5 ms·

OK, one person on this thread willing to actually have an intelligent conversation. This is very rare. Thank you. I take your point about these papers being

OK, one person on this thread willing to actually have an intelligent conversation. This is very rare. Thank you. I take your point about these papers being dated. The sad reality is that no researcher wants to touch this with a ten foot pole. Going against the political grain can be career ending, and they know it. If we do fund research to debunk some of these lofty projections, it is very, very rare. Let's steel-man your argument: Energy becomes so cheap that it makes some of the old ideas possible. OK. I don't know of a single proposal that has been shown to scale outside of a laboratory environment. In other words, if we were to say: We cannot deploy this today, yet, in 50 years, with cheaper energy, will will be able to. We ignore energy issues and show it working at a reasonable scale. The world then devotes the next N years to prepare for the deployment of such a solution. That would be sensible. And yet, that has not happened. We cannot construct an argument based on an imagined solution. And then there's the other "minor" problem: Solar is a horrible source of energy when you consider the entire life cycle. The procurement of materials, mining, transportation, manufacturing, shipping, installation and operation are not without massive ecological consequences, including CO2 production. Do anything at a massive scale --at a planetary scale-- and the consequences will track with the scale. And yet, it does not end there. I am not even going to explore the issues with energy storage technology at a planetary scale. From an ecological perspective, that is definitely unimaginable. Let's just agree that this is something significant enough not to ignore. Yet, onece again, that isn't the entire story. Solar technology is seriously problematic in one key metric: Operating life. A solar panel is only good for 15 to 25 years. The electronics, if lucky, could be on a similar time scale. RoHS introduced a degree of planned obsolescence in every single electronic device on the planet. Lovely. Talk about unintended consequences. The couple-decade lifespan of solar systems means that, if we were to install solar at the required planetary scale, the entire planet would have to rebuild the earths entire solar power generation infrastructure every twenty years or so. That cannot be ignored. The word "cataclysmic" comes to mind. Imagine China constantly producing replacement panels and systems for a planetary scale solar array rebuilding from here to eternity. We can't even begin to calculate just how much ecological damage this might cause. It won't be zero. It will not be trivial. It will be at a scale never before imagined. And that's the problem. It is easy to say "solar and wind" and point to a solution in isolation of the realities of the idea. It's like cleaning your floors with muddy water. The exercise does not result in cleaner floors.

kragen 2y ago
> OK, one person on this thread willing to actually have an intelligent conversation. This is very rare. Thank you. It's always pleasant to discuss issues with you, too! It's unfortunate that there's so much ideological battle on here that it's rare, in part because the structural incentives are for rapid replies rather than thoughtful ones. (The tone of some of your comments in this thread has perhaps been somewhat inflammatory, as well.) I forget, are you on the think.com hackers-l? > I don't know of a single proposal that has been shown to scale outside of a laboratory environment. Well, elsethread, I pointed at https://cobblab.eas.gatech.edu/energy/Readings/rochelle2009.pdf https://cobblab.eas.gatech.edu/energy/Readings/rochelle2009...., which mentions dozens of power plants that are using amine scrubbing at sub-gigawatt scales for flue gas decarbonization. While this is still three orders of magnitude lower than required, I hope you'll agree that it's outside of a laboratory environment! It's reasonable to wonder whether current sources of the required amines can scale to those levels. I suspect that with insanely cheap energy it will be more profitable to use cheaper, less energy-efficient materials such as lime. Hopefully it is clear that lime burning faces no obstacles in scaling to the necessary level to terraform Earth, but if you have any doubts, I'd be happy to answer them. I suspect we'll probably come up with something better than lime kilns; the reason that's what I outlined upthread is mostly because it's so simple and well understood, without any unknowns that might come back to bite us. > We cannot construct an argument based on an imagined solution. Well, as I understand it, the alternatives would be to construct an argument based on an already deployed solution, which unfortunately doesn't exist, or to construct an argument not based on any solutions, which won't get us anywhere. So constructing an argument based on an imagined solution is probably our best choice. While that approach didn't work so well at bringing about world peace, urban renewal, or Xanadu, its track record is pretty decent at building things like the Apollo program, the railroads, and the internet, so I don't think we should dismiss constructing an argument based on an imagined solution. We just have to be careful and rigorous in our calculations! > Solar is a horrible source of energy when you consider the entire life cycle. The procurement of materials, mining, transportation, manufacturing, shipping, installation and operation are not without massive ecological consequences. This turns out to be false, which should be clear from a simple calculation, almost a Fermi estimate. A square meter of 21%-efficient solar panel (the most common kind today) is 210 watts peak, 40 watts average at a 20% capacity factor. It's mostly (>90%) made of 3.2-mm-thick soda-lime glass, which is 2.44g/cc, so it's 7.8kg/m². If we take its rated life of 20 years, it produces 25GJ, which works out to 3200MJ/kg. We know that the procurement of materials, mining, transportation, and shipping to send coal to coal power plants is economically feasible—significant compared to the economic value of the coal as energy, but far from overwhelming, so many coal power plants are located far from coal mines. Coal's energy density tops out at 33MJ/kg, conveniently 