To make things a bit more blatant, Deep learning networks tend to be biased to find what they are taught to find. If the teacher is biased, so with the AI be.
Well, there *is* another approach that has been used, but I don't know on what scale. What you do it take a crystal (probably some kind of glass) and position certain tiny "fluorescent" "drops" carefully within it. These drops only fluoresce if stimulated appropriately. This is done by hitting them with two appropriately chosen laser frequencies from two different directions. Only where the beams intersect is there a visible light emission. You've got to scan rapidly. The tricky part is building the "crystal ball". This dates back at least to the 90's. They were talking about using it as a model for a diamond memory. I haven't heard about it since then. (I've heard about the diamond memory, but not the 3-D model.) My guess is that it's still technically too difficult (which partially means other ways are easier).
IIUC, sometimes the Japanese singer is projected onto a fog, but you're going to need SOME secondary radiator or you won't see the light. The only other way is to shine the light directly at you, and with a laser that's not exactly recommended. And using a secondary radiator doesn't mean that the original projection wasn't a hologram. (Was it? Possibly not technically. But words generally have a loose meaning as well as a tight meaning, and I wasn't talking about the mathematical definition, which nothing existing really quite fits.)
Well, ok, there's another approach, though you might not want to call that a hologram either: What you do is compute what light would be emitted at each point intersecting with a surface and then have a REALLY LARGE computer screen generate that light. You need to have LOTS of (very short) light pipes inside the screen, because each point needs to generate different lights headed in different directions. With this system each watcher needs to be tracked so you can generate the image that they should see. VR glasses are a better way to accomplish this.
Since what I was originally discussing was fake 3-D images of a finely grained ceiling, the fancy ways and precise definitions seem unreasonable. A fake moose head over the fireplace doesn't need to be precisely tuned, especially if the room is so designed that nobody can get within 3 feet of it. A wood paneled ceiling is even easier...and still beyond reasonable availability. (Remember, was I was asserting is that robot installed ceilings would get there first.)
No. The existing ones have to be counted as rudimentary as well as washing out on bright light and needing lot of maintenance. They are lab curiosities. But they *may* not stay that way. I was mentioning them as an example of something that probably won't be significant...but which *MIGHT* be if some unexpected things happen (which would need to include *both* technical improvement as fad).
Just remember that at one time the estimate was there might be a market for 6 computers in the whole world. Now granted, that estimate was made awhile ago... (1943.) It took a large number of technical changes to get from then to now, but it also took a bunch of social changes.
Well, the reason holographic projections are unlikely to replace actual things include that they tend to wash out in bright light and that the current projectors are expensive and require a lot of maintenance. So I think them unlikely. But they *might* happen. It would probably require a combination to technical improvements and fashion changes, but those are both possibilities, and it's possible that just the right ones could happen so that....
Which is what I meant about the difficulty in predicting the future. It's easy to look at things and say "The current situation is unstable, and it could go in this direction...", but saying it *will* go in this direction is something more difficult. Saying "It's not going to stay the same" is a safe prediction, but nearly useless. So.
Consider supermarket checkout clerks. Most supermarkets I go to have lots of unused lines, and they used to be full all the time. This is partially because of bar codes speeding up the job of the clerks, and partially because of self-checkout. And perhaps it's partially that you don't get people coming in all at the same time as much, and partially home deliver of items. So even after the fact I don't really know why many of those jobs disappeared, though I can spot contributing factors, some of which one could call automation, though hardly robots. Or think of the way the cashier job at a fast-food joint has been deskilled.
And this is the reason the article survey came up with that stupid opinion. Current robots can't do your job, so you don't believe future robots will cause you any problem. But they will. Just how they'll do it is uncertain, since there are multiple ways in which it could happen. They might just become better than you at detail work, flexibility, and aesthetic judgment, but that's only one way it could happen, and probably not the fastest. More likely they'll more quickly lead to a redesign of methods of building into ways that are easy for them and hard for you. There will still be a few really rich people who will want the old style enough to pay for it and also be wealthy enough, but the number of jobs will decrease by a huge factor, and most carpenters will end up unemployed. The real robot carpenters may be a decade behind the first wave.
