And spends a huge amount of energy pushing air aside.
As do the newer Japanese HSRs and maglev systems.
Nose shape doesn't actually matter (except to keep the ends on the ground). The vast majority of HSR energy losses come from skin friction along the sides of the train (At 200mph+ it's a very real phenomonenon for HSRs)
If you can partially evacuate a tube and _keep_ it partially evacuated for far less energy than running HSR then the speed is a bonus.
"but in order to do it we are going to build near your homes and then tell you that your property value went up"
Bollocks.
If you build a railway line across open countryside, businesses and homes will be built where they have access to the line and property prices go up from "unimproved" values. Most american railways made more money selling off parcels of land along the lines than from the lines themselves. This method was used in many other countries to incentivise line building (In most cases as soon as they ran out of land to sell, the railways went bust and the lines were nationalised, but the important part of the work had already been done.)
Russia never was. After centuries of being invaded they have a nationalistic tendency to assume the porcupine theory of defense.. It suits Putin and friends very well to play up the western threat to keep the russian population pliable (they myth of the strong man keeping the much richer and more powerful west at bay) and it suits the west to play up the Putin threat to keep us distracted from the hands raiding our pocketbooks.
"During that time (after 12 years of construction), we'd probably be hovering above ground with our own vehicles called HoverCraft using autopilot. "
You missed:
Energy consumption. Aircraft and most vehicles have to carry their own energy sources. Hyperloop (and HSR) can have the energy fed in along the length of the line.
Hyperloop is expensive to setup but cheap(ish) to run and FAST. Rail is expensive to setup but expensive to maintain and would be intermediate speed
Air transport is cheap to setup (runways) but expensive to operate (fuel), limited in payload and would be slower than Hyperloop over long distances. Aircraft are inherently fragile due to the need to be lightweight - which hyperloop pods don't need and can therefore be heavy enough to withstand sabotage more easily (armouring aircraft against most bombs is relatively straightforward but the tradeoff is a weight penalty that airlines don't want to pay - and liquid binary explosives are so hard to make that the vibration in an aircraft/train/ship/hyperloop pod is more than sufficient to ensure it will fail.
Energy supply is critical. As the world goes more-electric the amaount of hydrocarbons consumed by aircraft is going to become more and more frowned on. Hyperloop and HSR have the advantage that they can practically be fuelled by renewables or nuclear power (yay LFTRS).
What works (or not) in the USA with its very low population densities outside the coastal areas is entirely different to most of the rest of the world.
I can see a transport structure where HSR feeds into primary hyperloop lines with drayage from there and aircraft only used for transoceanic transport where ships are too slow. The important part becomes the switching systems to get hyperloop pods where they need to go with minimal disruption and transferring between HSR and Drayage with minimal disruption (we all hate airport transfers and even getting on/off HSRs is a pain in the neck as anyone who's had to change trains at Paris or Brussels will attest.)
"Never rely on a mathematical function in a software package that you can't cross check using other tools and techniques. It doesn't matter whether it is 'widely tested and reliable', which often just means other people assume it works. Software can and does have all kinds of untested corner cases and obscure bugs."
The same applies to just about anything in software.
I've had greybeards tell me they will not use XYZ newer package because ABC has been around forever and and "is widely tested and reliable" - which as you say, means that people assume that it works and mostly it's never been checked for bugs. I make a point of rubbing their faces in it when things like a 30 year old ntpd bug show up.
My argument for XYZ is "it's new, it's distrusted, it's been pulled apart by security researchers, here's the report on it. When you can find the security report for ABC, let me know" - knowing full well that in a large number of cases that when security researchers turn their attention to old code (particularly old networking stuff) a dozen things usually fall out within seconds of running a sanity checker over it.
"conveniently ignoring all that with the Hyperloop."
Hyperloop pods are small and the nature of the beast means any passenger-borne explosive is unlikely to damage the transportation system. An explosion into a partial vacuum isn't going to have much effect.
