Would you trust a computer with your life? Apple Car reportedly to have no steering wheel or pedals.

[wanderingfool2]

50 minutes ago, Kisai said:

I've yet to hear of a hacked car, let alone a Tesla, Prius, Leaf or Bolt starting up and driving itself into pedestrians. If we were going to have a widespread hacked car problem that wasn't an inside-job, we would have seen it by now.

Jeep comes to mind in regards to remote hacking.  You can kill the engine of the vehicle remotely (by knowing the sim number).  Tesla is pretty good in terms of keeping up with security, there are "hacked" version of Tesla, but at the same time the realistic threat level is quite a bit smaller...to do something like command a Tesla to accelerate and drive into someone would be stupidly hard compared to so many more trivial ways of injuring.

[mr moose]

3 hours ago, wanderingfool2 said:

I'm not ignoring anything...so what, they "drag and drop" to bring an out of service train into service.  You are foolish if you think that is commanding the train.  Applying the same analogy it's like getting into a self driving vehicle and putting in the destination.  670 people, that includes maintenance, attendants, managers, etc.  It's false to associate all of them with the ability for it to run.

 

yes, you are still ignoring the fact they all have mandatory monitoring and will not run unless a human can intervene.  Until you accept that you rest of your post is pointless.

 

3 hours ago, wanderingfool2 said:

 

I'm done arguing with you, since you obviously aren't applying critical thinking to what you are claiming.

Says the person who refuses to accept that the trains will not run without a human monitoring them and able to intervene.

 

2 hours ago, Kisai said:

 

SNIP

So you agree then, these trains will NOT run unless there are people in the control to monitor them and intervene when necessary.

2 hours ago, Kisai said:

 

So would any vehicle. Buses, Taxi's, fleet vehicle management for places that do construction, plenty of those jobs you'll see there too. If we're going to be super anal about this, this is only the Skytrain, not the Canada-line, who are under a separate company, and use different rolling stock and guideway technology, but use the same automation on the Expo line.

 

And in regards to the "drag and drop" comment. That's not how it works. The "drag and drop" is a simplification of how they can control the trains in realtime. When we had a power failure on the system years ago (caused by people working on the central control system, all the trains stopped immediately, they have batteries) one of the recommendations made was to update the software because they used to have manually enter all the trains. An automated re-entry was available and they had put off ever installing it. So that was installed to the best of my knowledge with the opening of the Evergreen line. The Expo/Millennium/Evergreen line are three lines that operate out of the same facilities using the same guideway. That's what enables them to do this.

 

On a more typical subway like the New York Subway or the Toronto Subway, they physically need to send two people out to the train and move it. There is nothing they can do in the control room but turn the power off. And the NY subway operates 24 hours, where they perform maintenance on the tracks next to powered running tracks.

 

Going back to the topic at hand. Since some of you seem to think that automated technology is new, or wireless is new.

 

You know what pre-dates the Skytrain? The San Francisco BART. In 1974, and they had the same whine about safety then as well.

 

So to put things in context. Automated technology has existed for 46 years in some shape or form and every time safety fears were overblown. Yet it's those continued myths that keeps saddling North America with trash-tier slow driver-driven light rail systems instead of safe automated systems that get people where they want to go quickly.

 

If Apple, or some other company says they want to take the manual controls off. Let them. They have to prove it has the safety redundancy to operate. Even automated trains have manual controls on them for when they need to be moved in an emergency, and I assume that any car produced would likely have the same, even if it's just two buttons and a knob. Cruise control in vehicles since the 90's have been able to accelerate vehicles without anyone putting their foot on the gas peddle, and recent generations of vehicles have lane keeping technology and auto-park. There's also been a widely advertised automatic emergency brake technology as standard for the last 5 years.

 

I've yet to hear of a hacked car, let alone a Tesla, Prius, Leaf or Bolt starting up and driving itself into pedestrians. If we were going to have a widespread hacked car problem that wasn't an inside-job, we would have seen it by now.

 

So far you are ignoring the reality of how the system works and I am not sure to what end.  Do you just not like the idea that there isn't a single AI controlled transport mechanism on the planet that doesn't have human intervention as a compulsory component?

