Let's say you're right, and there's a 50/50 chance of needing a follow up conversation. Great! That means you've avoided an intrusive time-wasting call 50% of the time, by sending a potentially standard pre-typed reply.
Granted, a more informative initial text would be good - but knowing clients/users/etc, a longer text is unlikely to actually be helpful, and even in a conversation it's likely to be a chore to extract useful information.
If nothing else their text is an invitation for you to call them, at *your* convenience. They're initiating communication in the least-intrusive way possible, and if it's important to them they'll likely answer your call.
Make it a federal law, and you don't have jurisdiction issues within the US. And nothing stops us from putting requirements on international calls - if a foreign phone company wants to be able to connect to US callers, they need to comply with US international-call regulations.
You missed the part about redistributing the tax income didn't you?
To use your example I didn't suggest that the government demands gold-plated plumbing - I suggested they put a $70,000 tax on all non-gold-plated plumbing, and then give that $70,000 to you. It's totally up to you whether you want to spend $75k on non-gold-plated plumbing, or spend $50k on gold-plated plumbing and the other $25k on other things.
Even logarithmic functions increase to infinity. Increase the temperature, you increase the CO2 levels, which further increases the temperature, which further increases the CO2 levels. A logarithmic increase will help that to eventually stabilize, but says nothing about *where* it stabilizes - that depends entirely on the strength of the positive feedback - and thawing permafrost, methane hydrates, etc. deliver extremely large positive feedback over a narrow temperature range - once the process begins in earnest, it's going to continue until the reserves are depleted. If we reach the point where the arctic is venting sequestered carbon at 10x the rate of human emissions, do you really think us stopping our emissions is going to stop the process? No - we will be beyond the tipping point.
And why don't my numbers make sense? Most ecologically sequestered carbon never become fossil fuels, it becomes more stable forms of rock. It takes very special circumstances to convert biomass into chemically unstable fossil fuels.
Recent studies (2009) estimate that northern circumpolar permafrost soil carbon content equals approximately 1672 Pg. (1 Pg = 1 Gt = 10^15g) This estimation of the amount of carbon stored in permafrost soils is more than double the amount currently in the atmosphere. ( https://en.wikipedia.org/wiki/... )
That's about 57x the roughly 29Gt amount we emit in a year - and worse, a whole lot of that is going to be initially released as methane, which is about 30x more potent a greenhouse gas than CO2, before eventually breaking down into CO2. And that's just circumpolar permafrost.
Meanwhile undersea methane hydrate reserves are estimated at somewhere between 100 and 530,000Gt (sea floor exploration is woefully spotty) 100 might not be that bad - under 4 years worth of human carbon emissions, though that 30x warming multiplier for methane won't help anything. 530,000 though, that's 18,000 years worrth of current human emissions, before the methane multiplier. We'd better hope the reality is a lot closer to the low-end estimates.
And no, study your paleontology a bit - there is NO guarantee a greenhouse Earth will be wetter - they also include some of the most severe global desert phases of our planet's history. Nor milder - more energy in the atmosphere, plus much weaker and more unstable/unpredictable jet streams sets the stage for far more extreme weather variation such as we're already seeing. There are no guarantees of anything. A Greenhouse Earth can be every bit as inhospitable as a glacier-covered Icehouse Earth.
I reduce the problem by having a different ringtone for unknown callers, and a voice message that tells people I screen unknown callers and please leave a message. If I get a message I know it immediately and can check, and it's rarely spam (political robocalls notwithstanding). Perhaps I've missed out on an old flame calling to reunite and chickening out at the message, but I doubt it. And when my friends change their number, or call from another phone, they generally still reach me within minutes of placing the call.
If you're going to lie to the client and and answer yes without knowing anything to shut them up, then is it really that likely that you'd be more helpful otherwise?
Here's a thought: Don't lie and answer "Yes" when you don't actually know anything - assuming you actually want to provide service to your client. If you don't know, answer something like "No. What's the issue and how big a problem is it for you?" Maybe even add a "Please call if it's serious."
The beauty of texting, in addition to being asynchronous so that it doesn't present a near-guaranteed interruption of your work flow, is that it can be very brief and minimally intrusive to both parties. Something has gone wrong at MZR - maybe it's causing lots of problems and you're in the middle of dealing with it. Do you really want to ALSO be dealing with all your clients calling in complaining about the MZR issue and asking for updates? Or would you rather get quick texts that you can occasionally stop to send a batch of quick "Yes, I'm working on it now" replies?
