For future usage, it sums up to “nobody knows”: AI companies are betting heavily on more demand. But one factor not mentioned is algorithmic improvements. The cost (in dollars) of an equivalent...
For future usage, it sums up to “nobody knows”:
Every researcher we spoke to said that we cannot understand the energy demands of this future by simply extrapolating from the energy used in AI queries today. And indeed, the moves by leading AI companies to fire up nuclear power plants and create data centers of unprecedented scale suggest that their vision for the future would consume far more energy than even a large number of these individual queries.
“The precious few numbers that we have may shed a tiny sliver of light on where we stand right now, but all bets are off in the coming years,” says Luccioni. “Generative AI tools are getting practically shoved down our throats and it’s getting harder and harder to opt out, or to make informed choices when it comes to energy and climate.”
AI companies are betting heavily on more demand. But one factor not mentioned is algorithmic improvements. The cost (in dollars) of an equivalent query is dropping rapidly (note logarithmic scale) and energy usage is a component of the price, so it should be dropping too.
If algorithmic trends continue, we could see more inference move from data centers to desktop and to mobile, where customers pay for their own electricity and power usage is limited by battery life. Apple and Google are building data centers while also doing their best to make mobile inference happen.
Meanwhile, there is also demand for smarter algorithms that use dramatically more tokens, like “reasoning” and “deep research” algorithms. They’re expensive, but some customers seem willing to pay.
When there are dramatic trends in opposite directions, they will cancel out to some extent and that makes predictions especially tricky. It’s difficult to say which trend will win in the end.
You're not wrong, but there's plenty of food for thought in the article and its the first major attempt I've seen to quantify the issue. I think the projections are incredibly valuable and I don't...
When there are dramatic trends in opposite directions, they will cancel out to some extent and that makes predictions especially tricky. It’s difficult to say which trend will win in the end.
You're not wrong, but there's plenty of food for thought in the article and its the first major attempt I've seen to quantify the issue. I think the projections are incredibly valuable and I don't think hand-waving away efficiency when we're talking about corporations which are chasing profit over everything else is a particularly sound argument.
A large problem is that if you bet on the human race using more energy you’re always right. It’s very hard to figure out what, if anything, AI might displace. I don’t love any of these systems...
A large problem is that if you bet on the human race using more energy you’re always right.
It’s very hard to figure out what, if anything, AI might displace.
I don’t love any of these systems like AI or Crypto that have questionable gain but the energy attacks are also lacking in rigor
The Jevons paradox describes a thing that can happen, but not a thing that must happen. Electricity consumption for lighting, for example, is declining due to increased efficiency....
The Jevons paradox describes a thing that can happen, but not a thing that must happen. Electricity consumption for lighting, for example, is declining due to increased efficiency.
Lighting is an example of the Jevons Paradox, which holds that as a good or service becomes less expensive, people consume more of it. So the cheaper the marginal lumen is, the more of them we use. This is certainly true, though it has limits — we don’t run lights all night simply because they are cheap, nor do we use three times the illumination in an already well-lit space because it costs a third of what it once did.
More importantly, thanks to the increasing efficiency of today’s LEDs and the speed with which they are reaching US homes, residential energy consumption for lighting has fallen by half in less than a decade. According to the Energy Information Administration, US households now use more electricity for refrigerators than for lighting, and about the same amount for lighting as for drying clothes. We are past a Jevons paradox here — consuming more lumens as the country’s population grows, but doing so with less and less energy.
I mean thermodynamics also applies. Population across the world is going up, thus energy use must go up. Keeping 8 billion people fed and alive takes more energy than 7 billlion and so on. While...
I mean thermodynamics also applies. Population across the world is going up, thus energy use must go up. Keeping 8 billion people fed and alive takes more energy than 7 billlion and so on. While population growth has slowed, and will likely peak at the end of the century, until then it basically MUST go up. The vast majority of advancements don't help that significantly with efficiency, especially when the majority of the worlds population doesn't actually live in the areas with the top infrastructure.
I think that generalization avoids the fact that there's a log-linear relationship between GDP per capita and energy consumption. As discussed elsewhere here, we could ensure comfortable lives for...
I think that generalization avoids the fact that there's a log-linear relationship between GDP per capita and energy consumption. As discussed elsewhere here, we could ensure comfortable lives for everyone with less energy than we're consuming now.
