Where do you see anything about "the largest source of CO2"? Nowhere in the summary or article. They mention "one of the hardest to clean", and that's not surprising, since the burning of carbon is actually used as part of the chemical reaction in the traditional process, and 5% is a massive percentage for a single narrow industry.
As I recall the Bitcoin stuff is all based on projected trend lines, which are poor predictors when looking at infant technologies, but does serve as a cautionary note as to potential real cost of trying to scale bitcoin up to common currency levels.
Presumably, just as with the other intentional impurities such as molybdenum, manganese, chromium, or nickel. The amount of carbon used in the alloy is insignificant compared to the amount used to generate the needed heat.
I don't think they're talking about the intentional carbon impurities, but the fossil fuel currently used to generate the heat in the furnace. I rather doubt all that energy comes from bio-charcoal.
Certainly. However, such inaccuracies are completely intolerable for many modern scientific purposes, and THOSE are the reason heavily protected standardized references are created. Nobody cares if the grocery store scale is off by a percent or two from the the international standard kilogram, but when you're trying to determine the mass of an electron to six decimal places, it becomes extremely important.
Originally perhaps weight was referenced, but by your own quote that was removed only a few years later, and over 200 years ago.
As for trying to use the old mass-of-water definition to recreate the reference: what kind of water? After all, we now know that there are three stable, naturally occurring isotopes of hydrogen, and three of oxygen, all of which will be present in varying amounts in a sample of distilled water that size, meaning that individual water molecules can potentially vary in mass from approximately 18 amu (1H, 16O) to 24amu (3H, 18O). That's an awful wide range of density to span. Even presuming 99.8(ish)% of the water is composed of H1, O16, the variance in the rest would be unacceptably large for a modern reference mass. No two randomly selected samples will weigh exactly the same amount to the limits of modern measurement tools.
That's the main reason all the reference units have been redefined in terms of absolute constants - we've realized that the world is far less homogeneous and far more volatile than we once imagined, and it's basically impossible for any physical object to be recreated, or even maintained in stasis.
Thanks, I had been wondering the same thing, and the solution is obvious in retrospect.
Presumably the new definition will include the precise reference acceleration to be used. You'll never find such a reference acceleration to use of course, but it's easy enough to adjust the rest of the parameters to compensate for the actual acceleration available.
If the difference has been measured, then they know the direction of the disagreement. What they don't know is which reference is the one that drifted.
Absolutely it should be their job, for one simple reason: they're the only ones who have the information.
Now, maybe you can get a scribe to interview the patient beforehand, and record all the peripheral information beforehand - but if someone is recording information about the Doctor's finding's and recommendations, they need to actually *know* that information. Which means that either the doctor is entering the information themselves, or the doctor is dictating the information to a scribe. You tell me which is likely to be more expensive and error-prone.
Of course, for it to work well, you need a well-designed interface. Preferably one that lets you input key details immediately, and then keeps nagging you incessantly until you finish the job. Maybe you don't need to enter everything while talking to the patient - but you should probably download all the details from your brain before seeing the next patient and introducing retroactive memory loss.
Why is free speech in the very first amendment in the bill of rights? Because it is the most powerful, the most dangerous to government, and thus the most necessary to be protected. Guns are important in overthrowing tyranny, but without the speech to impassion people to fight and die over an ideal, they're not worth a damned thing.
Words hold awesome power to reshape the world - little else ever has. Those who claim free speech as an important bastion democracy, while belittling its power to cause harm, are guilty of blatant hypocrisy.
Free speech was never about saying anything you wanted without consequence - just about not letting the government silence you. In person, the threat of a punch in the face generally acts as sufficient deterrent against the most egregious breaches of common courtesy, online we clearly need something to take over that roll.
Not saying my idea is the solution, but it's something off the cuff that would be very simple to implement.
It seems likely that their are two major contributing factors to Braes paradox:
- First is congestion cost - too many driver choosing a "shortcut" increases traffic beyond the level that the shortcut (or its outlet) can handle, and flow is reduced. That's a problem with human drivers though - a fully automated roadway will see minimal congestion issues, because the cars can all be operated according to congestion-minimization parameters rather then selfish time-minimizing ones. The primary congestion will be at loading points, giving potential users plenty of warning to avoid the problem.