Passenger scheduling requirements conflict greatly with freight operations. One or the other has to have priority on a line and when they run at markedly different speeds you have a major problem on your hands.
One proposal is to sling hyperloop above a rail corridor so that passenger and freight ops are segregated, but the speeds of hyperloop frequently mean this isn't practical unless you want to expose your passengers to multiple G turns.
My opinion (and only mine, yours may differ) The _only_ way that Hyperloop is viable for freight operations is if the lines are large enough to take the largest standard intermodal shipping containers and the pods that carry them can run at the same speed as everything else.
Anything which requires repacking reduces efficiency and makes Hyperloop uneconomic for freight. The only reason continual repacking is tolerable for airfreight is that it's only used for high value or perishable operations where the extra cost can be tolerated.
Bear in mind that hyperloop as proposed in the USA is built by private consortiums, whilst railway systems in Europe tend to be government-operated things (almost all railway companies in Europe went bust at one point or another due to the inefficiencies that using rail for passenger transport imposes on the entire system.)
The question is if a private organisation is willing to step forward and take on the risk.
"If you want a car that moves sideways, a better way to do it is to just make the steering pivot a full 90cdegrees. "
Given that you'd need this on all 4 wheels, it would probably be simpler to use some kind of droparm with castors mounted at 90 degrees to the drive wheels. The speed and distance requirements aren't onerous, although the contact pressures could be a bit high and if there's a high camber the drive power required could get awkward.
Tires actually give lots of warning before this happens and the primary cause is underinflation (which is why warnings systems are now mandatory in many areas)
"parts can break down unexpectedly"
Only if you ignore the warning signs. With more sensors and an AI at the helm instead of a "press on regardless" monkey, this is far less likely to happen (and auto insurance never covers parts breakdowns, only what happens if that causes a crash. Warranties and breakdown cover are separate items)
"If you live in Quebec, Canada then your vehicle might fall into a giant pothole"
That's the liability of the land or highway owner, although if you drive into it, then it's your fault for not paying attention.
"There's also incline weather"
an AI won't press-on regardless of the weather conditions (virtually every crash I've seen in snow has been the result of the driver failing to take conditions into account.)
" animals"
And pedestrians - with an AI paying 100% attention 100% of the time, the odds of being startled by a moose in your headlights is low. The car will likely stop before you're even aware of the danger.
"falling trees, power poles".
If you drive into one, then it's your own damned fault for not paying attention. See above about AIs
If one falls on your parked car then the liabilty is with the owner of the pole/tree, etc.
The number of _genuine_ accidents on the road is tiny. 99.9+% are the result of operator error, even when there are sinkholes, etc.
When a stationary smart vehicle is damaged, it will be awake and recording details instantly - which makes identifying the hit-and-run driver or the toerag who keyed your car that much easier. Only the terminally stupid will attempt to break into or vandalise an AI car, given vastly increased odds of being caught.
Even pothole damage is likely to be vastly reduced as an AI will assess the liklihood of causing damage & slow down to suit, PLUS record the exact location and image of the damage so any costs can be clawed back from the roadowner.
"But don't forget traffic cops and the revenue that citations bring municipalities. And the body shops that repair damaged cars. And driving professions. It will be interesting how this impacts all these segments of the economy and what the ripples are."
There are at least 400 million jobs worldwide which will be directly affected by automated driving - that's just the ones at the wheel.
Bear in mind that this change has already happened once (working horses) we can see what will happen over time is that drivers will retire without new replacements coming in. The USA is already short of at _least_ 30,000 long-haul truckers so there's already sufficient economic motivation to press ahead with automation.
"Actually, I figure that it'll end up being insurance, more than anything else, that destroys the market for human driven cars once self-driving versions can do everything necessary except maybe off-roading."
That, and once computers can demonstrate adequate(*) driving skills, humans are going to face much tougher driving tests than are currently the norm.
