The attacker sends specially crafted encrypted messages to his target, records the side-channel information from the attacks and uses that information to determine the target's private key. Now the attacker is able to impersonate the target. I would call that a real-world problem since the private key is supposed to remain private.
That is why the paper points at the vulnerability of plugins that automatically decrypt messages for notification purposes or any application where the decryption process is automated: the attacker may get to send hundreds of carefully crafted texts that exercise different parts of the modular exponentiation vulnerability before someone notices something weird is happening.
The numbers I remember reading for incandescent and halogen were 5% and 7% respectively... but those were many years ago so I would not be entirely surprised if lower thermal conductivity support wires and lower pressure nitrogen bulb fill could have improved incandescent's energy efficiency over time by 3-5% once energy efficiency became big enough of a concern either by making the 60W bulb brighter than it used to be for the same amount of power or make it draw less power for light equivalent to a standard 60W bulb.
In a book I read about electronic lamp ballasts, it was said that electronic ballasts can improve fluorescent tube efficiency from 10-11% to ~16% partly due to reducing losses in the ballast itself and also by by eliminating flicker / plasma extinction by stepping frequency up from 120Hz to 20+ kHz.
11% for LEDs seems awfully low since the most efficient monochrome emitters today have efficiencies in the neighborhood of 40% and some newer materials in labs might break 60% in the not-so-distant future.
You do realize that most of the things the flight crew "needs to check" is stuff a computer could do 100X more efficiently and accurately... course change from ATC? Check all nearby TCAS positions and vectors, cross-check those with ground and on-board radar, check topographical and weather maps/radar for obstacles, pick the safest practical path, notify ground control and potentially intersecting planes so they can adjust their paths accordingly if necessary... typical path-finding exercise, shouldn't take much longer to solve than the time it takes to gather the latest updates.
The main obstacle is people aren't comfortable with trusting machines with that degree of control but with fly-by-wire planes, we are getting into territory where we are entrusting total control to computers even if pilots are on-board since all flight inputs and outputs pass through electronics and algorithms that convert pilot and instrument inputs into adjustments across dozens of other controls to provide the most efficient response that matches the pilot's intent. It gets worse with future planes where things like stabilizers which are essential to make planes humanly flyable get replaced by software emulation using the remaining flight surfaces... and in some of those, most flight controls get replaced by a simple joystick - no point in having all the conventional controls when most of the traditional surfaces no longer exist or at the very least cannot be controlled individually anymore.
Even the jets aren't designed to be flown by a single person (though in a non-emergency situation they probably could be at a higher risk of missing something important).
In a non-emergency situation, many modern planes can be on auto-pilot from takeoff to landing inclusively... only thing they need is someone to set the autopilot course and even that could be done remotely were it not for fears of computer hijacking. The pilots are pretty much there just in case something does go wrong. With fly-by-wire making all on-board instrumentation and controls available to the computer, on-board pilots might eventually get replaced by remote pilots.
If a driver wants to get distracted, it will take more than a touch-interface lock-out to stop them from doing so... and looking at the GPS map even if you cannot interact with it already counts as a distraction on its own merits.
Changing channel or track on the car's radio is not illegal yet is still one of the most common causes of rear-endings.
The main reason for ergonomics in aircraft cockpits isn't so much due to complexity as it is due to the large number of crashes caused by seemingly insignificant layout flaws... like putting an autopilot disengage switch near a foot rest, an important indicator falling into a blind spot, similar indicators getting confused with something else, bad lighting causing control positions to get misread, etc. So they need to find the best spot for everything to reduce strain on pilots and potential for pilot errors.
Much of the apparent complexity of cockpits is because planes cannot stop on a cloud for cloud-side assistance so cockpits have tons of extra dials and gauges that serve absolutely no purpose during normal flights but still need to be in locations that make sense during a potential crisis where they may be necessary to troubleshoot and hopefully recover from a fault or failure instead of crashing. I bet car dashboards would be pretty complex too if everything under the hood was exposed to a similar extent.