- The second is opportunity cost, and that's a lot less tractable. If the expected cost is low to get from A to B, then many people will make the journey that would otherwise avoid it and go elsewhere - as presumably happened in the bridge fire example on that page, where accidentally disabling a bridge ended up halving the total number of daily river crossings.
It's a non-trivial problem, but hardly intractable - competent traffic flow experts with enough data can make good predictions - as several examples on that page refer to.
Also having high karma here, I've got to say the whole karma thing is apparently set up more as a game than anything else. Could it be made workable if the goal was actually just to silence the really undesirables? I suspect so.
One option would be to raise the cost of starting a new account. That would to a large degree happen automatically on a social networking site, since unlike here where everyone posts in a single thread, social networking sites are all about networking - until you've built up a network, nobody is going to hear anything you say. Start a new account, you need to rebuild your network from scratch.
You could also do things like put all new accounts on probation as probable trolls, effectively putting them just shy of banning-level "karma" so that *any* flagging will get their comments sent to trial, and not many guilty verdicts will get them banned. Force people to behave themselves for a while before they can use their account for trolling.
Would it be enough? Too much? No idea. But the current free-for-all is clearly bringing out the worst in a lot of people, so maybe its worth trying.
There is a middle ground - courteous speech. If I'm at the bar and see someone I don't like, I can go give them a ration of shit, and maybe chase them off, or maybe get a punch in the nose for my trouble. Having immediate potential repercussions for discourteous speech is one of the biggest differences between real-life and online conversations. Perhaps we need to bring that online.
Not necessarily as a top-down "official rules of conduct" sort of situation, such a thing is far too vulnerable to being used for agenda-based censorship by the appointed authorities. But a bottom-up implementation to mimic in-person interactions might be possible.
One possible off-the-cuff implementation that could be easily implemented by any "members only" social network: One of your comments gets flagged as inappropriate by someone reading it, so it and the surrounding context get anonymized and shown to X randomly selected people with no connection to anyone involved (the jury), who then vote Guilty or Not guilty on its inappropriateness. If it gets voted guilty by the threshold margin (1/2? 2/3?) that comment gets hidden, and you get a black mark on your posting record. Accumulate enough black marks, and your posting freedom begins being restricted. If it gets voted Not Guilty by a similar threshold, then the flagger gets a black mark on their flagging record, and accumulating enough of those similarly restricts their flagging abilities.
Other considerations could also be thrown in to help reduce overhead - e.g. a post needs to accumulate a certain number of flags before being sent to trial. That could actually work well with recording flagging histories as well: the more comments that you flag inappropriate that are found not-guilty, the fewer "flagging points" your flag is worth, some flag-happy busybody might be only worth 0.1 point, while someone who only ever flags comments bad enough that they're always found guilty might be worth several.
Similarly, if you make a lot of posts that get flagged and found guilty, you might get a few flagging points automatically applied to every post, so that very little additional flagging is needed to send a post to trial. Keep it up and you might even reach the point where your every posts is presumed guilty and hidden until a trial determines otherwise. Or you might just get banned outright.
There will still be some censorship of unpopular people and opinions, but it will be democratically applied censorship, just as is seen in any real-life social setting, rather than some top-down system designed to enforce the agenda of those in power. And the greater the supermajority required to be found guilty, the less the level of censorship.
Absolutely. Now, pardon me I'm about to drive my car through your house, while gunning down anyone I see. After all, any interference in my freedom to do that would be a compromise, and thus prove that freedom does not exist.
In case you missed it, that was sarcasm. The entirety of civilization is built on the compromise of freedoms in exchange for safety, efficiency, etc. As the old saying goes - My freedom to swing my arms stops where your nose begins.
It's fine if the next person arrives before the queue is clear, so long as at least one person is served before they arrive the queue will continue to shorten. But yes, sometimes queues build up - that's why queues were invented instead of just random mobs of people standing around waiting. Helps minimize the worst-case wait times, and perhaps more importantly enforce a certain amount of fairness, as primates like us appear to be hardwired to rebel against perceived unfairness. (Also, LIFO and FIFO queues actually have exactly the same average waiting time, they simply trade off between minimizing best-case and worst-case waiting times. Either way you've still served X total people in Y total time, so Y/X is the same.)