(*) Most hoomans drive badly. Even good drivers have bad days and the level of most driving tests isn't much more than "You can point it in the right direction, not crash into anything else and don't bully on other drivers. Here's your license." - which is often the high point of many peoples' driving skills.
"People keep forgetting about the comprehensive part, what makes them think that the manufacturer is going to pay if a tree branch falls on it or a rock gets kicked up into the windshield? "
Claims of this nature are relatively rare. You're looking at a $30-50 premium covering it.
The vast majority of claims are down to driver error (which is effectively eliminated) or vandalism/burglary (which won't be particularly smart when these cars are bristling with sensors and can snap clear photos of the miscreants that are likely to be beamed back to base even if the car is stolen for parts - my own insurance company makes it clear that in the event of vandalism, if the culprit can be identified they'll go after him/her for 100% of all costs regardless of any criminal charges which might be added - and they've followed through on that (complete with discounts for dashcams or other monitoring systems which can detect/record damage whilst parked).
If a tree branch falls on your car, you might make a claim via your insurer but the actual liability and cost falls on the owner of the tree. Your insurance company is merely a facilitator of getting the claim resolved more quickly. Ditto in road subsidence issues.
"Most folks will probably still want comprehensive car insurance for other things"
'most folks' will probably stop owning a car once the cost overhead of paying a driver goes away. Even for the ones who do keep cars, 2-3 vehicle households will probably become single-vehicle ones as the single car performs multiple functions or fleet services are called in as needed.
Barcelona is estimating that the number of vehicles on its roads will drop by 80% once automation becomes common for this reason. I have no reason to doubt their reasoning on this issue (in many cities you don't leave personal stuff in your vehicle because it's likely to be broken into and items stolen. That's another disincentive to private ownership and a plus point for fleet services)
A lot of "don't swerve" has to do with the fact that it's easier to brake in a straight line.
ABS and ESC changes that dynamic a _little_ - they do allow swerving & etc without going sideways, but the primary problem is between the pedals and the seat. We simply don't react fast enough to deal with situations that occur at standard urban speeds (those speeds are too fast for our monkey-brains. Seriously)
Automated vehicles likely won't need to swerve or brake hard because they're communicating with each other but also because even when in standalone mode they'll be paying 100% attention 100% of the time and not following too close (the primary cause of almost all urban road crashes) or driving too fast for the conditions (the primary cause of most highway road crashes)
"Given a sidewalk where people can unexpectedly enter the road"
People _don't_ unexpectedly enter the road, short of erupting from a shielded location and even then there's usually some sort of warning (like a ball bouncing onto the road - a classic warning that it's time to slow down and prepare for a kid on the road) They telegraph their intentions pretty clearly even if not aware they're doing so.
The problem is car _drivers_ expect to have right of way and _expect_ that pedestrians won't step onto the road. The truth of the matter is that the curb creates a "safety demarcation line" which only exists in the mind of the driver.
A self driving car will not only stop more quickly, it will probably already be slowing down before the pedestrian has even stepped onto the road. Additionally, without an impatient monkey at the controls, more suitable speed limits (20mph in residential areas) will become the norm. The tradeoff of smoother traffic flow will more than likely eliminate any potential time loss on most journeys and pedestrian overpasses will probably become a thing of the past (they're a stupid idea. People don't use them and the main reason for them existing is because drivers tend to have tunnel vision above 30mph)
Machines have 360 degree vision AND they don't spend time looking at the cute ass on that female walking down the sidewalk instead of what's happening on the road in front of them, or looking at the pedestrian looking like he _might_ step onto the road on the right whilst missing the 18 wheeler pulling out on the left (or the wobbly cyclist who demands your entire attention to the point of missing the pedestrian stepping out from the other side of the road)
WRT the "google car hit a bus" non-story from a few months back, the car assumed (rightly) that it had right of way. The software didn't take into account that the bus driver would bully his way through regardless (If you look, you'll see the car stopped well before the collision). If both were automated it wouldn't have happened and newer algorithms take more account of human drivers refusing to yield to right of way.