Driving does not require the same type of "intelligence" as desk work does... people with things like non-verbal learning disorders can have a PhD in physics and describe the physics of driving in exhaustive details but still be lost causes when you put them behind the wheel because they lack the spatial awareness to put the theory in practice while other people have no clue how the physics work... they simply look where they want to go, follow safe driving common sense and safely get there.
You do not need 2K for supraconduction: there is at least one class of supraconductor ceramics that works at temperatures as high as 135K and another (YBCO) that works at 92K which makes them relatively simple to cool by simply using liquid nitrogen. Most of the others operate in the 25-55K range.
Sure, some of the costs are (relatively) fixed. But in a mature market where there is almost no differentiation between years and manufacturers, many of them also go away due to no longer having any reasons to release, distribute and support multiple new models each year so the R&D and other costs can be spread over 2-3 years instead of being a recurring yearly expense across multiple slightly different versions of practically the same product.
And how many times would you buy LCD from a manufacturer whose models consistently fail after 2-3 years before looking elsewhere? Both of my LG LCDs started malfunctioning after about two years and in both cases, it turned out to be due to under-rated caps in their PSU. Replaced those caps and now they are into their 7th year... 5 years of extra life for a ~$0.10 difference on the original BOM to get caps rated for 2A RMS ripple instead of 0.7A or ~$1.50 worth of replacement parts. Cutting a $0.10 corner on a $300 device that makes such a drastic difference in usable life is quite insulting.
Cellphones and tablets are not quite there yet but they will be about on par with today's PCs in about three years and today's PCs are already grossly over-powered for most people's everyday needs, which is why desktop PC sales have started dropping 10-15% per year as more people go from 2-3 years upgrade cycles to 4-7 years. What effect does this have on manufacturers? Some are bailing out of the market, some go bankrupt because the market can no longer support as many players while others scrap redundant product lines, stretch product cycles, expand into different product categories, etc. to reduce their duplicated cost overheads. Cyclic market crunches have forced most HDD, DRAM, CPU, GPU and LCD manufacturers either into bankruptcies, sell-outs or mergers; future crunches will likely do the same for other components in due time.
The only reason phone and tablet manufacturers can afford wasting so many resources on product diversity is simply because the market's desire for new shiny toys with better specs has not been saturated yet. Give it another 3-5 years.
Agreed, the standard warranty on most consumer goods is a sad thing.
Countries genuinely seeking to improve the environment and reduce their landfill and recycling costs should increase minimum warranty to 3-7 years (except on batteries) based on the goods' obsolescence curve. Things like TVs, PC monitors and standard power supplies that usually remain practically unchanged for 10+ years at a time should have 7 years warranty. Things with relatively slow but steady progress curve like today's PCs should carry 5 year warranties, HDDs and other devices with generally steady wear characteristics such as OLED panels or are highly accident-prone like portable electronics could be at 3 years.
If a manufacturer makes a $10 margin on a $100 device that lasts an average of two years, I would gladly pay $120 for the same device that lasts 5+ years instead, assuming I do not expect to need or want a newer model before then. The manufacturers make $30 up-front + interests instead of 2.5x $10, I save at least $130 overall and the environment saves about two spots in a landfill somewhere.
Governments might not like the economic slowdown from people having to periodically replace goods that are currently engineered for failure or on the obsolescence fast path less than half as often though.
If you want your fiber cable to be rodent-resistant, use armored cable. Corning sells just about all their fiber cables in both armored and non-armored versions so I'm guessing most other major manufacturers have a fair selection of armored cables to choose from as well.
For residential and patch panel applications, Corning also has ClearCurve fiber cable that is nearly unbreakable - the fiber itself has less than 5mm bend radius and is encased in a densely packed sleeve that provides self-bend-limiting (nearly impossible to break the fiber by folding the cable on itself or making knots in it) and mechanical protection so the cable can be tacked to walls using regular cable staplers.
VDSL2 is great in real-life too as long as the outside plant is in reasonably good condition and people keep their expectations within the realm of what is theoretically possible. As long as the outside plant is in good shape, xDSL can deliver fairly close to its maximum "lab" speed for a given distance.
The theory behind xDSL can easily predict the performance drop-off with distance due to wire attenuation and other factors that scale with distance and would affect performance even under lab conditions just as much. xDSL is designed to work on less-than-ideal wiring and is becoming better at it as more cost-effective components and processing power become available to mitigate more of those factors that can be - such as pair crosstalk within wire binders using vectoring.