Phone calls are great for serious information exchange or when an immediate response is needed - but are complete intrusive overkill most of the time.
It would be easy enough if those in power actually wished to engorce them, rather than just passing symbolic legislation to make themselves look good without actually changing anything.
The phone company knows exactly who's calling you - they're charging them for the call. All they'd need to do is let you, e.g. dial *FU after a spam call to quickly report them. Allow a little leeway for false reports, but anyone who racks up a lot of reports gets prosecuted.
Increased temperature increases the rate of natural greenhouse gas stores - there's far more ecologically sequestered carbon and methane than humans would use in centuries - if our emissions warm the planet sufficiently, those reserves start being released faster than anything we've done. Same thing appears to have happened with every icehouse to greenhouse transition. Global climate is a bistable system - push it far enough and you hit the tipping point where the process becomes self-accelerating until it reaches the opposite state.
>And during that time, we have experienced massive climate change and temperature swings. Not compared to the greenhouse transition we haven't. That's the point - everything our species has seen is variation within one extreme, switching to the other extreme has just as much variation - from jungle to planet-spanning desert, but the distance between the two is sufficient that normal variations are insufficient to "toggle" the global climate.
A greenhouse world *might* be more hospitable, assuming we could avoid the global deserts, but the transition is going to be murder - literally. Adapting over the course of centuries is one thing - adapting from year to year because the weather is increasingly unpredictable, not so much. As for taking thousands of years to melt the ice caps - no. Just watch how much the arctic cap melted last year: https://www.nasa.gov/feature/g...
That depends entirely on the implementation details. There's lots of things we could put in place that would disincentivize fossil fuel use, without necessarily hampering the overall economy. For example, tax all fossil fuels increasingly heavily, and then redistribute the income equally to the population. Anybody with an average sized carbon footprint breaks even without changing their purchasing habits, most people are below average so come out ahead, and the rich can afford to pay more - they're already getting a steadily increasing percentage of the wealth.
That incentivizes consumers to reduce their carbon footprint, and the entire supply chain to shift to non-fossil fuels at every level, as it would give them an immediate price advantage over their competitors. Profit margins would shrink accordingly, but absolute profit would remain unchanged.
As I recall, the half-life of CO2 in the atmosphere is about 50 years, cease production today, and in about 50 years the elevated levels due to our contribution would be halved. Assuming we haven't crossed any tipping points.
>There is zero evidence that we have reached a tipping point. Well, not zero, but we really don't understand the system well enough to make that call, so yes, there's a good chance we haven't yet crossed any tipping point.
Which is exactly why it's important to curtail emissions as fast as possible. Because pretty much every expert agrees that the tipping point will be somewhere between 400 and 800ppm. We're already between the error bars, and with current extensive fossil fuel subsidies (wars, tax breaks, exemption from environmental regulations, etc.) we will quite likely peak somewhere beyond 800ppm. And once we hit the tipping point, then we'll have *major* expensive adaptations to deal with for the next several centuries, enough to dwarf the expense accelerating the transition now.
>There are no "geo-historical norms" Certainly there are - as you say, the planet has been cycling back and forth between ice-house and greenhouse states, and the transition in either direction has generally taken several thousand years (and still caused massive extinctions). We're currently on track to make the transition to a greenhouse world occur much faster than normal, possibly within only several centuries.
There's also historical temperature norms - all of human history has occurred within one icehouse period, our species has never seen a greenhouse world. There's been lots of fluctuations within that icehouse period, with glaciers advancing and receding, but none compare to the changes that accompany a transition to a greenhouse world.
As for >But the necessary government interventions would cast the world into a massive global economic depression and poverty That's just speculation - economics is worse than psychology when it comes to being riddled with unsupported pet "theories" and being unable to predict outcomes.
Actually, it might be very difficult. Introducing flaws is easy, but each manufacturing technique has it's own characteristic flaws, and the flawless techniques are mostly pretty finicky. Creating natural looking flaws might well prove an art form in itself. A rather stupid one at that.
>But we can make excellent predictions about the effects of economic development. No, we really, really can't. Economics, especially macroeconomics, makes climate science look positively rock solid and thoroughly-understood in comparison. There is precious little science there, it's all speculation and philosophy.