Yeah i wound up too busy for that topic, but yes if we were a hive mind with perfect resource distribution and absolutely no disagreements on who gets rights things would be better. In the reality...
Yeah i wound up too busy for that topic, but yes if we were a hive mind with perfect resource distribution and absolutely no disagreements on who gets rights things would be better.
In the reality where global politics, resource scarcity, literal geography, personal desires to live in the spot that's most optimal for you rather than the spot that's most optimal for everyone, clashing cultures, absolute dictators, and a million other factors matter it's not anywhere near that easy.
Feeding hungry people isn't hard. Getting the food to them is hard. ESPECIALLY when you start talking on a global scale. Idealistic nonsense like that article annoys me because it boils down to the oh so western centric "capitalism bad" discussions which I disagree with, but fine I get it, except it kinda ignores that huge portions of the world wish that was the only problem they're dealing with.
Even when everyone is on the same page of "these people probably need help" it's a massive undertaking to redistribute resources in a way that realistically helps, and that's before you get into "well look if you could just force-able relocate people to X Y and Z spot to live" conversations which look great on paper right up until you realize a lot of horrific practices have started with a similar thesis for justification.
So to your point, there is NO WAY, we could raise the quality of living across the world for less energy than we're using now. Just the energy cost of moving the energy and resources around to properly distribute would be tremendous.
That doesn't mean it's something that shouldn't be a goal of humanity, but i really don't think these idealistic "what if X just wasn't a problem!" things should be waved around as solutions. They are starting points for the discussion, which is "in the absolute best case, it's this, and we all know that's not what we're dealing with, so where can we go from there?"
I call bullshit on this point in particular. We're already shipping every good imaginable across the planet, constantly. To the point when that happens less frequently its ab economic crisis. We...
Just the energy cost of moving the energy and resources around to properly distribute would be tremendous.
I call bullshit on this point in particular. We're already shipping every good imaginable across the planet, constantly. To the point when that happens less frequently its ab economic crisis. We just pick and choose who gets what according to who has the most money. Which is why "The West" has an unbelievably luxurious life relative to the countries that the resources are extracted from. The rich get richer, the poor get poorer.
Yes, global geopolitics is hard. It doesn't have to be, but it does require shedding notions of nationalism and individualism. It means targeting the elimination of war and despots, and not arranging mineral deals in order to provide military aid.
I think the only way we get there is nuclear war decimating large enough numbers who decide that is to be the end of war.
Shockingly the majority of large successful cities and nations happen to be geographically located in areas that support having major warm water and easily defended ports and as such have been...
I call bullshit on this point in particular. We're already shipping every good imaginable across the planet, constantly.
Shockingly the majority of large successful cities and nations happen to be geographically located in areas that support having major warm water and easily defended ports and as such have been able to develop massive industrialized infrastructure around them.
The majority of the ones who actually need the help do not. Shipping car parts and phones is vastly easier than anything that might spoil. Shipping on well developed infrastructure and roads in countries that don't have active warlords and dictators is easier than the opposite.
So yes, the rich get richer. Absolutely in part due to all sorts of culture and political and societal issues, but also because they happen to be in the spots that make it easier for them to richer.
Before ports it was beasts of burden and before that it was rivers and before that it was easy access to shelter and game. There's a ton that goes into these situations and it's not all some guy with a bag of money. If you want to help the people who aren't rich get richer, you need to actually address the multitude of problems facing them,.
Yes, global geopolitics is hard. It doesn't have to be, but it does require shedding notions of nationalism and individualism
So an alternative way to read this is "if people just didn't have distinct cultures, languages, and beliefs things would be easier" correct? I never understand this point because I'll bet money you find that statement abhorrent and I don't see how it's any different than what you're advocating.
It means targeting the elimination of war and despots, and not arranging mineral deals in order to provide military aid.
Given we're not currently in a utopia the two can often be related. I've seen very few examples of anything approaching a just war, but the Iraq invasion did objectively eliminate a despot with war. Pretty sure it wasn't a net gain or as easy as you're making it sound.
I think the only way we get there is nuclear war decimating large enough numbers who decide that is to be the end of war.
So fuck it and let god sort them out? Because this massive world shattering event leading to the death of billions of humans and untold ecological damage will THIS TIME for some reason bring peace from the ashes? 2 world wars weren't enough? Humans don't exactly default to utopia when ripped from society into pure survival cultures. Quite the opposite if all the horrific local atrocities leading to people fighting to survive are anything to go by.
