I’ve noticed that recommendations for action based on a vision of the future are based on an idea that something must “eventually” occur. For example, eventually: We will run out of coal, so we’d better find replacements soon.
How many electronic switches are required to match the computational power of a neuron? We don't even have the resources to emulate a rat brain.
> 2) Electrical memory is 100 percent accurate, where human memory isn't even close.
How many terabytes would be required to keep a 100% accurate record of a lifetime of human experiences?
> 3) Computers can communicate with one another much more quickly than humans.
But still have substantial communication delays which are ultimately bounded by the speed of light. You will not get a real-time intelligent machine by wiring together one billion dumb computers scattered around the planet.
> 4) There is reason to believe humans can design massively parallel computers.
And, given equivalent computational resources, parallel computers perform worse than sequential computers. People are turning to parallelization because we can't make our computers run faster.
>5) Computers don't sleep. They only need electricity to eat.
Electricity doesn't exactly grow on trees. Currently, electricity expenditure is becoming a larger and larger fraction of the costs of companies that operate large clusters. Soon it may become the limiting factor of the ICT economy.Computers also need plastic, copper, aluminium, silicon and rare elements.
Human-level AI may be possible. But taking over the world, not so much.
"Hence it is in their interest to avoid conflict."
In the machines' interest? I don't see that as true...any more than it is in our interest to avoid conflict with ants. (We are the ants in this analogy.)
"IIUC, natural gas is much more scarce than coal, hence, even with the higher efficiency of its power stations (50% vs 38% of coal-fired power stations) it will not replace coal."
Jesse Ausubel thinks natural gas will replace coal.
Decarbonization: the next 100 years
I agree. I can install a fuel cell that operates on natural gas in my crawl space. (Or I will be able to do it in a decade or two, as prices come down.) Then I get both electricity and hot water from my natural gas.
In contrast, no way would I want coal in my crawl space, or even anywhere near my neighborhood.
Regarding photovoltaics, the EROEI is higher than one. That's basically all that matters, since the energy produced is electricity, which is very high quality.
"We won't dig up the last piece of coal, but eventually it will become uneconomic to keep using it. At that point, for all practicall purposes, we will have been run out of coal."
In the United States, natural gas is already beating coal on economics...especially considering the recent requirements to add much more air pollution control (for mercury, nitrogen oxides, sulfur dioxide, and particulate).
Plus, no one wants a coal mine in their yard. But lots of people have accepted hydrofracking natural gas wells in their yard
IIUC, natural gas is much more scarce than coal, hence, even with the higher efficiency of its power stations (50% vs 38% of coal-fired power stations) it will not replace coal.
Photovoltaic currently has an EROEI lower than coal. While technological developments might increase it, we don't expect large improvements.
As coal becomes more scarce, its EROEI will decrease, hence at some point other energy sources (photovoltaics, uranium breeders and thorium breeders) will become competitive. We won't dig up the last piece of coal, but eventually it will become uneconomic to keep using it. At that point, for all practicall purposes, we will have been run out of coal.
Do you mean that you think it’s possible that there is an unlimited amout of coal?Or that our yearly demand of coal falls to zero?I suppose the latter is possible, but how it would practically differ from having run out of coal?
As has been often noted, the Stone Age didn't end when we ran out of stone, and the Bronze Age didn't end when we ran out of bronze. The Coal Age is almost certainly going to end in this century. But not because there won't be any coal left in the ground. Instead, it will be because natural gas and photovoltaics are better sources of electricity.
Even when we are relatively certain that a cost now will result in benefits in the future (for example: using less gas now will mean there will be more gas to use in the future. duh)
Well...ummm...I don't agree. There may be less gas made from *fossil* fuels in the future. But there's no reason why using more gas now will result in appreciably less renewable gas in the future. For example, diesel can be made from algae. That essentially provides a never-ending source of diesel. (And high-quality, low-sulfur diesel, at that.)
If modern history teaches us anything, it’s that when the rich and powerful rule on the poor and disenfranchised, but when the poor and disenfranchised rebel, the rich and powerful usually end up dead, hence it is interest to avoid conflict.
Oh, I forgot one more basic principle why computers will take over the world: it takes humans ~9 months to grow a new human, and about 18 years to train it. It constrast, terminators can be built at prodigious rates, and are essentially ready to start killing right off the assembly line.