>Mass starvation is not required to lower population levels You're absolutely right. Mass warfare or mass-plagues will also do the trick. Climate change almost certainly isn't going to be slow enough for us to implement a global one-child-per-family policy and let the population fall naturally.
Global population is currently over 7 billion, with most projections expecting it to peak at 9-11 billion somewhere around mid-century. Meanwhile the sustainable carrying capacity of the planet's ecosystem is estimated at about 4-5 billion and falling (as we "spend the capital" and reduce biomass and ecosystem health) And that's *before* considering climate change, which will further reduce the carrying capacity during the multi-century transition, as well as greatly reducing agricultural effectiveness due to much less predictable weather.
We *may* be - that's one of the areas we don't know enough detail to be sure of. At present, we're still the primary source of increasing atmospheric CO2, and if we stopped all fossil fuel use tomorrow it might still be very possible to reverse the process. A tall order, but cheap and easy compared to trying to reverse the process once our total contributions become only a fraction of the increasing CO2 levels.
We're still very much the forcing factor, and it's not at all clear that the planet has crossed the tipping point where, if we stopped forcing, it would return to its historical stable state. I'll grant you though that we're quite possibly past the point where there's a realistic chance of cutting our emissions fast enough to not cross the tipping point. Our choice may now be down to slowing the transition to geo-historical norms, so that we and the planetary ecosystem have a decent chance to adapt.
In-use transistors? Do you have a source for that? I would be quite intrigued by the details.
I do recall hearing that that was a major contributing factor to many multicore CPU tiers being available - e.g. if two cores in an 8-core CPU don't pass inspection, then those two cores get disabled and the resulting CPU is sold as a lower-tier 6-core product. The individual core is still useless, but the remainder of the CPU can still be sold. That's why I specified cores, and not CPUs.
I could even see finer-grained versions of such a thing - e.g. a "full featured core" might have 4 integer adders, so if one malfunctioned it could be disabled and marketed as a 3-adder core. But that seems like it would quickly present such a non-homogeneous computing platform that compiler optimization would be hard pressed to deal with the variation, resulting in either abysmal mis-optimized performance on the flawed cores, or substantial unused performance potential on full-featured cores.
I suppose the "unused potential" option could still be done if yields were bad enough - e.g. make every core with 5 adders under the assumption that one would probably fail, and then only discard those where more than one failed, and disable the perfectly functional "extra" adder on those where it didn't fail. It seems like that could quite rapidly get excessive though: you'd have to build in enough redundancy so that a failure in any one functional unit could be replaced - easily adding a double-digit percentage cost increase to the production of every CPU. And of course, the largest most expensive functional units are also going to be the most likely to contain a malfunctioning transistor though perhaps, if such units comprise a large enough percentage of the total core cost, you simply don't put in redundancy for "major functions", and just treat any failure in them as a whole-core failure
I did say "relatively" for a reason. I was even thinking of less ornamental applications - numerous papers have detailed techniques by which computations from unreliable CPUs could still be useful. Just, apparently not useful enough to be worth implementing in the face of easy access to non-defective CPUs (unless there's actually a thriving market for CPU manufacturing "seconds" which I've just never heard of)
The extent of the problem is vague, because we're dealing with systems we don't understand well enough to make detailed predictions. However, the one thing pretty much all our models agree on, and that we're already seeing, is that global weather will become more much unpredictable as global warming increases. We're already seeing much extended droughts and flooding as uneven global warming weakens the equatorial-polar thermal engine that kept the jet streams relatively predictable, allowing them to instead meander and trap weather formations over certain areas (polar vortices extending down into the U.S., etc).
And I addressed your point directly. Yes, famine is caused by an imbalance between population size and agricultural productivity. And thus a sustained drop in agricultural productivity, which is looking very likely, directly translates to famine until the population size drops to newly sustainable levels.
Fair enough - I didn't intend to use the the term "runaway" in the context of a larger specific term, but rather in the usual context of "process that, once you've started, is beyond your ability to stop". Positive feedback alone doesn't convey the fact that there's a very real danger of pushing the global climate balance beyond the point where anything we can reasonably do will have a noticeable effect. (barring large-scale geoengineering - a proposition whose potential side effects may easily turn out to be a cure worse than the disease)
Not in the long term certainly, but they can potentially use the massive profits of their natural diamond monopoly to subsidize the prices of their artificial diamonds below the cost of production, and drive the competition out of business, at which point they can abandon artificial diamonds. Or at least inflate the price enough that they're not a serious threat to their core business, while maintaining production capacity as an ongoing threat to keep any new competitors out of the market. It's hardly an uncommon strategy.