Unless you plan to fit each home with medical-grade equipment - you do not want misdiagnosis due to incorrect color calibration, telemedicine would be mostly between medical offices and remote exam offices where adequate equipment is available. For a simple video consultations, even 5Mbps would get the job done in HD since encoding mostly static scenes requires almost no bandwidth.
As for higher learning, what does higher learning requires so much more bandwidth for? Classroom presentations show little more than a blackboard and the teacher which are practically static and require almost no bandwidth to encode even at UHD resolutions. Most other classroom materials are text and images, nothing specially bandwidth-intensive there either. The only sort-of-bandwidth-intensive application I can think of is remote desktop and for most CAD and other software I have used from remote, 5Mbps was already enough to be usable under most circumstances. With a lot of specialized software, you can also simply setup a campus license server to lease out floating licenses to students running the software locally on their PC instead of having to host, build and maintain a large server farm.
Simple: wiring the middle two pairs this way makes the wiring compatible with RJ11 phone plugs so the RJ41 jacks can be repurposed between data or standard phones by simply changing a patch cable in the wiring closet.
The "large band" (the one nearest to the cable/grip) is the last in and first out so that part of the plug itself does not cause the shorts since it never contacts anything else but contacts inside the jack may short across rings during insertion and removal.
A live 1/8" plug is certainly very short-prone - imagine if AC outlets were switched around so we had exposed live copper lugs in unused outlets. A live 1/8" jack is much safer with no exposed live metal bits. As for audio accessories that used to use 1/8" plugs for power, those adapters usually were simple iron core transformers and these things can take a fair amount of abuse... particularly those without built-in rectifiers - but even rectifiers can take some abuse too, as long as they get some time to cool off between events.
There is a simple fix for most wall-wart power strip issues: getting a power bar with transformer pads like APC's SurgeArrest 11 which has six sideways plugs spaced far enough apart to accommodate just about anything that might reasonably hang off a wall. I have two of those and do not remember running into an adapter that does not fit - though I did have to shuffle adapters to squeeze smaller ones next to oversized ones.
Normal Ethernet cables are wired exactly the same way at both ends. All modern Ethernet controllers and switches automatically detect wiring type (straight or cross-over) and configure themselves accordingly.
Duplicating all contacts on both sides would make the connectors wider than necessary so USB-C will most likely use similar cable detection. All this requires is a tiny bit of logic in the interface controller and fully-fledged RX/TX electronics on both data lines or analog switches to internally re-route signals between pads and circuitry.
Power would be the only thing that requires contact duplication so polarities remain the same regardless of orientation.
If USB3 is really intended to scale much beyond 10Gbps and deliver more power to active devices, I would expect the pinout to end up looking something like this: D0+ 0V 5V 0V D1- D0- 0V 5V 0V D1+
You may have "surplus power" but you are also out-of-position and will need to use that power to get back where you were supposed to be.
To generate a decent amount of power with a wind turbine, you have to ram it into headwinds and this generates considerable drag. If you try heading into headwinds with a sailing plane, you will bleed speed and crash pretty quickly.
Sure, sailing planes can use thermals and hopefully be able to switch wind currents to hopefully get back to your starting position without having to lose power and also recover altitude in the process... but there aren't many of those at 15km altitude.
And TFA's illustrations show the platforms using fans to generate lift so even if they used thermals to move between locations, a substantial chunk of power from the solar panel would be going to lift fans. This means every extra gram of hardware comes at a substantial premium on net output.
So, as-presented, the project has all the makings of a cartoon physics experiment.
As has been pointed by others, generating power using untethered wind turbines is cartoon physics: whatever work the wind does against the turbine needs to be countered by an equal and opposite force to prevent the turbine from drifting away. Since conversion from force to energy and from energy to thrust is far less than 100%, you end up with a net energy deficit.
So, the floating untethered wind turbine idea is doomed based on Newton's third law. Cannot be done unless they find ways to break fundamental laws of physics.
The ability to focus a beam depends upon frequency and antenna size. You simply can't focus a radio signal to within one degree. Not even microwave.