You're right about the historical "rich and powerful". But that's merely because, historically, the "rich and powerful" have been vastly outnumbered by the poor and desperate.
In contrast, the rich and powerful machines can easily outnumber us. That's because they need only for electricity to eat, need no clothes to wear, and need little or no shelter.
But I, for one, will welcome our new machine overlords.
We dumped our trash in the street, left our dead horses in the streets, left the horse shit in the streets because the streets were wide and our waste so small.
Plus it was a lot cheaper that hauling it out of town.
I think I heard on the radio, government dictates mandating hauling waste a mile out of town were first issued in ancient Greece.
The fossil fuel industry are still pretty much in the dumping waste in the streets because they are wide and its cheap stage.
And I guess you believe no one predicted demand increase outstripping supply increase in the 90s which resulted in the sharp run up in oil prices in the 00s rather than the price holding steady at the 1990 oil price. Shouldn't producers been holding back oil in the US in the 90s so in the 00s oil production would have increased instead of falling even as regulation was relaxed and prices skyrocketed? Why did production in the US increase as prices fell and regulation increased?
Can you explain how you deduce from established principles that AI will eventually take over the world?
First principles:
1) Electronics switch faster than neurons.
2) Electrical memory is 100 percent accurate, where human memory isn't even close.
3) Computers can communicate with one another much more quickly than humans.
4) There is reason to believe humans can design massively parallel computers.
5) Computers don't sleep. They only need electricity to eat. They can't be bargained with. They can't be reasoned with. They don't feel pity, or remorse, or fear. And they absolutely will not stop, ever, until we are dead.
It doesn’t matter what “people” do. What matters is what the people with money do.
But this is a distinction without a difference, since virtually anybody has money (or even without money partecipates in economic transactions of some sort). There is nothing that compels Average Joe to drive a car instead of riding a bicycle and to eat meat and fish instead of soy. A few billions Average Joes driving their cars and eating meat results in resource depletion and global warming.
A few billions Average Mohammeds and Average Manuels fathering six children each because they are too poor to accumulate assets for retirement and their countries offer no social security also contribute to the problem, but not as much as the Joes do.
Those that behave irrationally tend make poor investments or overconsume, and as such, their decisions have a smaller effect on the market, while those with the foresight to to make sound predictions have a much larger influence.
That why massive speculative bubbles that regularly inflate and burst don't exist, do they?
Belief that professional investors are essentially rational is an empirical falsehood. And even if they were rational, their ability to predict the markets would be limited by the intrinsic chaoticity of the system.
But, for the sake of the argument, let's assume that all economic agents, not just investors, everybody, are perfectionally rational and perfectly able to predict the markets.Time-inconsistent (e.g. hyperbolic) discounting and tragedy-of-the-commons competition still results in resource depletion.
Really? Did the Easter Island civilization operate in an efficient modern market?
Their market wasn't modern by definition, but why would you assume it was any less efficient than any modern market?
"The problem is that people tend to focus on short-term benefits, discounting (hyperbolically, perhaps) future costs. "
It doesn't matter what "people" do. What matters is what the people with money do. Those that behave irrationally tend make poor investments or overconsume, and as such, their decisions have a smaller effect on the market, while those with the foresight to to make sound predictions have a much larger influence.
"Consider the Easter Island civilization: one could be tempted to think that, when palm trees became scarce, rising prices should..."
Really? Did the Easter Island civilization operate in an efficient modern market? I don't think that's a very good example.
The problem is that people tend to focus on short-term benefits, discounting (hyperbolically, perhaps) future costs. Moreover, even if exponential discounting is used, competition can lead to "tragedy of the commons" situations.
Consider the Easter Island civilization: one could be tempted to think that, when palm trees became scarce, rising prices should have reduced consuption and hence exploitation, leading to an equilibrium at the replacement rate. It didn't happen. Easter Islanders cut down their last palm tree and their civilization disintegrated.
Hyperbolic discounting might have been evolutionary advantageous for hunter-gatherers who lived in a world full of unpredictable dangers.
Should we use it for modern public policy decision making? Our environment is more predictable than the ancestral environment.
While there are possible catastrophes that can cause the collapse of modern civilization or even human extinction (large asteroid impact, mega-eruption, world war with WMDs, etc.) they are unlikely.
It makes more sense to use exponential discounting, which is regret-free.
Ants can't disrupt our vital infrastructure.