>But there is no difference between a mined and a lab made diamond
Sure there is - the lab grown diamond is potentially far, far purer, more flawless, and larger, than anything you'll find in nature. (though not all production techniques can achieve all those benefits)
It's also far younger as a cohesive unit, though if the age of the stones in your jewelry is what's important to you, finely polished fossilized wood is potentially older and has a lot more character.
Why not? DeBeers already did it with diamonds themselves. What a racket, take control of the worlds supply of extra hard-and-sparkly gravel so that you can artificially restrict the supply, and then convince people all over the world that it's extremely valuable.
I believe those impurities are introduced *after* the diamond has been produced. As with silicon, first you need to grow a pure, non-conductive crystal, then slice it into finely polished wafers, and then introduce "n" and "p" impurity dopings at very precisely controlled locations to create transistors,etc. If there's any impurities in the original crystal you're sunk before you've even begun. After all, if even one of the (100s of) millions of transistors in a CPU core malfunctions, the whole core is relatively much useless.
Which is why nobody has bothered trying to make whole CPUs out of natural diamond (along with the ridiculously overinflated prices and generally tiny size) - the finest natural diamond is riddled through with defects and impurities that would totally destroy the results - you need lab-grade, truly perfect diamonds, flawless down to the atomic level, to make such things possible.
Actually it's very dependent on the manufacturing process - there's lots of ways to make synthetic diamonds, most of which introduce lots of impurities and/or lattice imperfections. It's only in the last few decades that we've worked out how to make flawless carbon-only diamonds, and those processes are mostly both more expensive (at least for now) and at least some are slower to scale due to physical limits of a cumulative process.
Slashdot Mirror
Let's say you're right, and there's a 50/50 chance of needing a follow up conversation. Great! That means you've avoided an intrusive time-wasting call 50% of the time, by sending a potentially standard pre-typed reply.
Granted, a more informative initial text would be good - but knowing clients/users/etc, a longer text is unlikely to actually be helpful, and even in a conversation it's likely to be a chore to extract useful information.
If nothing else their text is an invitation for you to call them, at *your* convenience. They're initiating communication in the least-intrusive way possible, and if it's important to them they'll likely answer your call.
Make it a federal law, and you don't have jurisdiction issues within the US. And nothing stops us from putting requirements on international calls - if a foreign phone company wants to be able to connect to US callers, they need to comply with US international-call regulations.
You missed the part about redistributing the tax income didn't you?
To use your example I didn't suggest that the government demands gold-plated plumbing - I suggested they put a $70,000 tax on all non-gold-plated plumbing, and then give that $70,000 to you. It's totally up to you whether you want to spend $75k on non-gold-plated plumbing, or spend $50k on gold-plated plumbing and the other $25k on other things.
Namely the part where it's completely false.
Even logarithmic functions increase to infinity. Increase the temperature, you increase the CO2 levels, which further increases the temperature, which further increases the CO2 levels. A logarithmic increase will help that to eventually stabilize, but says nothing about *where* it stabilizes - that depends entirely on the strength of the positive feedback - and thawing permafrost, methane hydrates, etc. deliver extremely large positive feedback over a narrow temperature range - once the process begins in earnest, it's going to continue until the reserves are depleted. If we reach the point where the arctic is venting sequestered carbon at 10x the rate of human emissions, do you really think us stopping our emissions is going to stop the process? No - we will be beyond the tipping point.
And why don't my numbers make sense? Most ecologically sequestered carbon never become fossil fuels, it becomes more stable forms of rock. It takes very special circumstances to convert biomass into chemically unstable fossil fuels.
Recent studies (2009) estimate that northern circumpolar permafrost soil carbon content equals approximately 1672 Pg. (1 Pg = 1 Gt = 10^15g) This estimation of the amount of carbon stored in permafrost soils is more than double the amount currently in the atmosphere. ( https://en.wikipedia.org/wiki/... )
That's about 57x the roughly 29Gt amount we emit in a year - and worse, a whole lot of that is going to be initially released as methane, which is about 30x more potent a greenhouse gas than CO2, before eventually breaking down into CO2. And that's just circumpolar permafrost.