OTOH, predicting the timeline is tricky. It depends on lots of other things. E.g., if the robots need to carry their smarts in their body it will take longer than if they are able to run off some wireless transmission, but the wavebands are already getting jammed, so it will need to be some fairly high bandwidth short range transmission. An optic link might do it, or infra-red or even low power microwave.
People are lousy at predicting the future because the actual future is the result of LOTS of things currently under development, most of which won't go anywhere, but some of which will be more spectacularly successful than anyone expects. Maybe holographic projectors will cause us not to care what anything we can't touch really looks like. Unlikely, but it's "one of the things that are under development". I doubt that VR would be able to get THAT extreme, but it could happen...given the proper changes in social custom.
OTOH, as a rule of thumb, if a prediction is for less than five years, expect it to take longer than they say, because of unexpected problems. And if it's for longer than 15 years expect it to take less time because of unexpected solutions. And note that there's a lot of uncertainty in this. And people's predictions of when something will happen are often biased in ways that confound this rule of thumb, you've also got to consider how reasonable the prediction is, which is hard if you aren't "skilled in the art".
P.S.: You, the carpenter, probably won't immediately be able to justify buying the robot. But the contractor who hires you will be able to justify it a lot sooner. Especially if it's also a plumber.
Money is not wealth. Money is something that can usually be exchanged for wealth, but the rate of conversion varies from time to time, situation to situation, and person to person.
Wealth is, basically, "degree of wellness", but I'm not talking only about health. And this means that for some problems money doesn't even help you achieve wealth. Sometimes wealth is "my good friends". How much money do you think would be a fair trade for ending up castaway on an island with nobody else on it and nobody looking for you?
OK, that's an extreme case, but not the only one. The problem is that it's easy to measure money, and hard to measure wealth, so people tend to focus on money as a useful surrogate. This is reasonable, but it's important to remember that they aren't the same.
There *are* no safe investments. Different failure modes wipe out different strategies, and some cause previously failing modes to succeed.
One thing you need to remember is that money is governmental accounting for your worth. And all governments cheat some group. (People disagree about exactly what "cheat" means, but for every definition I'm aware of, the prior statement remains true.) Bitcoin, etc., are parasitic on a working government...without a government you won't have an enabling internet. But survivalism is a bet that things are going to fall apart, and fails if they don't.
If the government allowed the creation of additional money at the exact same rate as the production of wealth happened, then there would be no inflation...but nobody even knows how to measure "wealth". It sure isn't money. And this doesn't even address the distribution problem. When robots increasingly do the work, justify a method of distribution. (I'm not talking about when they do all the work. I'm assuming that there will still be some necessary jobs that aren't automated.)
Now to add to the complexity of the problem, most wealth is intangible, and generated by societies rather than individuals. Knowledge of how to produce a hardened steel screw from raw materials, etc. The machinery needed for each way of making that screw is also wealth, but the knowledge of how to do that is the valuable, and took centuries to develop...and part of it was the knowledge that a hardened steel screw was valuable. Without knowing that you don't even aim to produce one. What any one person can add is a very thin layer on top of an immense existing structure. There is a reason that three people in different countries tried to patent the phone at the same time. "Winner takes all." is a hideous travesty. A patent should say something like "The government won't buy or pay for this device if made by someone else.". And it should require the production of a working model. And the patent should be specific. These idiotic generic patents are even an abuse of existing law.
Sorry, I forgot that rise isn't evenly distributed around the globe. Locally it's fractions of an inch/decade (though not small fractions) but on the US East Coast it's considerably more. But still a lot less than 2 inches/decade.
Solar panels will need more than hosing down. They'll need to be lifted vertically. (Of course if they're on a roof top, they'll get lifted automatically when the house is jacked up onto stilts...though I don't know how well that plays with hurricane wind resistance.