That depends on the size of dish you can afford. The nice thing about hypothetical/theoretical scenarios is you can throw practicality and economic viability aside much like all those wireless floating power plant plans do.
What is the beam width of the Arecibo radio-telescope? Probably beats 1 degree at 1GHz by a fair margin.
But kites are very much tethered: the cable between the kites and their operator is what keeps them at their relatively fixed location so kites do not need to waste energy trying to maintain lift or position.
TFA's hypothetical turbines are at 15km altitude, which crosses the ~10km altitude where commercial planes cruise at, which is impractical and highly hazardous for cables so those kites would need to somehow hold their position on their own. This is unlikely to leave them with much spare power to beam down.
Tesla's coils were omnidirectional and operating at relatively low frequencies.
Give the TX and RX coils an antenna with 1 degree beam width, electronic switches instead of spark gaps, operate at ~1GHz instead of ~1MHz, etc. and you can probably achieve orders of magnitude better efficiency than Tesla's design. Still won't beat copper by a long shot but likely good enough to provide fair amounts of usable power when nothing else is feasible.
Every time I read about wireless power, I cannot help thinking about one of those old SimCity versions where orbital solar panels would occasionally malfunction and burn off random city blocks around receiver stations with their off-target microwave transmitter.
The biggest problem with untethered airborne turbines is the amount of power spent on keeping them in the air at their designated location. The other major problem is the efficiency of beaming power down through either RF or light beams - not particularly efficient.
The tethered helium or hydrogen filled canvas baloon-turbines does sound more feasible to me as well: make the turbine large enough to lift the whole thing including cables, the gas fill provides the lift so no additional power needs to be wasted on that and the tether provides ground attachment to keep the turbines in the general area where they are supposed to be which means little to no power wasted on positioning either. Copper cables in the tether also eliminate the need for energy conversions for transmission. Maybe not as neat but at least it does not have as many logic-defying challenges.
Slashdot Mirror
It doesn't?
The attacker sends specially crafted encrypted messages to his target, records the side-channel information from the attacks and uses that information to determine the target's private key. Now the attacker is able to impersonate the target. I would call that a real-world problem since the private key is supposed to remain private.
That is why the paper points at the vulnerability of plugins that automatically decrypt messages for notification purposes or any application where the decryption process is automated: the attacker may get to send hundreds of carefully crafted texts that exercise different parts of the modular exponentiation vulnerability before someone notices something weird is happening.
The numbers I remember reading for incandescent and halogen were 5% and 7% respectively... but those were many years ago so I would not be entirely surprised if lower thermal conductivity support wires and lower pressure nitrogen bulb fill could have improved incandescent's energy efficiency over time by 3-5% once energy efficiency became big enough of a concern either by making the 60W bulb brighter than it used to be for the same amount of power or make it draw less power for light equivalent to a standard 60W bulb.
In a book I read about electronic lamp ballasts, it was said that electronic ballasts can improve fluorescent tube efficiency from 10-11% to ~16% partly due to reducing losses in the ballast itself and also by by eliminating flicker / plasma extinction by stepping frequency up from 120Hz to 20+ kHz.
11% for LEDs seems awfully low since the most efficient monochrome emitters today have efficiencies in the neighborhood of 40% and some newer materials in labs might break 60% in the not-so-distant future.
You do realize that most of the things the flight crew "needs to check" is stuff a computer could do 100X more efficiently and accurately... course change from ATC? Check all nearby TCAS positions and vectors, cross-check those with ground and on-board radar, check topographical and weather maps/radar for obstacles, pick the safest practical path, notify ground control and potentially intersecting planes so they can adjust their paths accordingly if necessary... typical path-finding exercise, shouldn't take much longer to solve than the time it takes to gather the latest updates.
The main obstacle is people aren't comfortable with trusting machines with that degree of control but with fly-by-wire planes, we are getting into territory where we are entrusting total control to computers even if pilots are on-board since all flight inputs and outputs pass through electronics and algorithms that convert pilot and instrument inputs into adjustments across dozens of other controls to provide the most efficient response that matches the pilot's intent. It gets worse with future planes where things like stabilizers which are essential to make planes humanly flyable get replaced by software emulation using the remaining flight surfaces... and in some of those, most flight controls get replaced by a simple joystick - no point in having all the conventional controls when most of the traditional surfaces no longer exist or at the very least cannot be controlled individually anymore.
