Some of them are sort of valid, but not relevant in practice. For example, it's true that Tesla's current service infrastructure can't handle high demands... but that's because the infrastructure is sized for the current customer base. Building a service infrastructure that can handle many more customers than you actually have is a waste of money, and it's completely unnecessary so long as you continue to scale that infrastructure as you grow.
The rates aren't the same, though. It may change in the future (I believe the stagnation of HDDs was due to lack of sufficient demand for larger drives), but for the past several years, SSD densities have been increasing far faster than HDD densities.
The Tesla car already has a bunch of shielding (both as a casing and to protect against impact damage) and cooling (active water cooling, specifically). Considering that most RTGs are designed to operate in a vacuum (where you have no medium to carry away heat) in zero gravity (where convection doesn't work), I have no idea what sort of cooling and shielding you'd require for terrestrial RTGs.
1. I realize that they're currently at 5%, the whole point of my scenario was examining what sort of changes a large increase in efficiency would produce... that's the whole point of the article, after all. Efficiency would need to be somewhere around 50% to justify replacing ICEs with thermoelectric engines. Is that possible? I've got no idea, TFA gives zero layman-friendly information about what sort of efficiency improvements are foreseen.
2. Supply isn't as big a problem as the incredible safety issues. I acknowledge in my post that the idea is totally insane, which is why I doubt that, even with a big improvement in efficiency, you'd probably never see RTGs used outside of military applications.
3. That's not necessarily a problem. They conveniently provide power that can be used for active cooling. Cooling them in a vacuum is an issue (hence the giant heat dissipation fins), cooling them in an atmosphere isn't as much of an issue.
I suspect that sufficiently efficient thermoelectrics might find their way into military UAVs, which could remain airborn for extended periods of time, for example. Or as an alternative to shipping diesel to remote outposts (although they're currently looking into robotic trucks to solve that problem).
It's not as silly as you might think. I believe you get roughly 500W of heat per kilo of plutonium-238. A Tesla Model S driving at normal speeds consumes something like 15KW. If you could get 50% efficiency for your thermoelectrics, you could build an RTG-powered model S with 60 kilos of plutonium. You'd need capacitors for surge demand, obviously.
Of course, this would be completely insane, but I don't see why it's not theoretically possible, since the battery pack on the car that you'd be replacing already weighs something like 600 kilos.
Then again, with sufficiently efficient thermoelectrics, you might see the military using RTGs.
While the marketing certainly played a big part in that, a bigger factor was likely that technology was finally ready. From the article, Nokia's tablet had a 4 hour battery life and weight more than four pounds... the bezel had 2.5x the surface area of the screen itself!
The thing wasn't a tablet, it was a portable Internet kiosk.
I'm not talking about enhanced field of view by turning around (although that's definitely a bonus), but about the subtle sensations that enhance presence by having the small movements of your head reflected in terms of parallax changes and such.
3D displays work poorly for the use case that you describe, because they all assume that your head is perfectly still, facing the monitor, dead-centre. It doesn't account for any movement or different position of your head whatsoever(so it probably doesn't work for anything but the central monitor in front of you).
You will not get a sense of presence sitting in front of a bunch of monitors, even if they surrounded you in a circle. This is something that's difficult to describe to someone who has never tried modern VR headsets, because if you'd tried them, you would understand the huge difference between them. The feeling of feeling like you're actually in a game world, rather than looking at it through windows. Of standing next to a virtual character and having the same sort of sense of the person being there, being a certain size, in relation to yourself. A multi-monitor setup doesn't provide any sense of presence at all.
Yes, varying the distance between the lens and the screen does the exact same thing on a display-based headset. The Oculus Rift sort of supports this today, by providing different lens cups. All the lenses that come with the Rift (the A/B/C lenses) are actually identical, the only difference is their plastic casing that varies their distance from the screen. It's possible that the consumer version will allow this to be fully adjustable rather than in discrete steps.
Having experienced both VR headsets (with 90-110 degree FOVs) and the surround-yourself-by-lots-of-2D-monitors approach, throwing LCD monitors at the problem doesn't hold a candle to the immersion/presence the VR headset gets. There's more to experiencing presence than a big horizontal FoV. A VR headset also gets you the horizontal FOV, gets rid of gaps between monitors, blocks out stuff outside the monitors, provides you with stereoscopy, the head tracking gives you the possibility of parallax, etc.
Trading the screen-door effect for the rainbow effect isn't a good trade-off. The screen-door effect can be solved by increasing display resolution and tweaking sub-pixel geometry (DK2 is a diamond matrix pentile-like display that reduces the screen door effect), while a single-chip DLP solution can't do much to improve on the rainbow effect short of cranking up switching speeds (which are already in the thousands or tens of thousands of hertz).