 

 

This entire conversation hinges on the fact that these trains have human intervention as an inescapable part of their functioning,  without that they don't,  and this is solely due to safety.    You guy's can pretend that isn't the case as much as you want and dance around it, but the fact is all transport mechanism have an operator that can hit a kill switch.  This presentation does not and there is no way current tech is good enough for that to be the case.

[Kisai]

8 hours ago, mr moose said:

 

 

 

This entire conversation hinges on the fact that these trains have human intervention as an inescapable part of their functioning,  without that they don't,  and this is solely due to safety.    You guy's can pretend that isn't the case as much as you want and dance around it, but the fact is all transport mechanism have an operator that can hit a kill switch.  This presentation does not and there is no way current tech is good enough for that to be the case.

And all vehicles require maintenance and a garage or dedicated "place" to be parked in when they're not in use. So there is no turning this argument in "fully automated, no humans involved whatsoever"

 

My point is the operation, Skytrain operates without anyone monitoring it. How long it would operate if the entire control center was locked out, and maintenance is neglected is anyone's guess, but my guess would be until significant debris collect in front of the train on the guideway that it catches fire. We've never had a situation like that on the Expo line. The Canada Line however, has had a vehicle "start smoking" due to damage in the guideway. 

 

Forget hacking, the easiest way to stop an automated vehicle without invoking the safety mechanism is to drop a barricade in front of it.

 

 

[mr moose]

16 hours ago, mr moose said:

 

 

This entire conversation hinges on the fact that these trains have human intervention as an inescapable part of their functioning,  without that they don't,  and this is solely due to safety.    You guy's can pretend that isn't the case as much as you want and dance around it, but the fact is all transport mechanism have an operator that can hit a kill switch.  This presentation does not and there is no way current tech is good enough for that to be the case.

I specifically point out human intervention is there because it is a key safety factor, I have done this repeatedly throughout this thread, my point has always been about the ability for humans to intervene at any stage as a matter of safety.  And then you respond with this, for the umpteenth time you completely ignore that fact.

 

7 hours ago, Kisai said:

And all vehicles require maintenance and a garage or dedicated "place" to be parked in when they're not in use. So there is no turning this argument in "fully automated, no humans involved whatsoever"

Tell me, yes or no,  Do you acknowledge that these trains have a controller monitoring them with the ability to intervene at any time?  And if so do you agree it is for safety?

 

7 hours ago, Kisai said:

My point is the operation, Skytrain operates without anyone monitoring it.

No it doesn't,   it may be able to but they will not let it,  So it doesn't.   There is a huge difference between a train being able to run back and forth on a track without human intervention and people actually allowing it to do that.  My point from the very start of this conversation was that no engineer in the world has allowed any form of transport to be without human monitoring and intervention.  I keep telling you and the others that the sky train and all other automated trains have a human controller who watches it and is ready to stop it at any time due to safety.   But you keep sidestepping this fact or ignoring it outright. 

 

The fact of the matter is, any automated transport that lacks human intervention is a legal nightmare waiting to happen and that why there is no such thing anywhere in the world. 

 

 

 

Now if you want to latch on to the fact that they might be able to walk away and the train will keep going, by all means, but that is irrelevant. It does not prove anything, what is important is the fact that no one will do that because it is not safe.  You may as well argue using a brick to hold the gas pedal down is safe because you can do it and the car will keep going.  To bad in the event of an incident you might have trouble removing the brick quick enough, just like if there is an issue with the train and no one is in the control they can't stop it.

 

 

 

[Kisai]

19 hours ago, mr moose said:

 

 

Now if you want to latch on to the fact that they might be able to walk away and the train will keep going, by all means, but that is irrelevant. It does not prove anything, what is important is the fact that no one will do that because it is not safe.  You may as well argue using a brick to hold the gas pedal down is safe because you can do it and the car will keep going.  To bad in the event of an incident you might have trouble removing the brick quick enough, just like if there is an issue with the train and no one is in the control they can't stop it.

 

 

 

The train's operate in UTO mode https://en.wikipedia.org/wiki/Automatic_train_operation , note that there are only 6 such systems in the world in GoA 4 on the wiki.

 

https://www.apta.com/wp-content/uploads/RC17-Keevill_Dave.pdf

 

The bottom photo in page 14 is the Skytrain.