I've written the rest of this 3 times because I honestly find this view insulting to the cause you claim to care about. For all my cynicism I think its actually possible for humanity to progress without catastrophic annihilation, and bluntly think you're naive if you think that's going to lead to any sort of progress. It's not happening in my lifetime, and it's not happening easily, and it's sure as shit not happening if half the people who claim they want things to be better don't bother to learn about all the blockers and issues. Yes the hording of wealth is a problem. Pretending it's the only one, or like it's something new, solves nothing.
Iraq was never about removing Saddam and liberating the people. That was, at best, a personal win for W Bush since his dad couldn't do it. It was always about the oil. There was no other reason to...
Iraq was never about removing Saddam and liberating the people. That was, at best, a personal win for W Bush since his dad couldn't do it. It was always about the oil. There was no other reason to lie about WMDs. The big oil companies (Exxon, BP, Chevron, Shell) spent more money to get W Bush elected than any other time in history, and by 2013 they owned almost all of the oil fields. GDP per capita is up in Iraq, but for some reason the standard of living is lower than it was before the war.
Heck, before the Iran-Iraq war in the 1980s, Iraq was believed to be one of the best education and health systems in the Middle East. It's amazing what decades of war (and the debts incurred from it) and economic sanctions will do to quality of life for an average person.
We didn't have to impose economic sanctions on Cuba. That was purely because we didn't like their politics, and were essentially trying to starve the population into revolution.
I 100% believe that we could, if willing to shed notions like nationalism and unifying behind the knowledge that if we don't, civilization will collapse. But it is more important to strengthen our borders. To keep the immigrants and refugees out. To impose trade deals that primarily benefit the wealthier country.
To be clear: I don't think we're going to make meaningful progress towards worldwide social parity in a short enough time window to preserve civilization as we know it. I know the unification is just that: a pipe dream. The prophesied nuclear war I spoke of is a Star Trek reference first and foremost. It's not a desirable thing by any means. But I do think that barring that level of destruction there will be no meaningful shedding of nationalism, which will only become stronger as wars for controlling scarce resources intensify.
Sorry to jump in, but now I'm a little confused. Are there enough resources for 8.5 billion to live comfortably with proper distribution, or are resources scarce? Seems like you're sort of arguing...
... which will only become stronger as wars for controlling scarce resources intensify.
Sorry to jump in, but now I'm a little confused. Are there enough resources for 8.5 billion to live comfortably with proper distribution, or are resources scarce? Seems like you're sort of arguing both ways.
When I say "scarce" I mean within the scope of any given nation being able to control the distribution of them and continue to distribute them as we do now. If we're willing to more equitably...
When I say "scarce" I mean within the scope of any given nation being able to control the distribution of them and continue to distribute them as we do now.
If we're willing to more equitably distribute said resources (ie oil rations per person across the world), there is plenty to go around. Just Americans will have to live with say 1 car instead of 2 or 3. Or a cellphone that they keep for 7-10 years instead of 2-4.
When have humans ever been satisfied with "plenty enough to go around"? I don't think rich, comfortable societies are in any way inherently better or more deserving than others. I don't subscribe...
When have humans ever been satisfied with "plenty enough to go around"?
I don't think rich, comfortable societies are in any way inherently better or more deserving than others. I don't subscribe to a Just World mindset. But I also recognize that animal populations tend to expand to use up available resources, making hay while the sun shines as it were. While our species isn't necessarily limited by the same thoughtless animal instinct that leads to the boom-bust cycles one learns about in biology 101, we are still animals, and our behavior is much more akin to the instinct-driven beasts than we care to acknowledge.
Supposing we convinced the genie to nod her head and make resource distribution equitable across the whole of humanity instantaneously, how many people would there be in 30 years time? Or 50? And more to the point, what would their quality of life be? What impact would 10+ billion comfortable people make on the rest of the world?
Again, that's not to say that the desire to improve the quality of life for the impoverished billions is dangerous or foolish–I don't believe in hell, but if there is one, I hope there's a special wing of it devoted to Malthusian demagogues–but I think it does suggest that the notion that simple redistribution of resources isn't sufficient in itself to lead to an equitable, sustainable world (and I'm firmly of the belief that you can't have one without the other). If that's true, then arguments like "we could feed, water and shelter 15 billion if we managed things right" don't really amount to a helluva lot, and the impulse to dismiss this observation as veiled fascism isn't really helpful.