Even the jets aren't designed to be flown by a single person (though in a non-emergency situation they probably could be at a higher risk of missing something important).
In a non-emergency situation, many modern planes can be on auto-pilot from takeoff to landing inclusively... only thing they need is someone to set the autopilot course and even that could be done remotely were it not for fears of computer hijacking. The pilots are pretty much there just in case something does go wrong. With fly-by-wire making all on-board instrumentation and controls available to the computer, on-board pilots might eventually get replaced by remote pilots.
If a driver wants to get distracted, it will take more than a touch-interface lock-out to stop them from doing so... and looking at the GPS map even if you cannot interact with it already counts as a distraction on its own merits.
Changing channel or track on the car's radio is not illegal yet is still one of the most common causes of rear-endings.
The main reason for ergonomics in aircraft cockpits isn't so much due to complexity as it is due to the large number of crashes caused by seemingly insignificant layout flaws... like putting an autopilot disengage switch near a foot rest, an important indicator falling into a blind spot, similar indicators getting confused with something else, bad lighting causing control positions to get misread, etc. So they need to find the best spot for everything to reduce strain on pilots and potential for pilot errors.
Much of the apparent complexity of cockpits is because planes cannot stop on a cloud for cloud-side assistance so cockpits have tons of extra dials and gauges that serve absolutely no purpose during normal flights but still need to be in locations that make sense during a potential crisis where they may be necessary to troubleshoot and hopefully recover from a fault or failure instead of crashing. I bet car dashboards would be pretty complex too if everything under the hood was exposed to a similar extent.
Driving does not require the same type of "intelligence" as desk work does... people with things like non-verbal learning disorders can have a PhD in physics and describe the physics of driving in exhaustive details but still be lost causes when you put them behind the wheel because they lack the spatial awareness to put the theory in practice while other people have no clue how the physics work... they simply look where they want to go, follow safe driving common sense and safely get there.
If you needed supraconduction to start a car, you could be in trouble unless you lived on one of Jupiter's moons or something..
You do not need 2K for supraconduction: there is at least one class of supraconductor ceramics that works at temperatures as high as 135K and another (YBCO) that works at 92K which makes them relatively simple to cool by simply using liquid nitrogen. Most of the others operate in the 25-55K range.
Sure, some of the costs are (relatively) fixed. But in a mature market where there is almost no differentiation between years and manufacturers, many of them also go away due to no longer having any reasons to release, distribute and support multiple new models each year so the R&D and other costs can be spread over 2-3 years instead of being a recurring yearly expense across multiple slightly different versions of practically the same product.
And how many times would you buy LCD from a manufacturer whose models consistently fail after 2-3 years before looking elsewhere? Both of my LG LCDs started malfunctioning after about two years and in both cases, it turned out to be due to under-rated caps in their PSU. Replaced those caps and now they are into their 7th year... 5 years of extra life for a ~$0.10 difference on the original BOM to get caps rated for 2A RMS ripple instead of 0.7A or ~$1.50 worth of replacement parts. Cutting a $0.10 corner on a $300 device that makes such a drastic difference in usable life is quite insulting.
Cellphones and tablets are not quite there yet but they will be about on par with today's PCs in about three years and today's PCs are already grossly over-powered for most people's everyday needs, which is why desktop PC sales have started dropping 10-15% per year as more people go from 2-3 years upgrade cycles to 4-7 years. What effect does this have on manufacturers? Some are bailing out of the market, some go bankrupt because the market can no longer support as many players while others scrap redundant product lines, stretch product cycles, expand into different product categories, etc. to reduce their duplicated cost overheads. Cyclic market crunches have forced most HDD, DRAM, CPU, GPU and LCD manufacturers either into bankruptcies, sell-outs or mergers; future crunches will likely do the same for other components in due time.
The only reason phone and tablet manufacturers can afford wasting so many resources on product diversity is simply because the market's desire for new shiny toys with better specs has not been saturated yet. Give it another 3-5 years.