I'm assuming that the flooding you are talking about is due to sea level rise (which is measured in fractions of an inch per decade)...but which does mean that some places that used to be usually dry now frequently end up under a bit of water. If you're talking about storm surges, this only gets amplified a little bit unless it's really true that move violent storms are becoming more frequent.
Sorry, but solar cells are more vulnerable to high winds than are nuclear plants. But if you are expecting a site to have high winds, they can be configured to have a low wind profile, and also be protectively covered in time of need. No big deal, if you're expecting the problem.
Still, I wouldn't like to see the kind of hurricane that could make ANY impression on a concrete confinement dome.
You are vastly overstating a valid point. Reactor problems can, and have, affected large areas. Solar panel problems only affect small areas. But neither noticeably affects 1/8 of the planet. Hundreds of miles, yes, and even that's in strongly decreased amount.
Chernobyl is a better extreme case than Fukishima, because it affected people as far away as Sweden. But the effects were minor.
There are worse case events than we have seen so far, but even one of them wouldn't noticeably affect 1/8th of the globe.
Still, if nuclear plants had to budget for all their externalized costs they'd be out of business. But so would coal. And oil would be a lot more limited. I don't have a good figure for solar or wind, because they both use lots of specialized materials that have a hefty environmental cost to extract, and they don't handle recycling their retired or non-functional products. I suspect both would easily double in costs. Society tends to hide the real costs of the sources of energy that it depends on for existence.
I've seen the claim before that molten salt reactors are inherently safe. It may be true. (It does seem plausible.)
This, however, is different from saying there won't be an accident. The claim I saw said something more like "and accident won't cause any problems outside the reactor".
It is also true that technology advances often yield greater reliability. But this isn't guaranteed. Sometimes they optimize something else, perhaps construction costs, or permit availability. I usually avoid using the first iteration of any technology. I wait for the first bug fix release, or at any rate wait until a track record has been established. And this is through long and bitter experience. But if only a few instances of a piece of technology is available to examine (say a new nuclear power plant design) on what basis should I assume that it is safe?
A secondary consideration is that companies siting nuclear plants seem to frequently chose horrendously inappropriate sites. A site which has historically repeatedly been swept by tsunami, for example, or at the bottom of a canyon in earthquake country. (Not that at the top of a canyon wall would be any better.) In these cases it hardly matters WHAT the plant technology is, that will only limit the magnitude of the catastrophe. And safely operating for a decade isn't an indication that it's a safe site. The known problems are also known to be both infrequent and unpredictable. In the case of Fukishima, e.g., the tsunami wasn't the maximum possible one, but a technology that allowed the plant to be a closed module would have limited the damage significantly. (In the current instance, even protecting the backup power supply would have limited the damage significantly, but it could have been more extreme.)
Actually, most of the regulations were put into place because of the possibility of incidents, not because of incidents that actually happened. This doesn't mean that they're a bad idea. It's also dubious when somebody claims "we've learned from this" because management keeps changing, and people have an extremely strong tendency to discount future costs in favor of short term gains.
Additionally, any particular incident usually has a string of antecedent causes, and singling one out as "the reason" is usually a mistake. Usually several different chains of causation have moved one closer to the incident, and any of several different subsequent chains of causation could move the incident into actuality. E.g.: In the case of Fukishima what was learned? Use a high seawall? Site the plant higher? Have redundant backup generators? Don't put your backup generators in the basement? Avoiding any one of those errors would have changed this incident into something MUCH less significant. But the thing is there are probably lots of things that didn't happen that could have, and any particular change would only deal with some of them.
In the case of Fukishima the thing that should *really* have been done is protect the backup generators. That would have been the cheap way to avoid the current problem. But what should really have been done is "All of the things mentioned, and a few you haven't thought of".