I fail to see how you could end up in a Nash equilibrium state though - that implies nothing individual actors do unilaterally can improve their situation. Whereas if some people take the Loop, they improve the situation of non-Loopers, since traffic congestion slowdown generally increases faster than the traffic level, AND they improve the situation for themselves, unless the expected wait time is longer than the time they would save on the Loop *and* time is more important to them than the free time they gain as a passenger.
Plus, if the expected wait time is greater than the expected benefit, they are free to simply decide not to take the Loop today and drive normally instead - this is after all visualized as primarily an "underground car ferry" service. That alone should help optimize the usage patterns.
Ah, yes, that's an issue. But as other's have pointed out, that rail is almost certainly left over from constructing the tunnel, and would NOT be used for high-speed trains.
After all, this is a test tunnel, for testing their refinements of the tunnel-boring machines and processes. NOT a prototype high-speed rail system. In fact, it's not even a sure thing they'd use rail for their proposed eventual Loop system at all - traveling at those speeds on pavement is no big deal. At least not for cars. Which brings up the other big difference - high speed trains are *trains* - you're moving huge amounts of mass all at once, which means the forces involved to guide it at a given speed are similarly huge, and the tolerances far tighter. If instead you're only looking to move individual flatbeds and minibuses things are a lot simpler.
Where are you getting 90 seconds from? From what I can find subway trains generally run every 2-10 minutes during rush hours, or 120-600 seconds between trains. Your basic passenger throughput comparison still holds, just not quite as dramatically.
Construction costs though I must disagree with - the primary purpose of The Boring Company is to revolutionize tunnel-boring technology, which has pretty much stagnated in terms of cost and speed for many, many decades. I believe they're targetting an eventual order of magnitude reduction in time and cost to build a tunnel. At which point yes, absolutely, forget the Loop nonsense, they would make great sense for tiny 12' subway tunnels as well. My sense is that the Loop system plan is still under-baked. If you forgot the outrageous speeds, and just figured your automated skate would take you wherever you wanted to go at highway speeds with bumper-to-bumper traffic but no congestion, then maybe it makes sense - highways have a pretty respectable throughput when running well.
One of the big issues with ferry and train queues is that they're large-batch endeavours. A bunch of people queue up as the loading time approaches, then the ferry arrives and has to wait for all the vehicles to be loaded (not helped by the fact that passengers generally don't ride in their car), and then they all get transported at once.
If instead you have a single vehicle arrives at any time and immediately gets loaded onto a 1-car ferry(skate) and departs, as is planned for Loop, then the queue depth is determined by the average frequency at which new cars arrive versus the time to load them on a 1-car ferry. If the load time is less than the average time it takes for the next vehicle to arrive, then you only get a queue when cars are arriving at above-average speeds. Which probably includes rush-hours, but if you're planning to take the Loop and don't want to deal with rush-hour queues you can likely adjust your plans to avoid the most congested times. After all, unlike with a train there's no waiting time except in the queue (or at least much less, you may have to wait for the next available skate to arrive, but not for a ore-scheduled departure time).
Tunnels certainly have limitations, but the biggest one is cost. And that is EXACTLY what The Boring Company is focused on - reducing the cost and construction time of tunnels. Tunnels currently cost ~10x what a comparable elevated road would cost to build - if they can make the construction costs roughly equivalent, as they're hoping, then you can get all the benefits of an elevated road, without any of the the eyesore or right-of-way issues, and much greater earthquake safety.
After all, they're not replacing roads, they're complementing them. Got a road that's heavily congested by through traffic? Replace it with a tunnel with very few on-off ramps, so that the surface streets can be used primarily by local traffic, and the underground throughfare doesn't get congested with cross-traffic.
Then there's all the Loop automated transportation network stuff, but that's a secondary / marketing goal.
Here's the thing, tunnels are round, and it doesn't much matter if you put the tracks (road bed, etc) on the floor or at 90* on the wall, so long as you don't have inexperienced human drivers trying to navigate them, and you can rely on traffic traveling at an appropriate speed for the amount of banking. Think roller coasters, which can be designed to give a very smooth ride at high speeds through hairpin turns (but usually aren't, since getting thrown around is part of the fun of the coaster)
Right. The Boring Company isn't focussed on making transportation systems, innovative or otherwise - they're focussed on digging tunnels, which Musk believes can be done at least an order of magnitude faster and cheaper than it currently is.