 

That's the skytrain again in the bottom photo.

 

There's also some comments in the presentation about automated cars and why rail is more expensive

 

Note that's a google car on the left and a Tesla on the Right

 

All the questions being asked about automated cars, were already answered about automated trains 35 years ago, and none of the fear-mongering has ever come true, but has instead held back adoption in North America, primarily by government incentives favoring cheap, slow, unprotected rail systems that are no better than a bus.

 

So here we are 35 years later, and people are fear mongering over cars not having features they're used to. Yet how many of those fears have come true? 

https://insideevs.com/news/528123/tesla-fire-versus-gas-car/

 

Vehicle fires are a consequence of poor maintenance, poor manufacturing, or driver carelessness, not because it's automated.

 

https://www.forbes.com/sites/jonathanponciano/2021/04/18/driverless-tesla-behind-crash-that-killed-two-in-texas-officials-believe/?sh=40a09c6e4824

 

Tesla Autopilot is not GoA4, it's GoA2. A driver is required to supervise the vehicle, because it will not avoid obstacles itself.

https://gerberinjurylaw.com/autonomous-vehicle-statistics/

Quote

• With autopilot disengaged and without active safety features, Tesla vehicles were involved in one accident for every 978,000 miles driven in Q1 2021, which is down from every 1.42 million miles driven in Q1 2020.
• With autopilot Engaged, Tesla vehicles were involved in one accident for every 4.19 million miles driven in Q1 2021, which is actually down from one every 4.68 million miles driven in Q1 2020.
• Up to date, there have been a total of 6 deaths from fatal car accidents where the driver was using autopilot.

Also

Quote

Waymo (Google) Self-Driving Car Statistics

• Initially, Google’s self-driving cars were not supposed to have a steering wheel, accelerator pedal, or brake pedal. The California Department of Motor Vehicles, on the other hand, refused to let this be used for testing.
• The National Highway Safety Administration (NHSA) accepted Google’s request for a self-driving automobile without a human driver in 2018. Until then, all self-driving cars had to have at least one human person onboard in case something went wrong.

So Google was aiming for GoA4 out the gate and got downgraded to GoA3 by California DMV and NHSA.

 

Perhaps a future generation of rail rolling stock will adopt some of the camera-based safety features from automated car driving to improve stopping due track intrusions. It would however be unable to do anything about a vehicle coming at it (eg t-boning a train in an unprotected railway crossing) , just like an automated car would be able to absolutely nothing about being t-boned by another vehicle operated manually. 

 

 

 

 

[CarlBar]

@Kisai The stuff you just linked specifically includes operator input in the event of specific rare exceptional situations, (failure modes of various kinds), it explicitly states they're required there. 

[Kisai]

1 hour ago, CarlBar said:

@Kisai The stuff you just linked specifically includes operator input in the event of specific rare exceptional situations, (failure modes of various kinds), it explicitly states they're required there. 

My argument is that people are crying "no automation, it's unsafe" when that's been proven to not be the case for 35 years. It's only people who fear that their ability to drive a car will be taken away from them. With the train's it was unions fearing their jobs would be taken away from them, and that's why Toronto ran their IDENTICAL RT (line 3) system in GoA2 because the drivers union wouldn't allow the TTC to allow UTO.

 

 

Moose is trying to make the argument that automation doesn't exist and automated vehicles are being controlled by people in a remote location like they were drones. That's not the case.

 

For all intents, the only reason there is a control room at all, is the same reason there are traffic control and 911 emergency call centers staffed by humans. You can not control the weather, you can not control accidents caused by people or improperly maintained vehicles. When a station needs to be closed due to a medical emergency, they bypass the station in the control room. If there was no control room, the train doesn't have the human capacity to rubberneck. Human drivers are expected to actually see flashing lights and pull over, but many do not, and cause additional accidents.

 

https://compactequip.com/archives/automated-traffic-control-devices-protect-flaggers-from-driving-hazards/

 

The technology exists to reduce deaths, but many places do not use it for cost reasons because many construction projects are lowest-bidder-wins. 

 

 

 

 

[Arika]

3 minutes ago, Kisai said:

My argument is that people are crying "no automation, it's unsafe"

Then your entire argument for this whole thread is pointless becaue you're not actually responding to anyone.