The existing generation of nuclear reactors is dangerously unsafe. Possibly the pebble bad or molten salt reactors would be inherently safe. Certainly I've never heard of a major problem with the swimming-pool reactors...but they must be too expensive. Still, it seems an existence proof for an "inherently safe reactor design". If something happens the water drains out or goes up as steam and then it no longer moderates the reactor so it shuts down. This might do a lot of damage to the reactor, but it wouldn't damage much outside the reactor.
OTOH, the problem about reactors that really bothers me is radioactive waste. Some reactor designs are designed to burn that stuff up within the reactor. That's what we need to be using even if there's a bit of damage to the "inherently safe design", and others have claimed that there doesn't need to be.
IIUC, a part of the problem was lack of adequate drainage. From my understanding the standby generator was not merely submerged by the tsunami, it was left standing in a pool of water that didn't (rapidly?) drain away. If the emergency generator had been able to be started within a couple of hours the damage would have been much less. I seem to remember that the proper evaluation is closer to *EXTREMELY* much less. As in it could have kept the fuel rods from melting.
If I'm correct, the problem was not only an improper sea wall, and not merely improper siting of the backup generators, but even basic things like drainage for the backup generators.
P.S.: Many of the plants being operated in the US are beyond their designed lifetime, and being run at higher power levels than they were rated for. I hope people are checking on their backup power supplies and refrigeration.
Actually, I think it is obviously true. Just not significant. Easier means a trifle less work, so there is less of a look-up required. Significantly isn't true, because library calls aren't that hard.
That's not what the incompleteness theorem says. It says: Any sufficiently complex system has true statements that cannot be proven true within the system. This is not the same as statements which are both true and false. If there are statements which are both true and false then the system is not necessarily incomplete, but merely inconsistent. (It may also be incomplete.)
Slashdot Mirror
To make things a bit more blatant,
Deep learning networks tend to be biased to find what they are taught to find. If the teacher is biased, so with the AI be.
It's probably not that nobody thought of it, but rather of strong perverse incentives...as you indicate in the first part of your post.
Well, there *is* another approach that has been used, but I don't know on what scale. What you do it take a crystal (probably some kind of glass) and position certain tiny "fluorescent" "drops" carefully within it. These drops only fluoresce if stimulated appropriately. This is done by hitting them with two appropriately chosen laser frequencies from two different directions. Only where the beams intersect is there a visible light emission. You've got to scan rapidly. The tricky part is building the "crystal ball".
This dates back at least to the 90's. They were talking about using it as a model for a diamond memory. I haven't heard about it since then. (I've heard about the diamond memory, but not the 3-D model.) My guess is that it's still technically too difficult (which partially means other ways are easier).
Only if you drop them right next to some hungry alligators. Most people will survive Irma.
From *this* government???
IIUC, sometimes the Japanese singer is projected onto a fog, but you're going to need SOME secondary radiator or you won't see the light. The only other way is to shine the light directly at you, and with a laser that's not exactly recommended. And using a secondary radiator doesn't mean that the original projection wasn't a hologram. (Was it? Possibly not technically. But words generally have a loose meaning as well as a tight meaning, and I wasn't talking about the mathematical definition, which nothing existing really quite fits.)
Well, ok, there's another approach, though you might not want to call that a hologram either: What you do is compute what light would be emitted at each point intersecting with a surface and then have a REALLY LARGE computer screen generate that light. You need to have LOTS of (very short) light pipes inside the screen, because each point needs to generate different lights headed in different directions. With this system each watcher needs to be tracked so you can generate the image that they should see. VR glasses are a better way to accomplish this.
Since what I was originally discussing was fake 3-D images of a finely grained ceiling, the fancy ways and precise definitions seem unreasonable. A fake moose head over the fireplace doesn't need to be precisely tuned, especially if the room is so designed that nobody can get within 3 feet of it. A wood paneled ceiling is even easier...and still beyond reasonable availability. (Remember, was I was asserting is that robot installed ceilings would get there first.)
https://www.google.com/search?...
I didn't see any refs there to the Japanese pop-singer who's a computer generated hologram, but it's also real.
I think VR glasses will be *MUCH* more likely, and therefore not as good an example of what I meant. Also, they'd have a very different use case.