Musk's dream, once cheap tunnels are available, is to build Loop transportation networks with them - but that's a long-term goal.
The plan for the tunnels is Loop, NOT Hyperloop - the physics are completely different when you're talking autonomous flatbeds and mini-busses, and no vacuum.
Depending on how cheaply the tunnels can be dug, they may eventually be well suited for Hyperloop partial-vacuum tunnels, but that's not part of the current plan. Hyperloop (arguably) makes sense for medium-to-long-range transportation(several 10s to 100s of miles), it definitely *doesn't* make sense for short-range transportation, which is what the Loop is trying to address.
Actually, yes, Musk *has* addressed those issues, and the solutions are fairly easy - eliminate the intersections and minimize stops.
Your tunnels are after all located in three-dimensional space - unlike on the mostly 2D ground surface there is negligible added cost for going higher or lower. So if you want a high-throughput "intersection" you do it the same way as for highways - an overpass with transfer ramps. It adds the cost of some extra tunnels length for the ramps, but that's about it.
And eliminating the stops is easy with small automated skates, since unlike a train you have relatively few passengers per vehicle (which are either a small "bus" or single-car "ferry") and you only need to stop when loading or unloading. And if you have lots of skates simultaneously serving the same basic route it's relatively simple to optimize the passenger loadout - maybe you wait an extra couple minutes for the "right" skate to arrive as others pass by, but you make up the time in high speeds and making very few stops.
Depends on the tunnel. The idea Musk is selling is a network of tunnels with frequent access points, and no human control within. You drive your car to an access point, it gets loaded on an electric flatbed "skate", and unloads you at your destination, without ever facing any of the congestion created by people driving exactly wrong way for promoting throughput.
Even with relatively uncommon access points, such a system could accelerate traffic throughput dramatically - put an access point every few miles and you can get pretty much anywhere the tunnels service while driving an average of only a couple miles, while reducing the burden of through traffic on the streets in between. And when you reduce the congestion on streets, traffic flows disproportionately better.
Granted, the initial tunnels probably won't deliver on that promise, any more than the early Falcon 9s delivered on the promise of reusable rockets. But you have to start somewhere.
Slashdot Mirror
Where do you see anything about "the largest source of CO2"? Nowhere in the summary or article. They mention "one of the hardest to clean", and that's not surprising, since the burning of carbon is actually used as part of the chemical reaction in the traditional process, and 5% is a massive percentage for a single narrow industry.
As I recall the Bitcoin stuff is all based on projected trend lines, which are poor predictors when looking at infant technologies, but does serve as a cautionary note as to potential real cost of trying to scale bitcoin up to common currency levels.
Presumably, just as with the other intentional impurities such as molybdenum, manganese, chromium, or nickel. The amount of carbon used in the alloy is insignificant compared to the amount used to generate the needed heat.
I don't think they're talking about the intentional carbon impurities, but the fossil fuel currently used to generate the heat in the furnace. I rather doubt all that energy comes from bio-charcoal.
Certainly. However, such inaccuracies are completely intolerable for many modern scientific purposes, and THOSE are the reason heavily protected standardized references are created. Nobody cares if the grocery store scale is off by a percent or two from the the international standard kilogram, but when you're trying to determine the mass of an electron to six decimal places, it becomes extremely important.
Originally perhaps weight was referenced, but by your own quote that was removed only a few years later, and over 200 years ago.
As for trying to use the old mass-of-water definition to recreate the reference: what kind of water? After all, we now know that there are three stable, naturally occurring isotopes of hydrogen, and three of oxygen, all of which will be present in varying amounts in a sample of distilled water that size, meaning that individual water molecules can potentially vary in mass from approximately 18 amu (1H, 16O) to 24amu (3H, 18O). That's an awful wide range of density to span. Even presuming 99.8(ish)% of the water is composed of H1, O16, the variance in the rest would be unacceptably large for a modern reference mass. No two randomly selected samples will weigh exactly the same amount to the limits of modern measurement tools.
That's the main reason all the reference units have been redefined in terms of absolute constants - we've realized that the world is far less homogeneous and far more volatile than we once imagined, and it's basically impossible for any physical object to be recreated, or even maintained in stasis.
Thanks, I had been wondering the same thing, and the solution is obvious in retrospect.