FWIW the average traffic speed in London is 10-12mph and it's a lot slower in peak hours. Parking is nearly impossible (or extremely expensive). When you factor these in (and that driving in such conditions is stressful) there's a strong argument that people will flock to automated vehicles and car services. Now factor in that large european cities account for 30% of car sales in europe, that's a large chunk of lost sales for manufacturers - balanced by vastly increased sales in developing countries.
Even if the government revokes copyright (or declares eminent domain), they cannot force the company to release the source code.
Without that, support is limited to instructing people how to do particular tasks and once something crops out the software can't handle, it's game over.
MS-DOS was/is y2k compliant. It was windows 3(.1)(1) that wasn't.
People are confusing support contracts with use licenses.
Plenty of software has time-limited use rules (annual licensing) - that used to be the norm until MS turned up with "good enough" software that was cheap and one-time-payment for use.
Don't confuse that with a one-time payment for use and annual support contracts (which is what Oracle, and even Redhat use)
"Or does a 5MW nuclear reactor sound very very small. Like small enough that it is one of those experimental ones that produce medical isotopes."
Small is relative.
5MW is enough to drive a nuclear sub or power a small military base. Civil reactors are in the 600-1400MW range (electrical, thermal is 3.5 to 4 times that)
The Oak Ridge Molten Salt experiment was 8MW (thermal)
Medical reactors only produce a few hundred watts of heat.
Slashdot Mirror
"the Shanghai maglev manages over 300mph"
And spends a huge amount of energy pushing air aside.
As do the newer Japanese HSRs and maglev systems.
Nose shape doesn't actually matter (except to keep the ends on the ground). The vast majority of HSR energy losses come from skin friction along the sides of the train (At 200mph+ it's a very real phenomonenon for HSRs)
If you can partially evacuate a tube and _keep_ it partially evacuated for far less energy than running HSR then the speed is a bonus.
"but in order to do it we are going to build near your homes and then tell you that your property value went up"
Bollocks.
If you build a railway line across open countryside, businesses and homes will be built where they have access to the line and property prices go up from "unimproved" values. Most american railways made more money selling off parcels of land along the lines than from the lines themselves. This method was used in many other countries to incentivise line building (In most cases as soon as they ran out of land to sell, the railways went bust and the lines were nationalised, but the important part of the work had already been done.)
"As long as Russia remains a mad dog"
Russia never was. After centuries of being invaded they have a nationalistic tendency to assume the porcupine theory of defense.. It suits Putin and friends very well to play up the western threat to keep the russian population pliable (they myth of the strong man keeping the much richer and more powerful west at bay) and it suits the west to play up the Putin threat to keep us distracted from the hands raiding our pocketbooks.
"During that time (after 12 years of construction), we'd probably be hovering above ground with our own vehicles called HoverCraft using autopilot. "
You missed:
Energy consumption. Aircraft and most vehicles have to carry their own energy sources. Hyperloop (and HSR) can have the energy fed in along the length of the line.
Hyperloop is expensive to setup but cheap(ish) to run and FAST. Rail is expensive to setup but expensive to maintain and would be intermediate speed
Air transport is cheap to setup (runways) but expensive to operate (fuel), limited in payload and would be slower than Hyperloop over long distances. Aircraft are inherently fragile due to the need to be lightweight - which hyperloop pods don't need and can therefore be heavy enough to withstand sabotage more easily (armouring aircraft against most bombs is relatively straightforward but the tradeoff is a weight penalty that airlines don't want to pay - and liquid binary explosives are so hard to make that the vibration in an aircraft/train/ship/hyperloop pod is more than sufficient to ensure it will fail.
Energy supply is critical. As the world goes more-electric the amaount of hydrocarbons consumed by aircraft is going to become more and more frowned on. Hyperloop and HSR have the advantage that they can practically be fuelled by renewables or nuclear power (yay LFTRS).