1% of the "energy density" of the solar panel, and we know that the majority of the massive ecological consequences of coal mining come from burning the coal, with a small but significant addition from strip mining. So, even for coal, procurement, and transportation/shipping are insignificant, and for solar panels they're roughly a hundred times smaller even than that. Mining the requisite materials similarly makes an impact that's only roughly a hundredth of the already small environmental impact of coal mining. (The only rare material used in solar panels is silver, which contributes about 10% of their cost. Everything else is very abundant, far more abundant than coal, and therefore doesn't have to be refined from low-grade deposits.) The ratio is actually close to 300:1 than the 100:1 I've been using, because the solar panels directly produce electricity, while about two thirds of the coal's energy is lost in the thermal power plant. Manufacturing, installation, and operation of solar panels can of course be carried out in arbitrarily harmful ways, but they don't carry any inherent environmental costs the way mining and shipping do, and overall they seem to be doing little or no environmental damage at the moment. Operating solar farms may actually be environmentally beneficial in desert and desertifying areas, although we should expect that to eventually be misaligned with some kind of economic objective. > I am not even going to explore the issues with energy storage technology at a planetary scale. That's okay! I've been exploring them for several years, and it turns out that, while the energy transition to renewables does create a planetary-scale energy storage problem, there are numerous viable planetary-scale solutions. So there's not really anything to worry about. > A solar panel is only good for 15 to 25 years. It turns out that this is only the rated life, and crystalline silicon solar panels (the kind that has virtually all of the current utility-scale market) have actually degraded by only about 15% at 25 years, unless they're broken (for example due to manufacturing defects or vandalism). They do continue to degrade thereafter, but only by a fraction of a percent per year. Crystalline silicon is the silicon analogue of diamond (but without diamond's metastability problem), and its surface oxide layer has a very low diffusivity for oxygen, so the relevant timescale for the solar cells themselves is probably millennia, not years. (The EVA adhesive would probably have to be replaced every century, though.) Most panels manufactured in the 01970s are still operational today and still produce most of their original power. > RoHS introduced a degree of planned obsolescence in every single electronic device on the planet. Lovely. Talk about unintended consequences. It turns out that there are exemptions from RoHS, and in particular solar cells are soldered with solder containing lead, avoiding the problem you're describing. Plausibly the power electronics that are connnected to the cells will eventually short out with tin whiskers, but even in RoHS-soldered devices, tin whiskers are not inevitable, and even when they occur in the power electronics, they generally don't damage the solar cells themselves, though they can. > The couple-decade lifespan of solar systems means that, if we were to install solar at the required planetary scale, the entire planet would have to rebuild the earth's entire solar power generation infrastructure every twenty years or so. Remember that the amount of material in solar panels is about 1/300 of the amount of coal that they replace over their lifetime. So if we did replace the whole solar power generation infrastructure every twenty years, the amount we'd have to replace is roughly equivalent to the amount of coal we'd have to dig up every month to supply the same amount of power through coal plants. Which is pretty much what we've been doing. However, this is a place where I think you're underestimating the impact. Right now installed solar power is increasing by about 25% per year, which is to say, every three years, we've been rebuilding the earth's entire solar power generation infrastructure! We just haven't been ripping the old stuff out. You could imagine (as many politicians and researchers do!) that once we have enough solar capacity installed to supply all of world marketed energy consumption (100 TW peak, costing about 11 trillion dollars at today's prices, to supply 20 TW) we will suddenly stop manufacturing solar panels except for those needed to replace out-of-warranty panels. At the current exponential rate, that should happen sometime around 02042, shortly after the 32-bit time_t rollover. But it seems implausible that a solar panel manufacturing industry manufacturing 25TW per year of solar panels (an amount worth US$3 trillion per year at today's prices, but will more likely only be worth on the order of US$200B/year at that time) will be willing to just roll over and die instead of looking for new markets. We should expect world marketed energy consumption to dramatically increase in the years leading up to that time, spurred by energy costs of a tenth or a hundredth of the lowest prices fossil fuels could ever deliver. That's when it gets, to use your word, cataclysmic. We're talking about deploying 100'000 km² of solar panels per year at that point, roughly the size of the UK, Italy, Ecuador, or (the land area of) the Philippines. That's still only 1/5000 of the planet per year, and much of it will surely be offshore like the new Shandong solar power plant, but it's sure to start to cause environmental damage as people start clear-cutting natural trees to make space for their artificial "trees". And that's why I think we should start deploying solar-orbit power satellites well before that point, to start transitioning the most energy-intensive industrial processes off of Earth. O'Neill cylinder habitats made from asteroids can provide abundant living space for extraterrestrial humans. That way, we can preserve its biosphere as a park and a workshop for those who want to continue practicing traditional human crafts such as farming, blacksmithing, CNC machining, and cooking. We probably can't keep the humans from practicing their less laudable traditional crafts, such as warfare and genocide, but we can at least prevent them from annihilating the biosphere.

