at 200 mph (average speed for the Indianapolis 500 this year was over 190 mph including pit stops and yellow flags). Are you telling me that in a system designed similar to railways AI can't go as fast? How many current railroads (especially in the US) go that fast?

 

I would like to see a system anyone can access at any time. Why would we continue to have mass transit with such a personal transit system in place?

[swansont]

1 minute ago, npts2020 said:

Well, humans drive cars around in circles (actually ovals if you want to be technical) at 200 mph (average speed for the Indianapolis 500 this year was over 190 mph including pit stops and yellow flags). Are you telling me that in a system designed similar to railways AI can't go as fast? How many current railroads (especially in the US) go that fast?

It’s really quite silly to address one point of contention by ignoring or discarding all the other issues. You have to do all of them for the proposed system to work. It’s not that you can’t get a car to go that fast, it’s that you can’t get a car to go that fast and do all the other things you’d need to do to have this be like a fast train.

At the indy 500 they scour the track for foreign objects, don’t run in bad weather, and have highly specialized vehicles. How much does it cost to run of those cars? It takes a pit crew to maintain it during the race. You get rid of the driver with an autonomous car but now you need a pit crew every ~100 miles to change the tires and do other maintenance? C’mon.

 

[npts2020]

6 minutes ago, swansont said:

At the indy 500 they scour the track for foreign objects, don’t run in bad weather, and have highly specialized vehicles. How much does it cost to run of those cars? It takes a pit crew to maintain it during the race. You get rid of the driver with an autonomous car but now you need a pit crew every ~100 miles to change the tires and do other maintenance? C’mon.

We'll just have to disagree, then. I say those engineering issues are solvable but apparently you do not. For one thing, vehicles in an automated system will not have to be subject to the same constant high speed turns, braking and accelerating as a race car. For another, there won't be any technical (just physical) restrictions on design, weight, power, etc. and if the system is enclosed to separate the system from pedestrians etc, things like weather and foreign objects should mostly not be a major consideration.

[iNow]

2 hours ago, npts2020 said:

I say those engineering issues are solvable

Yes, most everyone of these issues is solvable. No doubt, but you do your argument a disservice by ignoring the costs, time, effort, and lost opportunities elsewhere that come with solving some of these challenges. 

At some point it’s simply wiser and more mature to accept that it’s not worth it and begin exploring different solutions (or accepting some of the solutions already in place to address these same needs). 

ETA: You're also ignoring new problems this approach would likely create. Solving engineering problems on cars might create social problems and accessibility problems in neighborhoods, for example. 

[studiot]

Suggestions for using AI

 

Combined with suitable human oversight, monitoring and controlling dangerous situations, places or activities such as prisons.

[npts2020]

On 8/19/2023 at 10:12 PM, iNow said:

Yes, most everyone of these issues is solvable. No doubt, but you do your argument a disservice by ignoring the costs, time, effort, and lost opportunities elsewhere that come with solving some of these challenges. 

At some point it’s simply wiser and more mature to accept that it’s not worth it and begin exploring different solutions (or accepting some of the solutions already in place to address these same needs). 

ETA: You're also ignoring new problems this approach would likely create. Solving engineering problems on cars might create social problems and accessibility problems in neighborhoods, for example. 

What are the costs, time, effort and lost opportunities? Also, what are the new problems with society and accessibility? 

[iNow]

57 minutes ago, npts2020 said:

What are the costs, time, effort and lost opportunities? Also, what are the new problems with society and accessibility? 

Depends entirely on the details of what you're proposing. What details would you like to share so we may have that conversation?

[npts2020]

39 minutes ago, iNow said:

Depends entirely on the details of what you're proposing. What details would you like to share so we may have that conversation?

I thought you already knew all of that, thus enabling you to claim:

On 8/19/2023 at 10:12 PM, iNow said:

No doubt, but you do your argument a disservice by ignoring the costs, time, effort, and lost opportunities elsewhere that come with solving some of these challenges. 

At some point it’s simply wiser and more mature to accept that it’s not worth it and begin exploring different solutions (or accepting some of the solutions already in place to address these same needs). 

ETA: You're also ignoring new problems this approach would likely create. Solving engineering problems on cars might create social problems and accessibility problems in neighborhoods, for example. 

 

[iNow]

You want to replace the entire transportation infrastructure with autonomous cars. 

Gonna need much more to go on than that if you want a full review if costs and risks and timelines. 

[npts2020]

23 minutes ago, iNow said:

Gonna need much more to go on than that if you want a full review if costs and risks and timelines. 

Then why are you commenting on it. I really did not intend to hijack the thread beyond saying AI could be used to automate our transportation system. The engineering and cost problems are relatively minor compared to overcoming the "big gubmint is going to take away muh car (freedom)" mindset.

[iNow]

14 minutes ago, npts2020 said:

AI could be used to automate our transportation system.

It’s already in use to automate much of our transportation system, and has been for years. 

[Endy0816]

The big thing would be axing costly signage and signal lights. Millions of dollars could be saved. Could also have vehicles continually flow through intersections. We spend around 58 hours each year just sitting at stop lights.

 

19 hours ago, studiot said:

Combined with suitable human oversight, monitoring and controlling dangerous situations, places or activities such as prisons.

That's definitely a good use case

[wtf]

All,

I could not resist posting this video of a public-spirited philosopher making a subtle critique of the concept of self-driving cars.

 

 

Edited September 14, 2023 by wtf

[Prometheus]

I guess it's easier to take the law into your own hands and destroy things than it is collecting evidence and constructing a well-reasoned case against something.