The thing that makes me scratch my head about this topic is that some of the same crowd who criticize capitalism with "unlimited growth is the ideology of the cancer cell" refuse to apply it to...
The thing that makes me scratch my head about this topic is that some of the same crowd who criticize capitalism with "unlimited growth is the ideology of the cancer cell" refuse to apply it to human populations as well. I get why human overpopulation is a fraught subject, since it's a favorite hobby horse of so many with repellant ideologies, but it seems like the default response to the premise "maybe 10 billion people is just too damn many" is to shut down discussion and label the notion fascistic.
I think there's probably room to consider the implications of a growing world population that doesn't inevitably descend to a Thanos Snap or the extermination of the Global South as solutions, but I doubt we'll find it if we insist on equating the idea with Malthus and Hitler and nothing else.
How do you square that against the lowering birthrates as quality of life improves? Animals don't have access to birth control methods (chemical or surgical) or detailed knowledge about the...
Animals don't have access to birth control methods (chemical or surgical) or detailed knowledge about the fertility cycle. I'd wager population would flatten out if the world's wealth was equitably distributed.
Outside of the obvious "correlation does not equal causation," I square it by noticing that our species "isn't necessarily limited by the same thoughtless animal instinct that leads to the...
Outside of the obvious "correlation does not equal causation," I square it by noticing that our species "isn't necessarily limited by the same thoughtless animal instinct that leads to the boom-bust cycles one learns about in biology 101," as I did in my response to your other comment.
To my knowledge, nobody really knows why birth rates have declined in many developed nations. It might very well be a universal trait of our species that we moderate our populations as a spontaneous response to sufficiency, but it could also be a quirk of history. I don't think the phenomenon is understood well enough that we can take it as a universal verity of humanity. I don't think it matters much at present anyway, because believe it or not, I agree that inequitable distribution of resources actually is our most pressing problem.
What I object to are arguments that if we were to somehow overcome the technological and logistical hurdles to equitable distribution, and if we somehow miraculously got 8+ billion people on the same page about how the world's resources and living space should be apportioned (with an extreme preference for non-violence), then we might feed, clothe, water and shelter X billions of people. I don't find such arguments persuasive or helpful in any meaningful way, and I don't think people should accept them uncritically and use them as justification to shut down discussion, as I have seen quite a few times in the past.
It took a while to find a paper with a nice "up and to the right" figure, but I found one in Will Energy-Hungry AI Create a Baseload Power Demand Boom? This image doesn't really have historical...
Both of these look very much like exponential growth, and so, while some algorithms may become more efficient on a per-operation basis, the overall effect seems like it will be to increase total utilization, as happens with Jevons Paradox.
To be clear, it looks like we are currently generating 12,000 TWh in 2024 and doubling every 20 years here. We should be at about 500 TWh for data centers with a ~100-fold increase every 20 years. So this would predict that we will be using 50,000 TWh for computing out of a total available 24,000 TWh of energy produced in 2044. So, something's gotta give.
This is also a great spot to point out that at current growth rates civilization has had since about 1650, if we migrate all energy production to solar panels, we will consume all of Earth's solar...
This is also a great spot to point out that at current growth rates civilization has had since about 1650, if we migrate all energy production to solar panels, we will consume all of Earth's solar budget within 400 years.
I have noticed that reasoning models actually provide good output. I find that they catch their hallucinations during their reasoning process and fix them before showing you the final output. It’s...
I have noticed that reasoning models actually provide good output. I find that they catch their hallucinations during their reasoning process and fix them before showing you the final output. It’s not required for every prompt, but I tend to just default to reasoning models because I’m lazy and don’t want to have to correct it when it makes mistakes.
One thing that is being researched are models that read your prompt and automatically determine which model will best handle it. This kind of pre-processing could decrease the amount of energy wasted on simple prompts that get sent to reasoning models.
I am deliberately chasing a sort of naive optimism that this surge in demand for power — if it lasts — will finally, FINALLY be the kick in the pants we need to transition to sustainable green...
I am deliberately chasing a sort of naive optimism that this surge in demand for power — if it lasts — will finally, FINALLY be the kick in the pants we need to transition to sustainable green energy.
I know the political tides are moving in the opposite direction. I know history has time and again showed that moments like these just enrich the entrenched usual suspects. I know the techno-utopian ethos has all but evaporated from Silicon Valley. But I gotta have hope.