You're right, there are more proven reserves of natural gas (1.09 * 10^19 J) rather than coal (2.07 * 10^16 J).
Thanks for pointing out.
> 1) Electronics switch faster than neurons.
How many electronic switches are required to match the computational power of a neuron? We don't even have the resources to emulate a rat brain.
> 2) Electrical memory is 100 percent accurate, where human memory isn't even close.
How many terabytes would be required to keep a 100% accurate record of a lifetime of human experiences?
> 3) Computers can communicate with one another much more quickly than humans.
But still have substantial communication delays which are ultimately bounded by the speed of light. You will not get a real-time intelligent machine by wiring together one billion dumb computers scattered around the planet.
> 4) There is reason to believe humans can design massively parallel computers.
And, given equivalent computational resources, parallel computers perform worse than sequential computers. People are turning to parallelization because we can't make our computers run faster.
>5) Computers don't sleep. They only need electricity to eat.
Electricity doesn't exactly grow on trees. Currently, electricity expenditure is becoming a larger and larger fraction of the costs of companies that operate large clusters. Soon it may become the limiting factor of the ICT economy.Computers also need plastic, copper, aluminium, silicon and rare elements.
Human-level AI may be possible. But taking over the world, not so much.
"Hence it is in their interest to avoid conflict."
In the machines' interest? I don't see that as true...any more than it is in our interest to avoid conflict with ants. (We are the ants in this analogy.)
"IIUC, natural gas is much more scarce than coal, hence, even with the higher efficiency of its power stations (50% vs 38% of coal-fired power stations) it will not replace coal."
Jesse Ausubel thinks natural gas will replace coal.
Decarbonization: the next 100 years
I agree. I can install a fuel cell that operates on natural gas in my crawl space. (Or I will be able to do it in a decade or two, as prices come down.) Then I get both electricity and hot water from my natural gas.
In contrast, no way would I want coal in my crawl space, or even anywhere near my neighborhood.
Regarding photovoltaics, the EROEI is higher than one. That's basically all that matters, since the energy produced is electricity, which is very high quality.
"We won't dig up the last piece of coal, but eventually it will become uneconomic to keep using it. At that point, for all practicall purposes, we will have been run out of coal."
In the United States, natural gas is already beating coal on economics...especially considering the recent requirements to add much more air pollution control (for mercury, nitrogen oxides, sulfur dioxide, and particulate).
Plus, no one wants a coal mine in their yard. But lots of people have accepted hydrofracking natural gas wells in their yard
IIUC, natural gas is much more scarce than coal, hence, even with the higher efficiency of its power stations (50% vs 38% of coal-fired power stations) it will not replace coal.
Photovoltaic currently has an EROEI lower than coal. While technological developments might increase it, we don't expect large improvements.
As coal becomes more scarce, its EROEI will decrease, hence at some point other energy sources (photovoltaics, uranium breeders and thorium breeders) will become competitive. We won't dig up the last piece of coal, but eventually it will become uneconomic to keep using it. At that point, for all practicall purposes, we will have been run out of coal.
Do you mean that you think it’s possible that there is an unlimited amout of coal?Or that our yearly demand of coal falls to zero?I suppose the latter is possible, but how it would practically differ from having run out of coal?
As has been often noted, the Stone Age didn't end when we ran out of stone, and the Bronze Age didn't end when we ran out of bronze. The Coal Age is almost certainly going to end in this century. But not because there won't be any coal left in the ground. Instead, it will be because natural gas and photovoltaics are better sources of electricity.
Even when we are relatively certain that a cost now will result in benefits in the future (for example: using less gas now will mean there will be more gas to use in the future. duh)
Well...ummm...I don't agree. There may be less gas made from *fossil* fuels in the future. But there's no reason why using more gas now will result in appreciably less renewable gas in the future. For example, diesel can be made from algae. That essentially provides a never-ending source of diesel. (And high-quality, low-sulfur diesel, at that.)
If modern history teaches us anything, it’s that when the rich and powerful rule on the poor and disenfranchised, but when the poor and disenfranchised rebel, the rich and powerful usually end up dead, hence it is interest to avoid conflict.
Oh, I forgot one more basic principle why computers will take over the world: it takes humans ~9 months to grow a new human, and about 18 years to train it. It constrast, terminators can be built at prodigious rates, and are essentially ready to start killing right off the assembly line.
You're right about the historical "rich and powerful". But that's merely because, historically, the "rich and powerful" have been vastly outnumbered by the poor and desperate.