 

what every single person has been saying is actually "automation without some ability of human intervention is unsafe".

 

Quote

Moose is trying to make the argument that automation doesn't exist and automated vehicles are being controlled by people in a remote location like they were drones. That's not the case.

now you're arguing in bad faith, Moose is saying they are being MONITORED constantly by people

 

even in your own table you provided as some kind of proof of fully automated trains for GOA-4 it STILL has requirements for human intervention, I can guarantee people are watching the system at all times. they don't just hit a big green button that says "GO" on it, then lock that computer away in a closet.

 

Quote

Grade-of-Automation 4 (GoA4)

In this system, trains are capable of operating automatically at all times, including door closing, obstacle detection and emergency situations. On-board staff may be provided for other purposes, e.g. customer service, but are not required for safe operation. Controls are often provided to drive the train manually in the event of a computer failure.

hmmmm if the system is perfect, why would there ever be a computer failure?

another GOA4 train is the Sydney Metro. gee i wonder why they need all of this if it's fully automated?

 

Control room for JUST the automated trains on the sydney metro

 

new additional control room to monitor every single thing on sydney rails (including the automated trains)

 

I wonder if it has something to do with safety and the constantly need for monitoring what the automated trains are doing?

[CarlBar]

What Akira S said. Moose has never been arguing that automation itself is unsafe, but that you have to have some way for a human to step in during extreme cases for it to be safe. And there's no system in the world for transportation where thats not the case. 

[wanderingfool2]

35 minutes ago, CarlBar said:

What Akira S said. Moose has never been arguing that automation itself is unsafe, but that you have to have some way for a human to step in during extreme cases for it to be safe. And there's no system in the world for transportation where thats not the case. 

My issue is that monitoring like that is akin to humans sitting in the vehicle.  There wouldn't be a wheel but again if there is an emergency stop button then who cares?  [Because there will be a point where removing a steering wheel is safer than leaving it in]

 

I'd argue that it's about the logistics of it all..after all, in the Vancouver Control Room they even have the people monitoring the elevator.  Would that entitle me to claim that those elevators aren't needing human intervention?

[Kisai]

2 hours ago, Arika S said:

 

 

hmmmm if the system is perfect, why would there ever be a computer failure?

another GOA4 train is the Sydney Metro. gee i wonder why they need all of this if it's fully automated?

 

Control room for JUST the automated trains on the sydney metro

 

There clearly isn't 150 people driving the Unattended Trains, which was the bad-faith argument Moose provided.

 

There are 342 trainsets on the Vancouver Expo/Millenium/Evergreen line which share the same guideway. Contrary to the discussion here, they are operated as three phases, not "lines", from the same control room using the same signaling.

https://www.infrastructurebc.com/files-4/documents/PBCEvergreen.pdf

https://www.thalesgroup.com/en/canada/press-release/thales-technology-selected-vancouver-skytrain-fleet-expansion

That VOBC is the size of a fridge BTW.

 

https://www.thalesgroup.com/en/worldwide/transport/magazine/whats-driving-driverless-trains

Quote

Trains are automatically controlled by the Operation Control Centre and adjustments can be made quickly to meet changes in demand. Early in the morning, before the start of regular operation, trains are automatically positioned for service. Pre-scheduled in the timetable, or by operator command, trains can even be sent for a wash! All is done automatically with little to no human intervention.

https://www.thalesgroup.com/en/worldwide/magazine/thirty-years-and-counting-vancouvers-driverless-transit-revolution

Quote

Thales has carried out 5 “life-cycle extension projects” for the system, including the 11-kilometer Evergreen Extension to the Millennium Line. The computers on the train, the communications system, and the central operating function have all now been replaced. “Thales has nurtured these systems, replaced them sometimes, in order to keep the system growing while maintaining excellent reliability.”

 

Most of the humans employed by the Skytrain division are for things not related to controlling the train. 

https://www.cbc.ca/player/play/2657995661

 

And a new control center is being built since the SLS is being built as an extension.

https://dailyhive.com/vancouver/translink-new-skytrain-control-centre

 

You can see in the floor plan there are less than 20 seats.