Honestly I know nuclear is contentious but I’m a bit relieved (and simultaneously deeply unnerved) to see companies looking into constructing private reactors. It’s at least better than doubling down on more coal. I’ll take little victories where I can.
I do not want privately managed reactors unless they are held to an extremely high auditing process. Like USDA-level full-time government auditor in each one that can hit the panic button at any...
I do not want privately managed reactors unless they are held to an extremely high auditing process. Like USDA-level full-time government auditor in each one that can hit the panic button at any time if something is even a mouse-fart off.
I think they're far more likely to propagandize "clean coal" again.
The problem is also much like with crypto: Even if it is 100% green, we would have been better off collecting that expense in tax money to put it all on the public grid and ban building additional data centers.
Increasing green energy isn't good enough if it isn't eating into the backlog of dirty energy.
I'm not very informed, but I was under the impression a lot of the reactor designs (like China's thorium / molten salt reactor) were pretty resistant to things going wrong. Not that they shouldn't...
I'm not very informed, but I was under the impression a lot of the reactor designs (like China's thorium / molten salt reactor) were pretty resistant to things going wrong. Not that they shouldn't also have a lot of oversight.
If you can't trust the government to do it, you sure as shit can't trust the assholes funding said attack. I prefer my electric plants publicly owned by nonprofit NGOs, somewhat solated from...
If you can't trust the government to do it, you sure as shit can't trust the assholes funding said attack.
I prefer my electric plants publicly owned by nonprofit NGOs, somewhat solated from politics, but also removed from the profit motives.
Personally, I'd like to see these companies throw shitloads of money at nuclear fusion. If we finally crack that nut, so many possibilities will open up. It's a long play that AI companies would...
Personally, I'd like to see these companies throw shitloads of money at nuclear fusion. If we finally crack that nut, so many possibilities will open up. It's a long play that AI companies would likely only invest a little in because of time-delayed ROI.
If only we had the Apollo Program for nuclear fusion.
And then I saw Helion Energy raised half a billion USD with sizable investments from Altman.
No. More than that. Because the inspectors did their job: they repeatedly found and reported, but hey likely didn't have the authority to stop it themselves. They had to push it up the rung.
No. More than that. Because the inspectors did their job: they repeatedly found and reported, but hey likely didn't have the authority to stop it themselves. They had to push it up the rung.
As a resident of Virginia, it was "fun" to see us used as an example so much. heh. Especially the bit about maybe paying $37.50/mo to subsidize data centers. I know we've had some plans for action...
As a resident of Virginia, it was "fun" to see us used as an example so much. heh. Especially the bit about maybe paying $37.50/mo to subsidize data centers.
I know we've had some plans for action around here. In the article, there's a map of existing and planned datacenters, and I'm aware of the plans to build one in Virginia Beach - they're hoping to make a nice hub for undersea cables there.
I didn't see — or maybe the map was just too dense/small — the datacenter I've been hearing about planned near Smithfield (where the ham of the same name originates). So I wonder how many other plans they're missing.
On the energy usage… it seems like we're going to have to keep working on renewables. I think even Republicans realize this to some degree.
From everything I've read, nuclear is pretty darn safe. The design of Chernobyl had some major loopholes in safety margins that they've learned from. I think the technolgy has matured to the point of being pretty safe. But people are so scared of it. But I think we need more nuclear, and we need more solar+battery.
For future usage, it sums up to “nobody knows”:
AI companies are betting heavily on more demand. But one factor not mentioned is algorithmic improvements. The cost (in dollars) of an equivalent query is dropping rapidly (note logarithmic scale) and energy usage is a component of the price, so it should be dropping too.
If algorithmic trends continue, we could see more inference move from data centers to desktop and to mobile, where customers pay for their own electricity and power usage is limited by battery life. Apple and Google are building data centers while also doing their best to make mobile inference happen.
Meanwhile, there is also demand for smarter algorithms that use dramatically more tokens, like “reasoning” and “deep research” algorithms. They’re expensive, but some customers seem willing to pay.
When there are dramatic trends in opposite directions, they will cancel out to some extent and that makes predictions especially tricky. It’s difficult to say which trend will win in the end.