What works (or not) in the USA with its very low population densities outside the coastal areas is entirely different to most of the rest of the world.
I can see a transport structure where HSR feeds into primary hyperloop lines with drayage from there and aircraft only used for transoceanic transport where ships are too slow. The important part becomes the switching systems to get hyperloop pods where they need to go with minimal disruption and transferring between HSR and Drayage with minimal disruption (we all hate airport transfers and even getting on/off HSRs is a pain in the neck as anyone who's had to change trains at Paris or Brussels will attest.)
"Never rely on a mathematical function in a software package that you can't cross check using other tools and techniques. It doesn't matter whether it is 'widely tested and reliable', which often just means other people assume it works. Software can and does have all kinds of untested corner cases and obscure bugs."
The same applies to just about anything in software.
I've had greybeards tell me they will not use XYZ newer package because ABC has been around forever and and "is widely tested and reliable" - which as you say, means that people assume that it works and mostly it's never been checked for bugs. I make a point of rubbing their faces in it when things like a 30 year old ntpd bug show up.
My argument for XYZ is "it's new, it's distrusted, it's been pulled apart by security researchers, here's the report on it. When you can find the security report for ABC, let me know" - knowing full well that in a large number of cases that when security researchers turn their attention to old code (particularly old networking stuff) a dozen things usually fall out within seconds of running a sanity checker over it.
"conveniently ignoring all that with the Hyperloop."
Hyperloop pods are small and the nature of the beast means any passenger-borne explosive is unlikely to damage the transportation system. An explosion into a partial vacuum isn't going to have much effect.
Not just track conditions.
Passenger scheduling requirements conflict greatly with freight operations. One or the other has to have priority on a line and when they run at markedly different speeds you have a major problem on your hands.
One proposal is to sling hyperloop above a rail corridor so that passenger and freight ops are segregated, but the speeds of hyperloop frequently mean this isn't practical unless you want to expose your passengers to multiple G turns.
My opinion (and only mine, yours may differ) The _only_ way that Hyperloop is viable for freight operations is if the lines are large enough to take the largest standard intermodal shipping containers and the pods that carry them can run at the same speed as everything else.
Anything which requires repacking reduces efficiency and makes Hyperloop uneconomic for freight. The only reason continual repacking is tolerable for airfreight is that it's only used for high value or perishable operations where the extra cost can be tolerated.
Bear in mind that hyperloop as proposed in the USA is built by private consortiums, whilst railway systems in Europe tend to be government-operated things (almost all railway companies in Europe went bust at one point or another due to the inefficiencies that using rail for passenger transport imposes on the entire system.)
The question is if a private organisation is willing to step forward and take on the risk.
"If you want a car that moves sideways, a better way to do it is to just make the steering pivot a full 90cdegrees. "
Given that you'd need this on all 4 wheels, it would probably be simpler to use some kind of droparm with castors mounted at 90 degrees to the drive wheels. The speed and distance requirements aren't onerous, although the contact pressures could be a bit high and if there's a high camber the drive power required could get awkward.
kind of like wheel dollies, without the dolly.
The crash victim was known to record everything whilst driving.
What happened to his camera?
and, for that matter, how would a trucker in an 18 wheeler _hear_ a video playing in a car, over his own engine noise?
"Tires can blow"
Tires actually give lots of warning before this happens and the primary cause is underinflation (which is why warnings systems are now mandatory in many areas)
"parts can break down unexpectedly"
Only if you ignore the warning signs. With more sensors and an AI at the helm instead of a "press on regardless" monkey, this is far less likely to happen (and auto insurance never covers parts breakdowns, only what happens if that causes a crash. Warranties and breakdown cover are separate items)
"If you live in Quebec, Canada then your vehicle might fall into a giant pothole"
That's the liability of the land or highway owner, although if you drive into it, then it's your fault for not paying attention.