Presumably the new definition will include the precise reference acceleration to be used. You'll never find such a reference acceleration to use of course, but it's easy enough to adjust the rest of the parameters to compensate for the actual acceleration available.
If the difference has been measured, then they know the direction of the disagreement. What they don't know is which reference is the one that drifted.
Absolutely it should be their job, for one simple reason: they're the only ones who have the information.
Now, maybe you can get a scribe to interview the patient beforehand, and record all the peripheral information beforehand - but if someone is recording information about the Doctor's finding's and recommendations, they need to actually *know* that information. Which means that either the doctor is entering the information themselves, or the doctor is dictating the information to a scribe. You tell me which is likely to be more expensive and error-prone.
Of course, for it to work well, you need a well-designed interface. Preferably one that lets you input key details immediately, and then keeps nagging you incessantly until you finish the job. Maybe you don't need to enter everything while talking to the patient - but you should probably download all the details from your brain before seeing the next patient and introducing retroactive memory loss.
Why is free speech in the very first amendment in the bill of rights? Because it is the most powerful, the most dangerous to government, and thus the most necessary to be protected. Guns are important in overthrowing tyranny, but without the speech to impassion people to fight and die over an ideal, they're not worth a damned thing.
Words hold awesome power to reshape the world - little else ever has. Those who claim free speech as an important bastion democracy, while belittling its power to cause harm, are guilty of blatant hypocrisy.
Free speech was never about saying anything you wanted without consequence - just about not letting the government silence you. In person, the threat of a punch in the face generally acts as sufficient deterrent against the most egregious breaches of common courtesy, online we clearly need something to take over that roll.
Not saying my idea is the solution, but it's something off the cuff that would be very simple to implement.
It seems likely that their are two major contributing factors to Braes paradox:
- First is congestion cost - too many driver choosing a "shortcut" increases traffic beyond the level that the shortcut (or its outlet) can handle, and flow is reduced. That's a problem with human drivers though - a fully automated roadway will see minimal congestion issues, because the cars can all be operated according to congestion-minimization parameters rather then selfish time-minimizing ones. The primary congestion will be at loading points, giving potential users plenty of warning to avoid the problem.
- The second is opportunity cost, and that's a lot less tractable. If the expected cost is low to get from A to B, then many people will make the journey that would otherwise avoid it and go elsewhere - as presumably happened in the bridge fire example on that page, where accidentally disabling a bridge ended up halving the total number of daily river crossings.
It's a non-trivial problem, but hardly intractable - competent traffic flow experts with enough data can make good predictions - as several examples on that page refer to.
Also having high karma here, I've got to say the whole karma thing is apparently set up more as a game than anything else. Could it be made workable if the goal was actually just to silence the really undesirables? I suspect so.
One option would be to raise the cost of starting a new account. That would to a large degree happen automatically on a social networking site, since unlike here where everyone posts in a single thread, social networking sites are all about networking - until you've built up a network, nobody is going to hear anything you say. Start a new account, you need to rebuild your network from scratch.
You could also do things like put all new accounts on probation as probable trolls, effectively putting them just shy of banning-level "karma" so that *any* flagging will get their comments sent to trial, and not many guilty verdicts will get them banned. Force people to behave themselves for a while before they can use their account for trolling.
Would it be enough? Too much? No idea. But the current free-for-all is clearly bringing out the worst in a lot of people, so maybe its worth trying.
There is a middle ground - courteous speech. If I'm at the bar and see someone I don't like, I can go give them a ration of shit, and maybe chase them off, or maybe get a punch in the nose for my trouble. Having immediate potential repercussions for discourteous speech is one of the biggest differences between real-life and online conversations. Perhaps we need to bring that online.
Not necessarily as a top-down "official rules of conduct" sort of situation, such a thing is far too vulnerable to being used for agenda-based censorship by the appointed authorities. But a bottom-up implementation to mimic in-person interactions might be possible.
One possible off-the-cuff implementation that could be easily implemented by any "members only" social network:
One of your comments gets flagged as inappropriate by someone reading it, so it and the surrounding context get anonymized and shown to X randomly selected people with no connection to anyone involved (the jury), who then vote Guilty or Not guilty on its inappropriateness. If it gets voted guilty by the threshold margin (1/2? 2/3?) that comment gets hidden, and you get a black mark on your posting record. Accumulate enough black marks, and your posting freedom begins being restricted. If it gets voted Not Guilty by a similar threshold, then the flagger gets a black mark on their flagging record, and accumulating enough of those similarly restricts their flagging abilities.