Agreed, the standard warranty on most consumer goods is a sad thing.
Countries genuinely seeking to improve the environment and reduce their landfill and recycling costs should increase minimum warranty to 3-7 years (except on batteries) based on the goods' obsolescence curve. Things like TVs, PC monitors and standard power supplies that usually remain practically unchanged for 10+ years at a time should have 7 years warranty. Things with relatively slow but steady progress curve like today's PCs should carry 5 year warranties, HDDs and other devices with generally steady wear characteristics such as OLED panels or are highly accident-prone like portable electronics could be at 3 years.
If a manufacturer makes a $10 margin on a $100 device that lasts an average of two years, I would gladly pay $120 for the same device that lasts 5+ years instead, assuming I do not expect to need or want a newer model before then. The manufacturers make $30 up-front + interests instead of 2.5x $10, I save at least $130 overall and the environment saves about two spots in a landfill somewhere.
Governments might not like the economic slowdown from people having to periodically replace goods that are currently engineered for failure or on the obsolescence fast path less than half as often though.
If you want your fiber cable to be rodent-resistant, use armored cable. Corning sells just about all their fiber cables in both armored and non-armored versions so I'm guessing most other major manufacturers have a fair selection of armored cables to choose from as well.
For residential and patch panel applications, Corning also has ClearCurve fiber cable that is nearly unbreakable - the fiber itself has less than 5mm bend radius and is encased in a densely packed sleeve that provides self-bend-limiting (nearly impossible to break the fiber by folding the cable on itself or making knots in it) and mechanical protection so the cable can be tacked to walls using regular cable staplers.
VDSL2 is great in real-life too as long as the outside plant is in reasonably good condition and people keep their expectations within the realm of what is theoretically possible. As long as the outside plant is in good shape, xDSL can deliver fairly close to its maximum "lab" speed for a given distance.
The theory behind xDSL can easily predict the performance drop-off with distance due to wire attenuation and other factors that scale with distance and would affect performance even under lab conditions just as much. xDSL is designed to work on less-than-ideal wiring and is becoming better at it as more cost-effective components and processing power become available to mitigate more of those factors that can be - such as pair crosstalk within wire binders using vectoring.
Unless you plan to fit each home with medical-grade equipment - you do not want misdiagnosis due to incorrect color calibration, telemedicine would be mostly between medical offices and remote exam offices where adequate equipment is available. For a simple video consultations, even 5Mbps would get the job done in HD since encoding mostly static scenes requires almost no bandwidth.
As for higher learning, what does higher learning requires so much more bandwidth for? Classroom presentations show little more than a blackboard and the teacher which are practically static and require almost no bandwidth to encode even at UHD resolutions. Most other classroom materials are text and images, nothing specially bandwidth-intensive there either. The only sort-of-bandwidth-intensive application I can think of is remote desktop and for most CAD and other software I have used from remote, 5Mbps was already enough to be usable under most circumstances. With a lot of specialized software, you can also simply setup a campus license server to lease out floating licenses to students running the software locally on their PC instead of having to host, build and maintain a large server farm.
Simple: wiring the middle two pairs this way makes the wiring compatible with RJ11 phone plugs so the RJ41 jacks can be repurposed between data or standard phones by simply changing a patch cable in the wiring closet.
Well, with auto-MDX, it does not really matter since it makes patch and cross-over cables interchangeable.
The "large band" (the one nearest to the cable/grip) is the last in and first out so that part of the plug itself does not cause the shorts since it never contacts anything else but contacts inside the jack may short across rings during insertion and removal.
A live 1/8" plug is certainly very short-prone - imagine if AC outlets were switched around so we had exposed live copper lugs in unused outlets. A live 1/8" jack is much safer with no exposed live metal bits. As for audio accessories that used to use 1/8" plugs for power, those adapters usually were simple iron core transformers and these things can take a fair amount of abuse... particularly those without built-in rectifiers - but even rectifiers can take some abuse too, as long as they get some time to cool off between events.