kragen 2y ago

> OK, one person on this thread willing to actually have an intelligent conversation. This is very rare. Thank you. It's always pleasant to discuss issues with you, too! It's unfortunate that there's so much ideological battle on here that it's rare, in part because the structural incentives are for rapid replies rather than thoughtful ones. (The tone of some of your comments in this thread has perhaps been somewhat inflammatory, as well.) I forget, are you on the think.com hackers-l? > I don't know of a single proposal that has been shown to scale outside of a laboratory environment. Well, elsethread, I pointed at https://cobblab.eas.gatech.edu/energy/Readings/rochelle2009.pdf https://cobblab.eas.gatech.edu/energy/Readings/rochelle2009...., which mentions dozens of power plants that are using amine scrubbing at sub-gigawatt scales for flue gas decarbonization. While this is still three orders of magnitude lower than required, I hope you'll agree that it's outside of a laboratory environment! It's reasonable to wonder whether current sources of the required amines can scale to those levels. I suspect that with insanely cheap energy it will be more profitable to use cheaper, less energy-efficient materials such as lime. Hopefully it is clear that lime burning faces no obstacles in scaling to the necessary level to terraform Earth, but if you have any doubts, I'd be happy to answer them. I suspect we'll probably come up with something better than lime kilns; the reason that's what I outlined upthread is mostly because it's so simple and well understood, without any unknowns that might come back to bite us. > We cannot construct an argument based on an imagined solution. Well, as I understand it, the alternatives would be to construct an argument based on an already deployed solution, which unfortunately doesn't exist, or to construct an argument not based on any solutions, which won't get us anywhere. So constructing an argument based on an imagined solution is probably our best choice. While that approach didn't work so well at bringing about world peace, urban renewal, or Xanadu, its track record is pretty decent at building things like the Apollo program, the railroads, and the internet, so I don't think we should dismiss constructing an argument based on an imagined solution. We just have to be careful and rigorous in our calculations! > Solar is a horrible source of energy when you consider the entire life cycle. The procurement of materials, mining, transportation, manufacturing, shipping, installation and operation are not without massive ecological consequences. This turns out to be false, which should be clear from a simple calculation, almost a Fermi estimate. A square meter of 21%-efficient solar panel (the most common kind today) is 210 watts peak, 40 watts average at a 20% capacity factor. It's mostly (>90%) made of 3.2-mm-thick soda-lime glass, which is 2.44g/cc, so it's 7.8kg/m². If we take its rated life of 20 years, it produces 25GJ, which works out to 3200MJ/kg. We know that the procurement of materials, mining, transportation, and shipping to send coal to coal power plants is economically feasible—significant compared to the economic value of the coal as energy, but far from overwhelming, so many coal power plants are located far from coal mines. Coal's energy density tops out at 33MJ/kg, conveniently 1% of the "energy density" of the solar panel, and we know that the majority of the massive ecological consequences of coal mining come from burning the coal, with a small but significant addition from strip mining. So, even for coal, procurement, and transportation/shipping are insignificant, and for solar panels they're roughly a hundred times smaller even than that. Mining the requisite materials similarly makes an impact that's only roughly a hundredth of the already small environmental impact of coal mining. (The only rare material used in solar panels is silver, which contributes about 10% of their cost. Everything else is very abundant, far more abundant than coal, and therefore doesn't have to be refined from low-grade deposits.) The ratio is actually close to 300:1 than the 100:1 I've been using, because the solar panels directly produce electricity, while about two thirds of the coal's energy is lost in the thermal power plant. Manufacturing, installation, and operation of solar panels can of course be carried out in arbitrarily harmful ways, but they don't carry any inherent environmental costs the way mining and shipping do, and overall they seem to be doing little or no environmental damage at the moment. Operating solar farms may actually be environmentally beneficial in desert and desertifying areas, although we should expect that to eventually be misaligned with some kind of economic objective. > I am not even going to explore the issues with energy storage technology at a planetary