Other considerations could also be thrown in to help reduce overhead - e.g. a post needs to accumulate a certain number of flags before being sent to trial. That could actually work well with recording flagging histories as well: the more comments that you flag inappropriate that are found not-guilty, the fewer "flagging points" your flag is worth, some flag-happy busybody might be only worth 0.1 point, while someone who only ever flags comments bad enough that they're always found guilty might be worth several.
Similarly, if you make a lot of posts that get flagged and found guilty, you might get a few flagging points automatically applied to every post, so that very little additional flagging is needed to send a post to trial. Keep it up and you might even reach the point where your every posts is presumed guilty and hidden until a trial determines otherwise. Or you might just get banned outright.
There will still be some censorship of unpopular people and opinions, but it will be democratically applied censorship, just as is seen in any real-life social setting, rather than some top-down system designed to enforce the agenda of those in power. And the greater the supermajority required to be found guilty, the less the level of censorship.
Absolutely. Now, pardon me I'm about to drive my car through your house, while gunning down anyone I see. After all, any interference in my freedom to do that would be a compromise, and thus prove that freedom does not exist.
In case you missed it, that was sarcasm. The entirety of civilization is built on the compromise of freedoms in exchange for safety, efficiency, etc. As the old saying goes - My freedom to swing my arms stops where your nose begins.
It's fine if the next person arrives before the queue is clear, so long as at least one person is served before they arrive the queue will continue to shorten. But yes, sometimes queues build up - that's why queues were invented instead of just random mobs of people standing around waiting. Helps minimize the worst-case wait times, and perhaps more importantly enforce a certain amount of fairness, as primates like us appear to be hardwired to rebel against perceived unfairness. (Also, LIFO and FIFO queues actually have exactly the same average waiting time, they simply trade off between minimizing best-case and worst-case waiting times. Either way you've still served X total people in Y total time, so Y/X is the same.)
I fail to see how you could end up in a Nash equilibrium state though - that implies nothing individual actors do unilaterally can improve their situation. Whereas if some people take the Loop, they improve the situation of non-Loopers, since traffic congestion slowdown generally increases faster than the traffic level, AND they improve the situation for themselves, unless the expected wait time is longer than the time they would save on the Loop *and* time is more important to them than the free time they gain as a passenger.
Plus, if the expected wait time is greater than the expected benefit, they are free to simply decide not to take the Loop today and drive normally instead - this is after all visualized as primarily an "underground car ferry" service. That alone should help optimize the usage patterns.
Generally speaking using the best-case scenario in any comparison is useless.
Ah, yes, that's an issue. But as other's have pointed out, that rail is almost certainly left over from constructing the tunnel, and would NOT be used for high-speed trains.
After all, this is a test tunnel, for testing their refinements of the tunnel-boring machines and processes. NOT a prototype high-speed rail system. In fact, it's not even a sure thing they'd use rail for their proposed eventual Loop system at all - traveling at those speeds on pavement is no big deal. At least not for cars. Which brings up the other big difference - high speed trains are *trains* - you're moving huge amounts of mass all at once, which means the forces involved to guide it at a given speed are similarly huge, and the tolerances far tighter. If instead you're only looking to move individual flatbeds and minibuses things are a lot simpler.
Where are you getting 90 seconds from? From what I can find subway trains generally run every 2-10 minutes during rush hours, or 120-600 seconds between trains. Your basic passenger throughput comparison still holds, just not quite as dramatically.
Construction costs though I must disagree with - the primary purpose of The Boring Company is to revolutionize tunnel-boring technology, which has pretty much stagnated in terms of cost and speed for many, many decades. I believe they're targetting an eventual order of magnitude reduction in time and cost to build a tunnel. At which point yes, absolutely, forget the Loop nonsense, they would make great sense for tiny 12' subway tunnels as well. My sense is that the Loop system plan is still under-baked. If you forgot the outrageous speeds, and just figured your automated skate would take you wherever you wanted to go at highway speeds with bumper-to-bumper traffic but no congestion, then maybe it makes sense - highways have a pretty respectable throughput when running well.