Meanwhile undersea methane hydrate reserves are estimated at somewhere between 100 and 530,000Gt (sea floor exploration is woefully spotty) 100 might not be that bad - under 4 years worth of human carbon emissions, though that 30x warming multiplier for methane won't help anything. 530,000 though, that's 18,000 years worrth of current human emissions, before the methane multiplier. We'd better hope the reality is a lot closer to the low-end estimates.
And no, study your paleontology a bit - there is NO guarantee a greenhouse Earth will be wetter - they also include some of the most severe global desert phases of our planet's history. Nor milder - more energy in the atmosphere, plus much weaker and more unstable/unpredictable jet streams sets the stage for far more extreme weather variation such as we're already seeing. There are no guarantees of anything. A Greenhouse Earth can be every bit as inhospitable as a glacier-covered Icehouse Earth.
I reduce the problem by having a different ringtone for unknown callers, and a voice message that tells people I screen unknown callers and please leave a message. If I get a message I know it immediately and can check, and it's rarely spam (political robocalls notwithstanding). Perhaps I've missed out on an old flame calling to reunite and chickening out at the message, but I doubt it. And when my friends change their number, or call from another phone, they generally still reach me within minutes of placing the call.
If you're going to lie to the client and and answer yes without knowing anything to shut them up, then is it really that likely that you'd be more helpful otherwise?
Here's a thought: Don't lie and answer "Yes" when you don't actually know anything - assuming you actually want to provide service to your client.
If you don't know, answer something like "No. What's the issue and how big a problem is it for you?" Maybe even add a "Please call if it's serious."
The beauty of texting, in addition to being asynchronous so that it doesn't present a near-guaranteed interruption of your work flow, is that it can be very brief and minimally intrusive to both parties. Something has gone wrong at MZR - maybe it's causing lots of problems and you're in the middle of dealing with it. Do you really want to ALSO be dealing with all your clients calling in complaining about the MZR issue and asking for updates? Or would you rather get quick texts that you can occasionally stop to send a batch of quick "Yes, I'm working on it now" replies?
Phone calls are great for serious information exchange or when an immediate response is needed - but are complete intrusive overkill most of the time.
It would be easy enough if those in power actually wished to engorce them, rather than just passing symbolic legislation to make themselves look good without actually changing anything.
The phone company knows exactly who's calling you - they're charging them for the call. All they'd need to do is let you, e.g. dial *FU after a spam call to quickly report them. Allow a little leeway for false reports, but anyone who racks up a lot of reports gets prosecuted.
Increased temperature increases the rate of natural greenhouse gas stores - there's far more ecologically sequestered carbon and methane than humans would use in centuries - if our emissions warm the planet sufficiently, those reserves start being released faster than anything we've done. Same thing appears to have happened with every icehouse to greenhouse transition. Global climate is a bistable system - push it far enough and you hit the tipping point where the process becomes self-accelerating until it reaches the opposite state.
>And during that time, we have experienced massive climate change and temperature swings.
Not compared to the greenhouse transition we haven't. That's the point - everything our species has seen is variation within one extreme, switching to the other extreme has just as much variation - from jungle to planet-spanning desert, but the distance between the two is sufficient that normal variations are insufficient to "toggle" the global climate.
A greenhouse world *might* be more hospitable, assuming we could avoid the global deserts, but the transition is going to be murder - literally. Adapting over the course of centuries is one thing - adapting from year to year because the weather is increasingly unpredictable, not so much. As for taking thousands of years to melt the ice caps - no. Just watch how much the arctic cap melted last year: https://www.nasa.gov/feature/g...
That depends entirely on the implementation details. There's lots of things we could put in place that would disincentivize fossil fuel use, without necessarily hampering the overall economy. For example, tax all fossil fuels increasingly heavily, and then redistribute the income equally to the population. Anybody with an average sized carbon footprint breaks even without changing their purchasing habits, most people are below average so come out ahead, and the rich can afford to pay more - they're already getting a steadily increasing percentage of the wealth.