scale. That's okay! I've been exploring them for several years, and it turns out that, while the energy transition to renewables does create a planetary-scale energy storage problem, there are numerous viable planetary-scale solutions. So there's not really anything to worry about. > A solar panel is only good for 15 to 25 years. It turns out that this is only the rated life, and crystalline silicon solar panels (the kind that has virtually all of the current utility-scale market) have actually degraded by only about 15% at 25 years, unless they're broken (for example due to manufacturing defects or vandalism). They do continue to degrade thereafter, but only by a fraction of a percent per year. Crystalline silicon is the silicon analogue of diamond (but without diamond's metastability problem), and its surface oxide layer has a very low diffusivity for oxygen, so the relevant timescale for the solar cells themselves is probably millennia, not years. (The EVA adhesive would probably have to be replaced every century, though.) Most panels manufactured in the 01970s are still operational today and still produce most of their original power. > RoHS introduced a degree of planned obsolescence in every single electronic device on the planet. Lovely. Talk about unintended consequences. It turns out that there are exemptions from RoHS, and in particular solar cells are soldered with solder containing lead, avoiding the problem you're describing. Plausibly the power electronics that are connnected to the cells will eventually short out with tin whiskers, but even in RoHS-soldered devices, tin whiskers are not inevitable, and even when they occur in the power electronics, they generally don't damage the solar cells themselves, though they can. > The couple-decade lifespan of solar systems means that, if we were to install solar at the required planetary scale, the entire planet would have to rebuild the earth's entire solar power generation infrastructure every twenty years or so. Remember that the amount of material in solar panels is about 1/300 of the amount of coal that they replace over their lifetime. So if we did replace the whole solar power generation infrastructure every twenty years, the amount we'd have to replace is roughly equivalent to the amount of coal we'd have to dig up every month to supply the same amount of power through coal plants. Which is pretty much what we've been doing. However, this is a place where I think you're underestimating the impact. Right now installed solar power is increasing by about 25% per year, which is to say, every three years, we've been rebuilding the earth's entire solar power generation infrastructure! We just haven't been ripping the old stuff out. You could imagine (as many politicians and researchers do!) that once we have enough solar capacity installed to supply all of world marketed energy consumption (100 TW peak, costing about 11 trillion dollars at today's prices, to supply 20 TW) we will suddenly stop manufacturing solar panels except for those needed to replace out-of-warranty panels. At the current exponential rate, that should happen sometime around 02042, shortly after the 32-bit time_t rollover. But it seems implausible that a solar panel manufacturing industry manufacturing 25TW per year of solar panels (an amount worth US$3 trillion per year at today's prices, but will more likely only be worth on the order of US$200B/year at that time) will be willing to just roll over and die instead of looking for new markets. We should expect world marketed energy consumption to dramatically increase in the years leading up to that time, spurred by energy costs of a tenth or a hundredth of the lowest prices fossil fuels could ever deliver. That's when it gets, to use your word, cataclysmic. We're talking about deploying 100'000 km² of solar panels per year at that point, roughly the size of the UK, Italy, Ecuador, or (the land area of) the Philippines. That's still only 1/5000 of the planet per year, and much of it will surely be offshore like the new Shandong solar power plant, but it's sure to start to cause environmental damage as people start clear-cutting natural trees to make space for their artificial "trees". And that's why I think we should start deploying solar-orbit power satellites well before that point, to start transitioning the most energy-intensive industrial processes off of Earth. O'Neill cylinder habitats made from asteroids can provide abundant living space for extraterrestrial humans. That way, we can preserve its biosphere as a park and a workshop for those who want to continue practicing traditional human crafts such as farming, blacksmithing, CNC machining, and cooking. We probably can't keep the humans from practicing their less laudable traditional crafts, such as warfare and genocide, but we can at least prevent them from annihilating the biosphere.