In contrast, the rich and powerful machines can easily outnumber us. That's because they need only for electricity to eat, need no clothes to wear, and need little or no shelter.
But I, for one, will welcome our new machine overlords.
We dumped our trash in the street, left our dead horses in the streets, left the horse shit in the streets because the streets were wide and our waste so small.
Plus it was a lot cheaper that hauling it out of town.
I think I heard on the radio, government dictates mandating hauling waste a mile out of town were first issued in ancient Greece.
The fossil fuel industry are still pretty much in the dumping waste in the streets because they are wide and its cheap stage.
And I guess you believe no one predicted demand increase outstripping supply increase in the 90s which resulted in the sharp run up in oil prices in the 00s rather than the price holding steady at the 1990 oil price. Shouldn't producers been holding back oil in the US in the 90s so in the 00s oil production would have increased instead of falling even as regulation was relaxed and prices skyrocketed? Why did production in the US increase as prices fell and regulation increased?
Can you explain how you deduce from established principles that AI will eventually take over the world?
First principles:
1) Electronics switch faster than neurons.
2) Electrical memory is 100 percent accurate, where human memory isn't even close.
3) Computers can communicate with one another much more quickly than humans.
4) There is reason to believe humans can design massively parallel computers.
5) Computers don't sleep. They only need electricity to eat. They can't be bargained with. They can't be reasoned with. They don't feel pity, or remorse, or fear. And they absolutely will not stop, ever, until we are dead.
Kyle Reese (killed by computer)
If you think price signals are undervaluing some alternatives, then you should work on that, if you are right, you'll get rich.
It doesn’t matter what “people” do. What matters is what the people with money do.
But this is a distinction without a difference, since virtually anybody has money (or even without money partecipates in economic transactions of some sort). There is nothing that compels Average Joe to drive a car instead of riding a bicycle and to eat meat and fish instead of soy. A few billions Average Joes driving their cars and eating meat results in resource depletion and global warming.
A few billions Average Mohammeds and Average Manuels fathering six children each because they are too poor to accumulate assets for retirement and their countries offer no social security also contribute to the problem, but not as much as the Joes do.
Those that behave irrationally tend make poor investments or overconsume, and as such, their decisions have a smaller effect on the market, while those with the foresight to to make sound predictions have a much larger influence.
That why massive speculative bubbles that regularly inflate and burst don't exist, do they?
Belief that professional investors are essentially rational is an empirical falsehood. And even if they were rational, their ability to predict the markets would be limited by the intrinsic chaoticity of the system.
But, for the sake of the argument, let's assume that all economic agents, not just investors, everybody, are perfectionally rational and perfectly able to predict the markets.Time-inconsistent (e.g. hyperbolic) discounting and tragedy-of-the-commons competition still results in resource depletion.
Really? Did the Easter Island civilization operate in an efficient modern market?
Their market wasn't modern by definition, but why would you assume it was any less efficient than any modern market?
"The problem is that people tend to focus on short-term benefits, discounting (hyperbolically, perhaps) future costs. "
It doesn't matter what "people" do. What matters is what the people with money do. Those that behave irrationally tend make poor investments or overconsume, and as such, their decisions have a smaller effect on the market, while those with the foresight to to make sound predictions have a much larger influence.
"Consider the Easter Island civilization: one could be tempted to think that, when palm trees became scarce, rising prices should..."
Really? Did the Easter Island civilization operate in an efficient modern market? I don't think that's a very good example.
The problem is that people tend to focus on short-term benefits, discounting (hyperbolically, perhaps) future costs. Moreover, even if exponential discounting is used, competition can lead to "tragedy of the commons" situations.
Consider the Easter Island civilization: one could be tempted to think that, when palm trees became scarce, rising prices should have reduced consuption and hence exploitation, leading to an equilibrium at the replacement rate. It didn't happen. Easter Islanders cut down their last palm tree and their civilization disintegrated.
Hyperbolic discounting might have been evolutionary advantageous for hunter-gatherers who lived in a world full of unpredictable dangers.
Should we use it for modern public policy decision making? Our environment is more predictable than the ancestral environment.
While there are possible catastrophes that can cause the collapse of modern civilization or even human extinction (large asteroid impact, mega-eruption, world war with WMDs, etc.) they are unlikely.
It makes more sense to use exponential discounting, which is regret-free.