That incentivizes consumers to reduce their carbon footprint, and the entire supply chain to shift to non-fossil fuels at every level, as it would give them an immediate price advantage over their competitors. Profit margins would shrink accordingly, but absolute profit would remain unchanged.
As I recall, the half-life of CO2 in the atmosphere is about 50 years, cease production today, and in about 50 years the elevated levels due to our contribution would be halved. Assuming we haven't crossed any tipping points.
>There is zero evidence that we have reached a tipping point.
Well, not zero, but we really don't understand the system well enough to make that call, so yes, there's a good chance we haven't yet crossed any tipping point.
Which is exactly why it's important to curtail emissions as fast as possible. Because pretty much every expert agrees that the tipping point will be somewhere between 400 and 800ppm. We're already between the error bars, and with current extensive fossil fuel subsidies (wars, tax breaks, exemption from environmental regulations, etc.) we will quite likely peak somewhere beyond 800ppm. And once we hit the tipping point, then we'll have *major* expensive adaptations to deal with for the next several centuries, enough to dwarf the expense accelerating the transition now.
>There are no "geo-historical norms"
Certainly there are - as you say, the planet has been cycling back and forth between ice-house and greenhouse states, and the transition in either direction has generally taken several thousand years (and still caused massive extinctions). We're currently on track to make the transition to a greenhouse world occur much faster than normal, possibly within only several centuries.
There's also historical temperature norms - all of human history has occurred within one icehouse period, our species has never seen a greenhouse world. There's been lots of fluctuations within that icehouse period, with glaciers advancing and receding, but none compare to the changes that accompany a transition to a greenhouse world.
As for
>But the necessary government interventions would cast the world into a massive global economic depression and poverty
That's just speculation - economics is worse than psychology when it comes to being riddled with unsupported pet "theories" and being unable to predict outcomes.
Actually, it might be very difficult. Introducing flaws is easy, but each manufacturing technique has it's own characteristic flaws, and the flawless techniques are mostly pretty finicky. Creating natural looking flaws might well prove an art form in itself. A rather stupid one at that.
>But we can make excellent predictions about the effects of economic development.
No, we really, really can't. Economics, especially macroeconomics, makes climate science look positively rock solid and thoroughly-understood in comparison. There is precious little science there, it's all speculation and philosophy.
>Mass starvation is not required to lower population levels
You're absolutely right. Mass warfare or mass-plagues will also do the trick. Climate change almost certainly isn't going to be slow enough for us to implement a global one-child-per-family policy and let the population fall naturally.
Global population is currently over 7 billion, with most projections expecting it to peak at 9-11 billion somewhere around mid-century. Meanwhile the sustainable carrying capacity of the planet's ecosystem is estimated at about 4-5 billion and falling (as we "spend the capital" and reduce biomass and ecosystem health) And that's *before* considering climate change, which will further reduce the carrying capacity during the multi-century transition, as well as greatly reducing agricultural effectiveness due to much less predictable weather.
We *may* be - that's one of the areas we don't know enough detail to be sure of. At present, we're still the primary source of increasing atmospheric CO2, and if we stopped all fossil fuel use tomorrow it might still be very possible to reverse the process. A tall order, but cheap and easy compared to trying to reverse the process once our total contributions become only a fraction of the increasing CO2 levels.
We're still very much the forcing factor, and it's not at all clear that the planet has crossed the tipping point where, if we stopped forcing, it would return to its historical stable state. I'll grant you though that we're quite possibly past the point where there's a realistic chance of cutting our emissions fast enough to not cross the tipping point. Our choice may now be down to slowing the transition to geo-historical norms, so that we and the planetary ecosystem have a decent chance to adapt.
Yep. It's been an open secret for at least several decades.
I figured. Should have added HAH! at the beginning of my reply.
In-use transistors? Do you have a source for that? I would be quite intrigued by the details.
I do recall hearing that that was a major contributing factor to many multicore CPU tiers being available - e.g. if two cores in an 8-core CPU don't pass inspection, then those two cores get disabled and the resulting CPU is sold as a lower-tier 6-core product. The individual core is still useless, but the remainder of the CPU can still be sold. That's why I specified cores, and not CPUs.
I could even see finer-grained versions of such a thing - e.g. a "full featured core" might have 4 integer adders, so if one malfunctioned it could be disabled and marketed as a 3-adder core. But that seems like it would quickly present such a non-homogeneous computing platform that compiler optimization would be hard pressed to deal with the variation, resulting in either abysmal mis-optimized performance on the flawed cores, or substantial unused performance potential on full-featured cores.