No. The existing ones have to be counted as rudimentary as well as washing out on bright light and needing lot of maintenance. They are lab curiosities. But they *may* not stay that way. I was mentioning them as an example of something that probably won't be significant...but which *MIGHT* be if some unexpected things happen (which would need to include *both* technical improvement as fad).
Just remember that at one time the estimate was there might be a market for 6 computers in the whole world. Now granted, that estimate was made awhile ago... (1943.) It took a large number of technical changes to get from then to now, but it also took a bunch of social changes.
Well, the reason holographic projections are unlikely to replace actual things include that they tend to wash out in bright light and that the current projectors are expensive and require a lot of maintenance. So I think them unlikely. But they *might* happen. It would probably require a combination to technical improvements and fashion changes, but those are both possibilities, and it's possible that just the right ones could happen so that....
Which is what I meant about the difficulty in predicting the future. It's easy to look at things and say "The current situation is unstable, and it could go in this direction...", but saying it *will* go in this direction is something more difficult. Saying "It's not going to stay the same" is a safe prediction, but nearly useless. So.
Consider supermarket checkout clerks. Most supermarkets I go to have lots of unused lines, and they used to be full all the time. This is partially because of bar codes speeding up the job of the clerks, and partially because of self-checkout. And perhaps it's partially that you don't get people coming in all at the same time as much, and partially home deliver of items. So even after the fact I don't really know why many of those jobs disappeared, though I can spot contributing factors, some of which one could call automation, though hardly robots. Or think of the way the cashier job at a fast-food joint has been deskilled.
Don't know. I don't watch TV. If so I'm sure they jiggered the situation for dramatic tension.
And this is the reason the article survey came up with that stupid opinion. Current robots can't do your job, so you don't believe future robots will cause you any problem. But they will. Just how they'll do it is uncertain, since there are multiple ways in which it could happen. They might just become better than you at detail work, flexibility, and aesthetic judgment, but that's only one way it could happen, and probably not the fastest. More likely they'll more quickly lead to a redesign of methods of building into ways that are easy for them and hard for you. There will still be a few really rich people who will want the old style enough to pay for it and also be wealthy enough, but the number of jobs will decrease by a huge factor, and most carpenters will end up unemployed. The real robot carpenters may be a decade behind the first wave.
OTOH, predicting the timeline is tricky. It depends on lots of other things. E.g., if the robots need to carry their smarts in their body it will take longer than if they are able to run off some wireless transmission, but the wavebands are already getting jammed, so it will need to be some fairly high bandwidth short range transmission. An optic link might do it, or infra-red or even low power microwave.
People are lousy at predicting the future because the actual future is the result of LOTS of things currently under development, most of which won't go anywhere, but some of which will be more spectacularly successful than anyone expects. Maybe holographic projectors will cause us not to care what anything we can't touch really looks like. Unlikely, but it's "one of the things that are under development". I doubt that VR would be able to get THAT extreme, but it could happen...given the proper changes in social custom.
OTOH, as a rule of thumb, if a prediction is for less than five years, expect it to take longer than they say, because of unexpected problems. And if it's for longer than 15 years expect it to take less time because of unexpected solutions. And note that there's a lot of uncertainty in this. And people's predictions of when something will happen are often biased in ways that confound this rule of thumb, you've also got to consider how reasonable the prediction is, which is hard if you aren't "skilled in the art".
P.S.: You, the carpenter, probably won't immediately be able to justify buying the robot. But the contractor who hires you will be able to justify it a lot sooner. Especially if it's also a plumber.
Money is not wealth. Money is something that can usually be exchanged for wealth, but the rate of conversion varies from time to time, situation to situation, and person to person.
Wealth is, basically, "degree of wellness", but I'm not talking only about health. And this means that for some problems money doesn't even help you achieve wealth. Sometimes wealth is "my good friends". How much money do you think would be a fair trade for ending up castaway on an island with nobody else on it and nobody looking for you?