 

 

 

[mr moose]

Honestly, This thread is about a car that is planned to have zero human intervention.  All I have pointed out is that:

 

 

1. Trains systems are easily automated, but NEVER without human monitoring and intervention.

2. The reason they are permanently monitored is for safety because no system is perfect.

3. Roads systems are orders of magnitude more complex and thus by the same standards of safety and with current technology, they are incapable of being automated safely without the ability for human intervention.

 

Not a single post in this thread has been able to prove these points false,  all we have seen is a lot of red herring information (that frankly we all already knew) that does not change the above points.  We have seen people try to argue away the need for human intervention by pointing to google cars (which still have human intervention),  the number of people employed in automated railways (which still run ONLY with human intervention) and some completely irrelevant points about capacity and costs. 

 

I even asked a very simple question which was completely ignored.  I'll ask it again because it seems everyone trying to argue against me can't answer it:

 

Yes or no, Do you acknowledge that there is no train in the world that runs fully automated without a human watching it and able to intervene at any time?

 

 

 

 

[wanderingfool2]

7 hours ago, mr moose said:

1. Trains systems are easily automated, but NEVER without human monitoring and intervention.

2. The reason they are permanently monitored is for safety because no system is perfect.

3. Roads systems are orders of magnitude more complex and thus by the same standards of safety and with current technology, they are incapable of being automated safely without the ability for human intervention.

Again, #2 is more about the logistics, traffic flow etc.  Also #1 is effectively the case when you have 3 people monitoring over 12 trains and no quick "shutdown" button.  Also, for cases where a sensor malfunctioned on a door; so they need to override that to keep the train moving (or to determine if the door is still open).  I consider that still driving without human intervention if 99.999% of runtime it has no humans really involved.  Need I point out that the same people who are monitoring the trains are also monitoring the elevators, and escalators for issues.

 

Like I've said, these kinds of cars will still have an emergency shutdown and I consider that to be "without human intervention".  Again as well, there will reach a point where having a steering wheel is more dangerous than not including it.  Again, when they talk about without human intervention, they aren't saying it will be 100% perfect, just that it's better than what humans could do and most of the time the only human interaction involved is putting in point A to point B (which conveniently is what the train operators are essentially doing).

 

Also

Gee look, a vehicle without any steering wheel or brake pedal running autonomously

[Sarra]

Apple Car: "I am sorry, <user>, but the destination you have selected is not on the Apple Corporate Whitelist for destinations. Please, select another destination."

[CarlBar]

4 hours ago, wanderingfool2 said:

Gee look, a vehicle without any steering wheel or brake pedal running autonomously

 

A research vehicle. Getting approval for that is very different from mass deployment. I also expect that it's being monitored whilst-ever it's in operation, but research projects do somtimes mess the safety stuff up, (usually with terminal career consequences for those involved if it goes wrong).

[mr moose]

7 hours ago, wanderingfool2 said:

Again, #2 is more about the logistics, traffic flow etc.  Also #1 is effectively the case when you have 3 people monitoring over 12 trains and no quick "shutdown" button.  Also, for cases where a sensor malfunctioned on a door; so they need to override that to keep the train moving (or to determine if the door is still open).  I consider that still driving without human intervention if 99.999% of runtime it has no humans really involved.  Need I point out that the same people who are monitoring the trains are also monitoring the elevators, and escalators for issues.

 

Like I've said, these kinds of cars will still have an emergency shutdown and I consider that to be "without human intervention".  Again as well, there will reach a point where having a steering wheel is more dangerous than not including it.  Again, when they talk about without human intervention, they aren't saying it will be 100% perfect, just that it's better than what humans could do and most of the time the only human interaction involved is putting in point A to point B (which conveniently is what the train operators are essentially doing).

 

Also

Gee look, a vehicle without any steering wheel or brake pedal running autonomously

You still haven't said anything that contradicts what I have said.  It appears you are reading into my posts and concluding some argument that I never made.

 

 

[wanderingfool2]

6 hours ago, CarlBar said:

 

A research vehicle. Getting approval for that is very different from mass deployment. I also expect that it's being monitored whilst-ever it's in operation, but research projects do somtimes mess the safety stuff up, (usually with terminal career consequences for those involved if it goes wrong).