There is a simple fix for most wall-wart power strip issues: getting a power bar with transformer pads like APC's SurgeArrest 11 which has six sideways plugs spaced far enough apart to accommodate just about anything that might reasonably hang off a wall. I have two of those and do not remember running into an adapter that does not fit - though I did have to shuffle adapters to squeeze smaller ones next to oversized ones.
Normal Ethernet cables are wired exactly the same way at both ends. All modern Ethernet controllers and switches automatically detect wiring type (straight or cross-over) and configure themselves accordingly.
Duplicating all contacts on both sides would make the connectors wider than necessary so USB-C will most likely use similar cable detection. All this requires is a tiny bit of logic in the interface controller and fully-fledged RX/TX electronics on both data lines or analog switches to internally re-route signals between pads and circuitry.
Power would be the only thing that requires contact duplication so polarities remain the same regardless of orientation.
If USB3 is really intended to scale much beyond 10Gbps and deliver more power to active devices, I would expect the pinout to end up looking something like this:
D0+ 0V 5V 0V D1-
D0- 0V 5V 0V D1+
You may have "surplus power" but you are also out-of-position and will need to use that power to get back where you were supposed to be.
To generate a decent amount of power with a wind turbine, you have to ram it into headwinds and this generates considerable drag. If you try heading into headwinds with a sailing plane, you will bleed speed and crash pretty quickly.
Sure, sailing planes can use thermals and hopefully be able to switch wind currents to hopefully get back to your starting position without having to lose power and also recover altitude in the process... but there aren't many of those at 15km altitude.
And TFA's illustrations show the platforms using fans to generate lift so even if they used thermals to move between locations, a substantial chunk of power from the solar panel would be going to lift fans. This means every extra gram of hardware comes at a substantial premium on net output.
So, as-presented, the project has all the makings of a cartoon physics experiment.
As has been pointed by others, generating power using untethered wind turbines is cartoon physics: whatever work the wind does against the turbine needs to be countered by an equal and opposite force to prevent the turbine from drifting away. Since conversion from force to energy and from energy to thrust is far less than 100%, you end up with a net energy deficit.
So, the floating untethered wind turbine idea is doomed based on Newton's third law. Cannot be done unless they find ways to break fundamental laws of physics.
The ability to focus a beam depends upon frequency and antenna size. You simply can't focus a radio signal to within one degree. Not even microwave.
That depends on the size of dish you can afford. The nice thing about hypothetical/theoretical scenarios is you can throw practicality and economic viability aside much like all those wireless floating power plant plans do.
What is the beam width of the Arecibo radio-telescope? Probably beats 1 degree at 1GHz by a fair margin.
But kites are very much tethered: the cable between the kites and their operator is what keeps them at their relatively fixed location so kites do not need to waste energy trying to maintain lift or position.
TFA's hypothetical turbines are at 15km altitude, which crosses the ~10km altitude where commercial planes cruise at, which is impractical and highly hazardous for cables so those kites would need to somehow hold their position on their own. This is unlikely to leave them with much spare power to beam down.
Tesla's coils were omnidirectional and operating at relatively low frequencies.
Give the TX and RX coils an antenna with 1 degree beam width, electronic switches instead of spark gaps, operate at ~1GHz instead of ~1MHz, etc. and you can probably achieve orders of magnitude better efficiency than Tesla's design. Still won't beat copper by a long shot but likely good enough to provide fair amounts of usable power when nothing else is feasible.
Every time I read about wireless power, I cannot help thinking about one of those old SimCity versions where orbital solar panels would occasionally malfunction and burn off random city blocks around receiver stations with their off-target microwave transmitter.
The biggest problem with untethered airborne turbines is the amount of power spent on keeping them in the air at their designated location. The other major problem is the efficiency of beaming power down through either RF or light beams - not particularly efficient.
The tethered helium or hydrogen filled canvas baloon-turbines does sound more feasible to me as well: make the turbine large enough to lift the whole thing including cables, the gas fill provides the lift so no additional power needs to be wasted on that and the tether provides ground attachment to keep the turbines in the general area where they are supposed to be which means little to no power wasted on positioning either. Copper cables in the tether also eliminate the need for energy conversions for transmission. Maybe not as neat but at least it does not have as many logic-defying challenges.