For what it's trying to do, which is to simulate big screen without having to carry around a big screen, the Glyph can put up with some rainbow effect. For VR, it's a death sentence.
The stuff at affordable prices (still double what VR headsets will go for when the Rift or Morpheus launch) 10+ years ago was high-latency, low-detail/resolution, low-precision, bulky, with a tiny depth of field. The units that solved some or most of those problems cost tens or hundreds of thousands of dollars.
The two biggest things today are the much higher amount of compute performance available, as well as the existence of modern smartphone displays (small, high-resolution, low-latency), which didn't exist ten years ago.
There is no real difference between using DLP to shine light in your eyes versus looking at an LED screen. The Avegant Glyph (and the LED/DLP setup you describe) is not a virtual retina display, and it doesn't "paint an image directly onto your retina".
Any advantages that it gains from increased pixel fill (and the three subpixels overlapping) are undone by the massive issues they have with the rainbow effect, since they have to stagger the red/green/blue images in time instead of in space. Solving that requires much higher update rates from the digital micro mirror device than we have today, and those things are already above ten thousand hertz...
There are no 4K OLED displays of the appropriate size in existence, let alone ready for integration into a product (Palmer's said this on Reddit to boot), so it's likely that CV1 will be 2.5K.
Huh? Most mobile phones use OLED displays, most upcoming VR headsets (including the Oculus Rift DK2 and consumer version and Sony's Project Morpheus) use or will be using OLED displays, and you can now buy OLED TVs at various sizes (cost still hasn't come down to consumer levels yet).
Did you neglect the part about how the temperature must not exceed 22 degrees, excepting when the air is dry, at which point 23 degrees is acceptable? You're trying to spin that as a simple "care and feeding", when in reality it goes far beyond that. He gets incredibly precise and specific on a huge number of points, showing little flexibility in some respects, and a bizarrely rigid flexibility in others, if that makes sense. I did take that specific quote without the context, but including the context doesn't do anything to change how incredibly specific the document gets.
My implication was that if he's so difficult to accommodate for speaking arrangements, it's entirely possible that he has some extremely specific requirements for remote interviews such as this. Something along the lines of "I won't answer questions sent to me via a non-free e-mail client, lest it make acceptable the use of the very thing we're trying to change" or some such.
And before you say I'm exaggerating, the rider for speaking arrangements does specifically say you can't use non-free software to record or stream his speeches, despite the act of recording and streaming being encouraged. He really does take this sort of stuff to the logical extreme.
The most efficient diesel engines in cars can hit 50% efficiency, while you're saying that a fuel cell car can hit about 76%. That's a nice improvement, but not enough to revolutionize anything.
I've heard about such things being a possibility, but there are two issues. First, they don't exist yet (batteries exist today, gasoline-powered fuel cells in cars are many years from production). Second, they don't solve most of the problems with gasoline (cost, dependence on foreign oil).
Have you SEEN his rider agreement? Sure, that's for speaking, rather than remote interviews, but the rider for giving a simple speech includes such gems as "If you buy a captured wild parrot, you will promote a cruel and devastating practice, and the parrot will be emotionally scarred before you get it."
Fuel cells have their own set of problems. There's no distribution infrastructure for hydrogen, while there is distribution infrastructure for electricity. It also has efficiency issues, since producing hydrogen isn't all that energy efficient.
Fuel cells may be practical in the long term, but batteries are practical today.
Power cables typically use stranded conductors specifically to avoid this problem. Drop your time requirement to five minutes (half a megawatt) and use the kind of power conductors that normal people use and you've got something practical.
Current charge stations (with their existing cabling) are expected to be able to do up to 150 kW. Worst case, you use two charge cables per car, and going from 150 kW to 250 kW is suddenly not such a big leap.
Tesla's plan is to have large amounts of grid storage on-site, powered mostly by solar (by building roofs over the charge stations). Tesla claims that they should be a net-positive in terms of grid power (that they produce more power than they consume). I'm skeptical that would work once they get popular, but it does still offset a chunk of the power draw, and the grid storage on-site smooths out the surges in demand.
Even with a queue, an individual gas hose isn't in use 100% of the time. There's the time you take to pull up, get out of your car, connect the hose, disconnect the hose, pay, and drive off. Even if capacitors supported a 1 minute charge every 3 minutes, that'd probably be enough.
Personally, I think a 30 second or 1 minute target is unnecessary. It takes longer than that to refuel a regular car anyhow. Five or ten minutes is probably fine, and charging 85 kWh in 10 minutes can be done with ~500 kW. That's high, but a heck of a lot less insane than the multi-megawatt charging people are talking about here. Current high-speed chargers are hitting 120 to 135 kW today.