The general point is that it has happened.  There's no point in saying it's "monitored" because that's the same case as having no steering wheel sitting in a self-driving car.  Again, I have constantly said that when it's statistically better to remove the wheel it should be removed.

 

5 hours ago, mr moose said:

You still haven't said anything that contradicts what I have said.  It appears you are reading into my posts and concluding some argument that I never made.

Like I said, under your logic and argument the elevators then can't be operated because of safety reasons...since they are monitored and brought in and out of service by the same controllers that deal with the trains.  Claiming it will never happen because of safety is inherently wrong because it will be all about statistics.

 

I clearly said the reasoning for #2 which does contradict, they are there because they are the ones that try keeping trains on schedule and dispatching people to deal with situations on the station...You seem to fail to grasp that it's the same concept of someone sitting in a car (even without a steering wheel, with only an emergency stop button)..because guess what, yes to an extent the rail might be partially safety, but one of the main reasons as well (which is even mentioned subtly in the video you posted) is that they keep things flowing (bringing in more vehicles in response to added demand).  I'm not going to pick apart every argument from someone who literally thought there were 500 people monitoring for "safety" the Expo/Millennium lines when the current control room holds less than 20 people.

 

Waymo is already effectively driving around "without a steering wheel"...but not really to the point where it doesn't get into situations it doesn't know how to handle.  Even then a remote operator can come in and get the car "unstuck" sometimes without someone getting into the drivers seat.  Are we at a point where steering wheels should be removed, no (I'm not saying that at least), but like I've been saying there will be a time where the AI driver is safer than a human driver statistically, and the concept of removing active steering will be removed.

 

Will the vehicle still be in fatal accidents when that happens, absolutely.  May there be a few cases where human driver would have prevented it, sure...but if it saves more lives than are lost then it is worthwhile.  [Because again, having manual override in steering can be more dangerous than not having it].  This is again talking about future technology, but really it's a lot closer than I think most people believe.  Autopilot, on a highway, already has less accidents per mile than humans.  The newest FSD, while still stupid in areas, has made leaps and bounds compared to what it was a year ago.  From where it could barely navigate streets, to now even being able to do some of the drives with the driver only hitting the accelerator to prompt the car to proceed.

[saltycaramel]

Self driving cars sound like a problem that could be solved on an infinite time scale and with infinite computational power, honestly. Doesn’t sound unsolvable a priori.

 

Humans aren’t that good. People of all shapes, intelligence, reaction times, eyesight, alertness, mood, mental status, etc. drive on our roads and life goes on. 

 

Would you trust 1 human with your life? Nope, that’s why we need to put 2 in the cockpit or at the head of a train. And sometimes even that isn’t enough. 

 

The question is if this “infinite time scale” is within our lifetimes or not. Two of the most advanced consumer tech companies seem to be convinced that it is. Or they seem to think that if you have infinite money for R&D (like Apple), that’s the problem you need to tackle right now.

[mr moose]

2 hours ago, wanderingfool2 said:

The general point is that it has happened.  There's no point in saying it's "monitored" because that's the same case as having no steering wheel sitting in a self-driving car.  Again, I have constantly said that when it's statistically better to remove the wheel it should be removed.

 

Like I said, under your logic and argument the elevators then can't be operated because of safety reasons...since they are monitored and brought in and out of service by the same controllers that deal with the trains.  Claiming it will never happen because of safety is inherently wrong because it will be all about statistics.

 

I clearly said the reasoning for #2 which does contradict, they are there because they are the ones that try keeping trains on schedule and dispatching people to deal with situations on the station...You seem to fail to grasp that it's the same concept of someone sitting in a car (even without a steering wheel, with only an emergency stop button)..because guess what, yes to an extent the rail might be partially safety, but one of the main reasons as well (which is even mentioned subtly in the video you posted) is that they keep things flowing (bringing in more vehicles in response to added demand).  I'm not going to pick apart every argument from someone who literally thought there were 500 people monitoring for "safety" the Expo/Millennium lines when the current control room holds less than 20 people.

 

Waymo is already effectively driving around "without a steering wheel"...but not really to the point where it doesn't get into situations it doesn't know how to handle.  Even then a remote operator can come in and get the car "unstuck" sometimes without someone getting into the drivers seat.  Are we at a point where steering wheels should be removed, no (I'm not saying that at least), but like I've been saying there will be a time where the AI driver is safer than a human driver statistically, and the concept of removing active steering will be removed.