"There's also incline weather"
an AI won't press-on regardless of the weather conditions (virtually every crash I've seen in snow has been the result of the driver failing to take conditions into account.)
" animals"
And pedestrians - with an AI paying 100% attention 100% of the time, the odds of being startled by a moose in your headlights is low. The car will likely stop before you're even aware of the danger.
"falling trees, power poles".
If you drive into one, then it's your own damned fault for not paying attention. See above about AIs
If one falls on your parked car then the liabilty is with the owner of the pole/tree, etc.
The number of _genuine_ accidents on the road is tiny. 99.9+% are the result of operator error, even when there are sinkholes, etc.
When a stationary smart vehicle is damaged, it will be awake and recording details instantly - which makes identifying the hit-and-run driver or the toerag who keyed your car that much easier. Only the terminally stupid will attempt to break into or vandalise an AI car, given vastly increased odds of being caught.
Even pothole damage is likely to be vastly reduced as an AI will assess the liklihood of causing damage & slow down to suit, PLUS record the exact location and image of the damage so any costs can be clawed back from the roadowner.
Your assumption is that insurance companies will need to maintain the same level of staffing.
They won't.
"A bank will never outsource their security to a contractor in a foreign country"
UK banks already have.
"It's all going overseas, where it's a lot cheaper."
Only until computers make them redundant and this will happen eventually too.
"But don't forget traffic cops and the revenue that citations bring municipalities. And the body shops that repair damaged cars. And driving professions. It will be interesting how this impacts all these segments of the economy and what the ripples are."
There are at least 400 million jobs worldwide which will be directly affected by automated driving - that's just the ones at the wheel.
Bear in mind that this change has already happened once (working horses) we can see what will happen over time is that drivers will retire without new replacements coming in. The USA is already short of at _least_ 30,000 long-haul truckers so there's already sufficient economic motivation to press ahead with automation.
"Actually, I figure that it'll end up being insurance, more than anything else, that destroys the market for human driven cars once self-driving versions can do everything necessary except maybe off-roading."
That, and once computers can demonstrate adequate(*) driving skills, humans are going to face much tougher driving tests than are currently the norm.
(*) Most hoomans drive badly. Even good drivers have bad days and the level of most driving tests isn't much more than "You can point it in the right direction, not crash into anything else and don't bully on other drivers. Here's your license." - which is often the high point of many peoples' driving skills.
"People keep forgetting about the comprehensive part, what makes them think that the manufacturer is going to pay if a tree branch falls on it or a rock gets kicked up into the windshield? "
Claims of this nature are relatively rare. You're looking at a $30-50 premium covering it.
The vast majority of claims are down to driver error (which is effectively eliminated) or vandalism/burglary (which won't be particularly smart when these cars are bristling with sensors and can snap clear photos of the miscreants that are likely to be beamed back to base even if the car is stolen for parts - my own insurance company makes it clear that in the event of vandalism, if the culprit can be identified they'll go after him/her for 100% of all costs regardless of any criminal charges which might be added - and they've followed through on that (complete with discounts for dashcams or other monitoring systems which can detect/record damage whilst parked).
If a tree branch falls on your car, you might make a claim via your insurer but the actual liability and cost falls on the owner of the tree. Your insurance company is merely a facilitator of getting the claim resolved more quickly. Ditto in road subsidence issues.
"Most folks will probably still want comprehensive car insurance for other things"
'most folks' will probably stop owning a car once the cost overhead of paying a driver goes away. Even for the ones who do keep cars, 2-3 vehicle households will probably become single-vehicle ones as the single car performs multiple functions or fleet services are called in as needed.
Barcelona is estimating that the number of vehicles on its roads will drop by 80% once automation becomes common for this reason. I have no reason to doubt their reasoning on this issue (in many cities you don't leave personal stuff in your vehicle because it's likely to be broken into and items stolen. That's another disincentive to private ownership and a plus point for fleet services)
A lot of "don't swerve" has to do with the fact that it's easier to brake in a straight line.