OK, that's an extreme case, but not the only one. The problem is that it's easy to measure money, and hard to measure wealth, so people tend to focus on money as a useful surrogate. This is reasonable, but it's important to remember that they aren't the same.
There *are* no safe investments. Different failure modes wipe out different strategies, and some cause previously failing modes to succeed.
One thing you need to remember is that money is governmental accounting for your worth. And all governments cheat some group. (People disagree about exactly what "cheat" means, but for every definition I'm aware of, the prior statement remains true.) Bitcoin, etc., are parasitic on a working government...without a government you won't have an enabling internet. But survivalism is a bet that things are going to fall apart, and fails if they don't.
If the government allowed the creation of additional money at the exact same rate as the production of wealth happened, then there would be no inflation...but nobody even knows how to measure "wealth". It sure isn't money. And this doesn't even address the distribution problem. When robots increasingly do the work, justify a method of distribution. (I'm not talking about when they do all the work. I'm assuming that there will still be some necessary jobs that aren't automated.)
Now to add to the complexity of the problem, most wealth is intangible, and generated by societies rather than individuals. Knowledge of how to produce a hardened steel screw from raw materials, etc. The machinery needed for each way of making that screw is also wealth, but the knowledge of how to do that is the valuable, and took centuries to develop...and part of it was the knowledge that a hardened steel screw was valuable. Without knowing that you don't even aim to produce one. What any one person can add is a very thin layer on top of an immense existing structure. There is a reason that three people in different countries tried to patent the phone at the same time. "Winner takes all." is a hideous travesty. A patent should say something like "The government won't buy or pay for this device if made by someone else.". And it should require the production of a working model. And the patent should be specific. These idiotic generic patents are even an abuse of existing law.
Sorry, I forgot that rise isn't evenly distributed around the globe. Locally it's fractions of an inch/decade (though not small fractions) but on the US East Coast it's considerably more. But still a lot less than 2 inches/decade.
Thanks. I thought that /proc/version was returning a text string, though in that case one has to wonder why I even tried kwrite.
Solar panels will need more than hosing down. They'll need to be lifted vertically. (Of course if they're on a roof top, they'll get lifted automatically when the house is jacked up onto stilts...though I don't know how well that plays with hurricane wind resistance.
I'm assuming that the flooding you are talking about is due to sea level rise (which is measured in fractions of an inch per decade)...but which does mean that some places that used to be usually dry now frequently end up under a bit of water. If you're talking about storm surges, this only gets amplified a little bit unless it's really true that move violent storms are becoming more frequent.
Sorry, but solar cells are more vulnerable to high winds than are nuclear plants. But if you are expecting a site to have high winds, they can be configured to have a low wind profile, and also be protectively covered in time of need. No big deal, if you're expecting the problem.
Still, I wouldn't like to see the kind of hurricane that could make ANY impression on a concrete confinement dome.
You are vastly overstating a valid point. Reactor problems can, and have, affected large areas. Solar panel problems only affect small areas. But neither noticeably affects 1/8 of the planet. Hundreds of miles, yes, and even that's in strongly decreased amount.
Chernobyl is a better extreme case than Fukishima, because it affected people as far away as Sweden. But the effects were minor.
There are worse case events than we have seen so far, but even one of them wouldn't noticeably affect 1/8th of the globe.
Still, if nuclear plants had to budget for all their externalized costs they'd be out of business. But so would coal. And oil would be a lot more limited. I don't have a good figure for solar or wind, because they both use lots of specialized materials that have a hefty environmental cost to extract, and they don't handle recycling their retired or non-functional products. I suspect both would easily double in costs. Society tends to hide the real costs of the sources of energy that it depends on for existence.
I've seen the claim before that molten salt reactors are inherently safe. It may be true. (It does seem plausible.)
This, however, is different from saying there won't be an accident. The claim I saw said something more like "and accident won't cause any problems outside the reactor".