One of the big issues with ferry and train queues is that they're large-batch endeavours. A bunch of people queue up as the loading time approaches, then the ferry arrives and has to wait for all the vehicles to be loaded (not helped by the fact that passengers generally don't ride in their car), and then they all get transported at once.
If instead you have a single vehicle arrives at any time and immediately gets loaded onto a 1-car ferry(skate) and departs, as is planned for Loop, then the queue depth is determined by the average frequency at which new cars arrive versus the time to load them on a 1-car ferry. If the load time is less than the average time it takes for the next vehicle to arrive, then you only get a queue when cars are arriving at above-average speeds. Which probably includes rush-hours, but if you're planning to take the Loop and don't want to deal with rush-hour queues you can likely adjust your plans to avoid the most congested times. After all, unlike with a train there's no waiting time except in the queue (or at least much less, you may have to wait for the next available skate to arrive, but not for a ore-scheduled departure time).
Tunnels certainly have limitations, but the biggest one is cost. And that is EXACTLY what The Boring Company is focused on - reducing the cost and construction time of tunnels. Tunnels currently cost ~10x what a comparable elevated road would cost to build - if they can make the construction costs roughly equivalent, as they're hoping, then you can get all the benefits of an elevated road, without any of the the eyesore or right-of-way issues, and much greater earthquake safety.
After all, they're not replacing roads, they're complementing them. Got a road that's heavily congested by through traffic? Replace it with a tunnel with very few on-off ramps, so that the surface streets can be used primarily by local traffic, and the underground throughfare doesn't get congested with cross-traffic.
Then there's all the Loop automated transportation network stuff, but that's a secondary / marketing goal.
Two words: Banked curve.
Here's the thing, tunnels are round, and it doesn't much matter if you put the tracks (road bed, etc) on the floor or at 90* on the wall, so long as you don't have inexperienced human drivers trying to navigate them, and you can rely on traffic traveling at an appropriate speed for the amount of banking. Think roller coasters, which can be designed to give a very smooth ride at high speeds through hairpin turns (but usually aren't, since getting thrown around is part of the fun of the coaster)
Right. The Boring Company isn't focussed on making transportation systems, innovative or otherwise - they're focussed on digging tunnels, which Musk believes can be done at least an order of magnitude faster and cheaper than it currently is.
Musk's dream, once cheap tunnels are available, is to build Loop transportation networks with them - but that's a long-term goal.
The plan for the tunnels is Loop, NOT Hyperloop - the physics are completely different when you're talking autonomous flatbeds and mini-busses, and no vacuum.
Depending on how cheaply the tunnels can be dug, they may eventually be well suited for Hyperloop partial-vacuum tunnels, but that's not part of the current plan. Hyperloop (arguably) makes sense for medium-to-long-range transportation(several 10s to 100s of miles), it definitely *doesn't* make sense for short-range transportation, which is what the Loop is trying to address.
Actually, yes, Musk *has* addressed those issues, and the solutions are fairly easy - eliminate the intersections and minimize stops.
Your tunnels are after all located in three-dimensional space - unlike on the mostly 2D ground surface there is negligible added cost for going higher or lower. So if you want a high-throughput "intersection" you do it the same way as for highways - an overpass with transfer ramps. It adds the cost of some extra tunnels length for the ramps, but that's about it.
And eliminating the stops is easy with small automated skates, since unlike a train you have relatively few passengers per vehicle (which are either a small "bus" or single-car "ferry") and you only need to stop when loading or unloading. And if you have lots of skates simultaneously serving the same basic route it's relatively simple to optimize the passenger loadout - maybe you wait an extra couple minutes for the "right" skate to arrive as others pass by, but you make up the time in high speeds and making very few stops.
Depends on the tunnel. The idea Musk is selling is a network of tunnels with frequent access points, and no human control within. You drive your car to an access point, it gets loaded on an electric flatbed "skate", and unloads you at your destination, without ever facing any of the congestion created by people driving exactly wrong way for promoting throughput.
Even with relatively uncommon access points, such a system could accelerate traffic throughput dramatically - put an access point every few miles and you can get pretty much anywhere the tunnels service while driving an average of only a couple miles, while reducing the burden of through traffic on the streets in between. And when you reduce the congestion on streets, traffic flows disproportionately better.
Granted, the initial tunnels probably won't deliver on that promise, any more than the early Falcon 9s delivered on the promise of reusable rockets. But you have to start somewhere.