ABS and ESC changes that dynamic a _little_ - they do allow swerving & etc without going sideways, but the primary problem is between the pedals and the seat. We simply don't react fast enough to deal with situations that occur at standard urban speeds (those speeds are too fast for our monkey-brains. Seriously)
Automated vehicles likely won't need to swerve or brake hard because they're communicating with each other but also because even when in standalone mode they'll be paying 100% attention 100% of the time and not following too close (the primary cause of almost all urban road crashes) or driving too fast for the conditions (the primary cause of most highway road crashes)
"Given a sidewalk where people can unexpectedly enter the road"
People _don't_ unexpectedly enter the road, short of erupting from a shielded location and even then there's usually some sort of warning (like a ball bouncing onto the road - a classic warning that it's time to slow down and prepare for a kid on the road) They telegraph their intentions pretty clearly even if not aware they're doing so.
The problem is car _drivers_ expect to have right of way and _expect_ that pedestrians won't step onto the road. The truth of the matter is that the curb creates a "safety demarcation line" which only exists in the mind of the driver.
A self driving car will not only stop more quickly, it will probably already be slowing down before the pedestrian has even stepped onto the road. Additionally, without an impatient monkey at the controls, more suitable speed limits (20mph in residential areas) will become the norm. The tradeoff of smoother traffic flow will more than likely eliminate any potential time loss on most journeys and pedestrian overpasses will probably become a thing of the past (they're a stupid idea. People don't use them and the main reason for them existing is because drivers tend to have tunnel vision above 30mph)
Machines have 360 degree vision AND they don't spend time looking at the cute ass on that female walking down the sidewalk instead of what's happening on the road in front of them, or looking at the pedestrian looking like he _might_ step onto the road on the right whilst missing the 18 wheeler pulling out on the left (or the wobbly cyclist who demands your entire attention to the point of missing the pedestrian stepping out from the other side of the road)
WRT the "google car hit a bus" non-story from a few months back, the car assumed (rightly) that it had right of way. The software didn't take into account that the bus driver would bully his way through regardless (If you look, you'll see the car stopped well before the collision). If both were automated it wouldn't have happened and newer algorithms take more account of human drivers refusing to yield to right of way.
FWIW the average traffic speed in London is 10-12mph and it's a lot slower in peak hours. Parking is nearly impossible (or extremely expensive). When you factor these in (and that driving in such conditions is stressful) there's a strong argument that people will flock to automated vehicles and car services. Now factor in that large european cities account for 30% of car sales in europe, that's a large chunk of lost sales for manufacturers - balanced by vastly increased sales in developing countries.
Even if the government revokes copyright (or declares eminent domain), they cannot force the company to release the source code.
Without that, support is limited to instructing people how to do particular tasks and once something crops out the software can't handle, it's game over.
MS-DOS was/is y2k compliant. It was windows 3(.1)(1) that wasn't.
People are confusing support contracts with use licenses.
Plenty of software has time-limited use rules (annual licensing) - that used to be the norm until MS turned up with "good enough" software that was cheap and one-time-payment for use.
Don't confuse that with a one-time payment for use and annual support contracts (which is what Oracle, and even Redhat use)
Limited liability shielding is for the SHAREHOLDERS.
NOT the directors.
Intentionally violating laws can (and does) put management on the hook.
Emphasis on "caught"
Without a full audit of their orders it's unknowable how many times they actually sent dangerous goods.
"Or does a 5MW nuclear reactor sound very very small. Like small enough that it is one of those experimental ones that produce medical isotopes."
Small is relative.
5MW is enough to drive a nuclear sub or power a small military base.
Civil reactors are in the 600-1400MW range (electrical, thermal is 3.5 to 4 times that)
The Oak Ridge Molten Salt experiment was 8MW (thermal)
Medical reactors only produce a few hundred watts of heat.