 

Will the vehicle still be in fatal accidents when that happens, absolutely.  May there be a few cases where human driver would have prevented it, sure...but if it saves more lives than are lost then it is worthwhile.  [Because again, having manual override in steering can be more dangerous than not having it].  This is again talking about future technology, but really it's a lot closer than I think most people believe.  Autopilot, on a highway, already has less accidents per mile than humans.  The newest FSD, while still stupid in areas, has made leaps and bounds compared to what it was a year ago.  From where it could barely navigate streets, to now even being able to do some of the drives with the driver only hitting the accelerator to prompt the car to proceed.

 

Huh?  I really wish you'd actually address my point if you think it is wrong rather than dancing around it.

 

I think I've made the point too many times now for your responses to be genuine.

 

 

[jagdtigger]

2 hours ago, saltycaramel said:

Humans aren’t that good.

AI's even worse, their sensors just cant cope with bad weather.... There are plenty of conditions where a human driver can safely continue but an AI would flip out because their sensors cant see anything.

[saltycaramel]

4 minutes ago, jagdtigger said:

AI's even worse, their sensors just cant cope with bad weather.... There are plenty of conditions where a human driver can safely continue but an AI would flip out because their sensors cant see anything.

 

Humans are still gonna be humans in 20 years.

 

AI and raw computational power at our disposal will be so exponentially higher that we can’t even fathom it right now (humans are also bad at imagining exponential growth, among other things, as covid unfortunately showed us).

[jagdtigger]

1 hour ago, saltycaramel said:

AI and raw computational power at our disposal will be so exponentially higher that we can’t even fathom it right now (humans are also bad at imagining exponential growth, among other things, as covid unfortunately showed us).

That is just wishful thinking based on a wild projection that we wont hit any brick walls on the way....

[wanderingfool2]

7 hours ago, mr moose said:

 

Huh?  I really wish you'd actually address my point if you think it is wrong rather than dancing around it.

 

I think I've made the point too many times now for your responses to be genuine.

 

 

 

23 hours ago, wanderingfool2 said:

Again, #2 is more about the logistics, traffic flow etc. 

That alone addresses it

 

But fine, you want me to pick it apart.  I don't think you grasp what actually happens inside the control room.

1) The automation that Canada Line is at the level of a self-driving car without a steering wheel or pedals (like I've always said, and you yourself never addressed it) [Which I've addressed multiple times].  So there isn't a point in bringing this up.  Your point is partially wrong because see my point 2.

2) The reason for the people is primarily about the logistics, traffic flow, etc. [Again I've addressed this multiple times]

3) I've already addressed this.  It's not comparable to "trains", as human mistakes can make it more dangerous than automation.

The "level" of no human intervention at that stage would be similar to a self driving car.  The above aren't red herrings, your whole [oh the trains] argument is the red herring.

 

It's akin to me arguing any automated car will require a phone line, because elevators and trains require it for safety.

 

[GhostRoadieBL]

this is very dependent on how the car is implemented. The lack of details on the implementation of the car (hopeful thinking here) is it would be limited to their own facilities similar to how some Japanese hotels have driverless carts for luggage and even some with driverless golfcarts for taking you to your room. There's really no saying what the 2025+ driverless car regulations will be and obviously advancements in the tech will only make things more reliable.

 

Still only going to buy one if it holds the company responsible for any and all accidents and has a "save the occupants at all costs" mode permanently enabled.

[Kisai]

This is an example of an automated vehicle not being designed for the climate. These were basically designed for California.

 

https://apnews.com/article/coronavirus-robot-food-delivery-technology-business-5b7dbb782f89ef3d586eefb9a33d5e1c

 

And unlike the train conversation earlier, these are actually monitored and controlled like a drone.

Quote

Remote operators keep tabs on multiple robots at a time but they say they rarely need to hit the brakes or steer around an obstacle. When a robot arrives at its destination, customers type a code into their phones to open the lid and retrieve their food.

 

The robots have drawbacks that limit their usefulness for now. They’re electric, so they must recharge regularly. They’re slow, and they generally stay within a small, pre-mapped radius.