It is also true that technology advances often yield greater reliability. But this isn't guaranteed. Sometimes they optimize something else, perhaps construction costs, or permit availability. I usually avoid using the first iteration of any technology. I wait for the first bug fix release, or at any rate wait until a track record has been established. And this is through long and bitter experience. But if only a few instances of a piece of technology is available to examine (say a new nuclear power plant design) on what basis should I assume that it is safe?
A secondary consideration is that companies siting nuclear plants seem to frequently chose horrendously inappropriate sites. A site which has historically repeatedly been swept by tsunami, for example, or at the bottom of a canyon in earthquake country. (Not that at the top of a canyon wall would be any better.) In these cases it hardly matters WHAT the plant technology is, that will only limit the magnitude of the catastrophe. And safely operating for a decade isn't an indication that it's a safe site. The known problems are also known to be both infrequent and unpredictable. In the case of Fukishima, e.g., the tsunami wasn't the maximum possible one, but a technology that allowed the plant to be a closed module would have limited the damage significantly. (In the current instance, even protecting the backup power supply would have limited the damage significantly, but it could have been more extreme.)
Actually, most of the regulations were put into place because of the possibility of incidents, not because of incidents that actually happened. This doesn't mean that they're a bad idea. It's also dubious when somebody claims "we've learned from this" because management keeps changing, and people have an extremely strong tendency to discount future costs in favor of short term gains.
Additionally, any particular incident usually has a string of antecedent causes, and singling one out as "the reason" is usually a mistake. Usually several different chains of causation have moved one closer to the incident, and any of several different subsequent chains of causation could move the incident into actuality.
E.g.: In the case of Fukishima what was learned? Use a high seawall? Site the plant higher? Have redundant backup generators? Don't put your backup generators in the basement? Avoiding any one of those errors would have changed this incident into something MUCH less significant. But the thing is there are probably lots of things that didn't happen that could have, and any particular change would only deal with some of them.
In the case of Fukishima the thing that should *really* have been done is protect the backup generators. That would have been the cheap way to avoid the current problem. But what should really have been done is "All of the things mentioned, and a few you haven't thought of".
The existing generation of nuclear reactors is dangerously unsafe. Possibly the pebble bad or molten salt reactors would be inherently safe. Certainly I've never heard of a major problem with the swimming-pool reactors...but they must be too expensive. Still, it seems an existence proof for an "inherently safe reactor design". If something happens the water drains out or goes up as steam and then it no longer moderates the reactor so it shuts down. This might do a lot of damage to the reactor, but it wouldn't damage much outside the reactor.
OTOH, the problem about reactors that really bothers me is radioactive waste. Some reactor designs are designed to burn that stuff up within the reactor. That's what we need to be using even if there's a bit of damage to the "inherently safe design", and others have claimed that there doesn't need to be.
IIUC, a part of the problem was lack of adequate drainage. From my understanding the standby generator was not merely submerged by the tsunami, it was left standing in a pool of water that didn't (rapidly?) drain away. If the emergency generator had been able to be started within a couple of hours the damage would have been much less. I seem to remember that the proper evaluation is closer to *EXTREMELY* much less. As in it could have kept the fuel rods from melting.
If I'm correct, the problem was not only an improper sea wall, and not merely improper siting of the backup generators, but even basic things like drainage for the backup generators.
P.S.: Many of the plants being operated in the US are beyond their designed lifetime, and being run at higher power levels than they were rated for. I hope people are checking on their backup power supplies and refrigeration.
What's weird though is that echo /proc/version yielded /proc/version even though kwrite /proc/version got the correct info.
I'm not expert, but...
Actually, I think it is obviously true. Just not significant. Easier means a trifle less work, so there is less of a look-up required. Significantly isn't true, because library calls aren't that hard.
That's not what the incompleteness theorem says. It says:
Any sufficiently complex system has true statements that cannot be proven true within the system.
This is not the same as statements which are both true and false. If there are statements which are both true and false then the system is not necessarily incomplete, but merely inconsistent. (It may also be incomplete.)