Slashdot Mirror
Some of them are sort of valid, but not relevant in practice. For example, it's true that Tesla's current service infrastructure can't handle high demands... but that's because the infrastructure is sized for the current customer base. Building a service infrastructure that can handle many more customers than you actually have is a waste of money, and it's completely unnecessary so long as you continue to scale that infrastructure as you grow.
The rates aren't the same, though. It may change in the future (I believe the stagnation of HDDs was due to lack of sufficient demand for larger drives), but for the past several years, SSD densities have been increasing far faster than HDD densities.
The Tesla car already has a bunch of shielding (both as a casing and to protect against impact damage) and cooling (active water cooling, specifically). Considering that most RTGs are designed to operate in a vacuum (where you have no medium to carry away heat) in zero gravity (where convection doesn't work), I have no idea what sort of cooling and shielding you'd require for terrestrial RTGs.
This is different to throwing explosive ordinance at current vehicles powered by nuclear reactors how?
1. I realize that they're currently at 5%, the whole point of my scenario was examining what sort of changes a large increase in efficiency would produce... that's the whole point of the article, after all. Efficiency would need to be somewhere around 50% to justify replacing ICEs with thermoelectric engines. Is that possible? I've got no idea, TFA gives zero layman-friendly information about what sort of efficiency improvements are foreseen.
2. Supply isn't as big a problem as the incredible safety issues. I acknowledge in my post that the idea is totally insane, which is why I doubt that, even with a big improvement in efficiency, you'd probably never see RTGs used outside of military applications.
3. That's not necessarily a problem. They conveniently provide power that can be used for active cooling. Cooling them in a vacuum is an issue (hence the giant heat dissipation fins), cooling them in an atmosphere isn't as much of an issue.
I suspect that sufficiently efficient thermoelectrics might find their way into military UAVs, which could remain airborn for extended periods of time, for example. Or as an alternative to shipping diesel to remote outposts (although they're currently looking into robotic trucks to solve that problem).
It's not as silly as you might think. I believe you get roughly 500W of heat per kilo of plutonium-238. A Tesla Model S driving at normal speeds consumes something like 15KW. If you could get 50% efficiency for your thermoelectrics, you could build an RTG-powered model S with 60 kilos of plutonium. You'd need capacitors for surge demand, obviously.
Of course, this would be completely insane, but I don't see why it's not theoretically possible, since the battery pack on the car that you'd be replacing already weighs something like 600 kilos.
Then again, with sufficiently efficient thermoelectrics, you might see the military using RTGs.
While the marketing certainly played a big part in that, a bigger factor was likely that technology was finally ready. From the article, Nokia's tablet had a 4 hour battery life and weight more than four pounds... the bezel had 2.5x the surface area of the screen itself!
The thing wasn't a tablet, it was a portable Internet kiosk.
I'm not talking about enhanced field of view by turning around (although that's definitely a bonus), but about the subtle sensations that enhance presence by having the small movements of your head reflected in terms of parallax changes and such.
3D displays work poorly for the use case that you describe, because they all assume that your head is perfectly still, facing the monitor, dead-centre. It doesn't account for any movement or different position of your head whatsoever(so it probably doesn't work for anything but the central monitor in front of you).
You will not get a sense of presence sitting in front of a bunch of monitors, even if they surrounded you in a circle. This is something that's difficult to describe to someone who has never tried modern VR headsets, because if you'd tried them, you would understand the huge difference between them. The feeling of feeling like you're actually in a game world, rather than looking at it through windows. Of standing next to a virtual character and having the same sort of sense of the person being there, being a certain size, in relation to yourself. A multi-monitor setup doesn't provide any sense of presence at all.
Yes, varying the distance between the lens and the screen does the exact same thing on a display-based headset. The Oculus Rift sort of supports this today, by providing different lens cups. All the lenses that come with the Rift (the A/B/C lenses) are actually identical, the only difference is their plastic casing that varies their distance from the screen. It's possible that the consumer version will allow this to be fully adjustable rather than in discrete steps.
Having experienced both VR headsets (with 90-110 degree FOVs) and the surround-yourself-by-lots-of-2D-monitors approach, throwing LCD monitors at the problem doesn't hold a candle to the immersion/presence the VR headset gets. There's more to experiencing presence than a big horizontal FoV. A VR headset also gets you the horizontal FOV, gets rid of gaps between monitors, blocks out stuff outside the monitors, provides you with stereoscopy, the head tracking gives you the possibility of parallax, etc.
Trading the screen-door effect for the rainbow effect isn't a good trade-off. The screen-door effect can be solved by increasing display resolution and tweaking sub-pixel geometry (DK2 is a diamond matrix pentile-like display that reduces the screen door effect), while a single-chip DLP solution can't do much to improve on the rainbow effect short of cranking up switching speeds (which are already in the thousands or tens of thousands of hertz).