I suppose the "unused potential" option could still be done if yields were bad enough - e.g. make every core with 5 adders under the assumption that one would probably fail, and then only discard those where more than one failed, and disable the perfectly functional "extra" adder on those where it didn't fail. It seems like that could quite rapidly get excessive though: you'd have to build in enough redundancy so that a failure in any one functional unit could be replaced - easily adding a double-digit percentage cost increase to the production of every CPU. And of course, the largest most expensive functional units are also going to be the most likely to contain a malfunctioning transistor though perhaps, if such units comprise a large enough percentage of the total core cost, you simply don't put in redundancy for "major functions", and just treat any failure in them as a whole-core failure
I did say "relatively" for a reason. I was even thinking of less ornamental applications - numerous papers have detailed techniques by which computations from unreliable CPUs could still be useful. Just, apparently not useful enough to be worth implementing in the face of easy access to non-defective CPUs (unless there's actually a thriving market for CPU manufacturing "seconds" which I've just never heard of)
The extent of the problem is vague, because we're dealing with systems we don't understand well enough to make detailed predictions. However, the one thing pretty much all our models agree on, and that we're already seeing, is that global weather will become more much unpredictable as global warming increases. We're already seeing much extended droughts and flooding as uneven global warming weakens the equatorial-polar thermal engine that kept the jet streams relatively predictable, allowing them to instead meander and trap weather formations over certain areas (polar vortices extending down into the U.S., etc).
And I addressed your point directly. Yes, famine is caused by an imbalance between population size and agricultural productivity. And thus a sustained drop in agricultural productivity, which is looking very likely, directly translates to famine until the population size drops to newly sustainable levels.
Fair enough - I didn't intend to use the the term "runaway" in the context of a larger specific term, but rather in the usual context of "process that, once you've started, is beyond your ability to stop". Positive feedback alone doesn't convey the fact that there's a very real danger of pushing the global climate balance beyond the point where anything we can reasonably do will have a noticeable effect. (barring large-scale geoengineering - a proposition whose potential side effects may easily turn out to be a cure worse than the disease)
I doubt even Switzerland has banks large enough to store those huge mountains of gravel.
Not in the long term certainly, but they can potentially use the massive profits of their natural diamond monopoly to subsidize the prices of their artificial diamonds below the cost of production, and drive the competition out of business, at which point they can abandon artificial diamonds. Or at least inflate the price enough that they're not a serious threat to their core business, while maintaining production capacity as an ongoing threat to keep any new competitors out of the market. It's hardly an uncommon strategy.
>But there is no difference between a mined and a lab made diamond
Sure there is - the lab grown diamond is potentially far, far purer, more flawless, and larger, than anything you'll find in nature. (though not all production techniques can achieve all those benefits)
It's also far younger as a cohesive unit, though if the age of the stones in your jewelry is what's important to you, finely polished fossilized wood is potentially older and has a lot more character.
Why not? DeBeers already did it with diamonds themselves. What a racket, take control of the worlds supply of extra hard-and-sparkly gravel so that you can artificially restrict the supply, and then convince people all over the world that it's extremely valuable.
I believe those impurities are introduced *after* the diamond has been produced. As with silicon, first you need to grow a pure, non-conductive crystal, then slice it into finely polished wafers, and then introduce "n" and "p" impurity dopings at very precisely controlled locations to create transistors,etc. If there's any impurities in the original crystal you're sunk before you've even begun. After all, if even one of the (100s of) millions of transistors in a CPU core malfunctions, the whole core is relatively much useless.
Which is why nobody has bothered trying to make whole CPUs out of natural diamond (along with the ridiculously overinflated prices and generally tiny size) - the finest natural diamond is riddled through with defects and impurities that would totally destroy the results - you need lab-grade, truly perfect diamonds, flawless down to the atomic level, to make such things possible.
Actually it's very dependent on the manufacturing process - there's lots of ways to make synthetic diamonds, most of which introduce lots of impurities and/or lattice imperfections. It's only in the last few decades that we've worked out how to make flawless carbon-only diamonds, and those processes are mostly both more expensive (at least for now) and at least some are slower to scale due to physical limits of a cumulative process.