For what it's trying to do, which is to simulate big screen without having to carry around a big screen, the Glyph can put up with some rainbow effect. For VR, it's a death sentence.
The stuff at affordable prices (still double what VR headsets will go for when the Rift or Morpheus launch) 10+ years ago was high-latency, low-detail/resolution, low-precision, bulky, with a tiny depth of field. The units that solved some or most of those problems cost tens or hundreds of thousands of dollars.
The two biggest things today are the much higher amount of compute performance available, as well as the existence of modern smartphone displays (small, high-resolution, low-latency), which didn't exist ten years ago.
There is no real difference between using DLP to shine light in your eyes versus looking at an LED screen. The Avegant Glyph (and the LED/DLP setup you describe) is not a virtual retina display, and it doesn't "paint an image directly onto your retina".
Any advantages that it gains from increased pixel fill (and the three subpixels overlapping) are undone by the massive issues they have with the rainbow effect, since they have to stagger the red/green/blue images in time instead of in space. Solving that requires much higher update rates from the digital micro mirror device than we have today, and those things are already above ten thousand hertz...
There are no 4K OLED displays of the appropriate size in existence, let alone ready for integration into a product (Palmer's said this on Reddit to boot), so it's likely that CV1 will be 2.5K.
Huh? Most mobile phones use OLED displays, most upcoming VR headsets (including the Oculus Rift DK2 and consumer version and Sony's Project Morpheus) use or will be using OLED displays, and you can now buy OLED TVs at various sizes (cost still hasn't come down to consumer levels yet).
Simpler to just use a monochrome LCD with a large single pixel (same idea as active 3D glasses) as your backing.
Did you neglect the part about how the temperature must not exceed 22 degrees, excepting when the air is dry, at which point 23 degrees is acceptable? You're trying to spin that as a simple "care and feeding", when in reality it goes far beyond that. He gets incredibly precise and specific on a huge number of points, showing little flexibility in some respects, and a bizarrely rigid flexibility in others, if that makes sense. I did take that specific quote without the context, but including the context doesn't do anything to change how incredibly specific the document gets.
My implication was that if he's so difficult to accommodate for speaking arrangements, it's entirely possible that he has some extremely specific requirements for remote interviews such as this. Something along the lines of "I won't answer questions sent to me via a non-free e-mail client, lest it make acceptable the use of the very thing we're trying to change" or some such.
And before you say I'm exaggerating, the rider for speaking arrangements does specifically say you can't use non-free software to record or stream his speeches, despite the act of recording and streaming being encouraged. He really does take this sort of stuff to the logical extreme.
The most efficient diesel engines in cars can hit 50% efficiency, while you're saying that a fuel cell car can hit about 76%. That's a nice improvement, but not enough to revolutionize anything.
I've heard about such things being a possibility, but there are two issues. First, they don't exist yet (batteries exist today, gasoline-powered fuel cells in cars are many years from production). Second, they don't solve most of the problems with gasoline (cost, dependence on foreign oil).
Have you SEEN his rider agreement? Sure, that's for speaking, rather than remote interviews, but the rider for giving a simple speech includes such gems as "If you buy a captured wild parrot, you will promote a cruel and devastating practice, and the parrot will be emotionally scarred before you get it."
Fuel cells have their own set of problems. There's no distribution infrastructure for hydrogen, while there is distribution infrastructure for electricity. It also has efficiency issues, since producing hydrogen isn't all that energy efficient.
Fuel cells may be practical in the long term, but batteries are practical today.
Power cables typically use stranded conductors specifically to avoid this problem. Drop your time requirement to five minutes (half a megawatt) and use the kind of power conductors that normal people use and you've got something practical.
Current charge stations (with their existing cabling) are expected to be able to do up to 150 kW. Worst case, you use two charge cables per car, and going from 150 kW to 250 kW is suddenly not such a big leap.
Tesla's plan is to have large amounts of grid storage on-site, powered mostly by solar (by building roofs over the charge stations). Tesla claims that they should be a net-positive in terms of grid power (that they produce more power than they consume). I'm skeptical that would work once they get popular, but it does still offset a chunk of the power draw, and the grid storage on-site smooths out the surges in demand.
Even with a queue, an individual gas hose isn't in use 100% of the time. There's the time you take to pull up, get out of your car, connect the hose, disconnect the hose, pay, and drive off. Even if capacitors supported a 1 minute charge every 3 minutes, that'd probably be enough.
Personally, I think a 30 second or 1 minute target is unnecessary. It takes longer than that to refuel a regular car anyhow. Five or ten minutes is probably fine, and charging 85 kWh in 10 minutes can be done with ~500 kW. That's high, but a heck of a lot less insane than the multi-megawatt charging people are talking about here. Current high-speed chargers are hitting 120 to 135 kW today.
Tesla's cars also have 7200 cells or so, so your comparison is flawed.