You're not wrong, but there's plenty of food for thought in the article and its the first major attempt I've seen to quantify the issue. I think the projections are incredibly valuable and I don't think hand-waving away efficiency when we're talking about corporations which are chasing profit over everything else is a particularly sound argument.
Sure, I don’t have a crystal ball either. Most people are betting on increased energy usage and there are reasons for that.
A large problem is that if you bet on the human race using more energy you’re always right.
It’s very hard to figure out what, if anything, AI might displace.
I don’t love any of these systems like AI or Crypto that have questionable gain but the energy attacks are also lacking in rigor
Hurray for Jevon's paradox, which is why energy consumption won't go down without regulation and/or punitive taxes/fines.
The Jevons paradox describes a thing that can happen, but not a thing that must happen. Electricity consumption for lighting, for example, is declining due to increased efficiency.
https://archive.is/fqfjE
I mean thermodynamics also applies. Population across the world is going up, thus energy use must go up. Keeping 8 billion people fed and alive takes more energy than 7 billlion and so on. While population growth has slowed, and will likely peak at the end of the century, until then it basically MUST go up. The vast majority of advancements don't help that significantly with efficiency, especially when the majority of the worlds population doesn't actually live in the areas with the top infrastructure.
I think that generalization avoids the fact that there's a log-linear relationship between GDP per capita and energy consumption. As discussed elsewhere here, we could ensure comfortable lives for everyone with less energy than we're consuming now.
Yeah i wound up too busy for that topic, but yes if we were a hive mind with perfect resource distribution and absolutely no disagreements on who gets rights things would be better.
In the reality where global politics, resource scarcity, literal geography, personal desires to live in the spot that's most optimal for you rather than the spot that's most optimal for everyone, clashing cultures, absolute dictators, and a million other factors matter it's not anywhere near that easy.
Feeding hungry people isn't hard. Getting the food to them is hard. ESPECIALLY when you start talking on a global scale. Idealistic nonsense like that article annoys me because it boils down to the oh so western centric "capitalism bad" discussions which I disagree with, but fine I get it, except it kinda ignores that huge portions of the world wish that was the only problem they're dealing with.
Even when everyone is on the same page of "these people probably need help" it's a massive undertaking to redistribute resources in a way that realistically helps, and that's before you get into "well look if you could just force-able relocate people to X Y and Z spot to live" conversations which look great on paper right up until you realize a lot of horrific practices have started with a similar thesis for justification.
So to your point, there is NO WAY, we could raise the quality of living across the world for less energy than we're using now. Just the energy cost of moving the energy and resources around to properly distribute would be tremendous.
That doesn't mean it's something that shouldn't be a goal of humanity, but i really don't think these idealistic "what if X just wasn't a problem!" things should be waved around as solutions. They are starting points for the discussion, which is "in the absolute best case, it's this, and we all know that's not what we're dealing with, so where can we go from there?"
I call bullshit on this point in particular. We're already shipping every good imaginable across the planet, constantly. To the point when that happens less frequently its ab economic crisis. We just pick and choose who gets what according to who has the most money. Which is why "The West" has an unbelievably luxurious life relative to the countries that the resources are extracted from. The rich get richer, the poor get poorer.
Yes, global geopolitics is hard. It doesn't have to be, but it does require shedding notions of nationalism and individualism. It means targeting the elimination of war and despots, and not arranging mineral deals in order to provide military aid.
I think the only way we get there is nuclear war decimating large enough numbers who decide that is to be the end of war.
Shockingly the majority of large successful cities and nations happen to be geographically located in areas that support having major warm water and easily defended ports and as such have been able to develop massive industrialized infrastructure around them.
The majority of the ones who actually need the help do not. Shipping car parts and phones is vastly easier than anything that might spoil. Shipping on well developed infrastructure and roads in countries that don't have active warlords and dictators is easier than the opposite.
So yes, the rich get richer. Absolutely in part due to all sorts of culture and political and societal issues, but also because they happen to be in the spots that make it easier for them to richer.
Before ports it was beasts of burden and before that it was rivers and before that it was easy access to shelter and game. There's a ton that goes into these situations and it's not all some guy with a bag of money. If you want to help the people who aren't rich get richer, you need to actually address the multitude of problems facing them,.
So an alternative way to read this is "if people just didn't have distinct cultures, languages, and beliefs things would be easier" correct? I never understand this point because I'll bet money you find that statement abhorrent and I don't see how it's any different than what you're advocating.
Given we're not currently in a utopia the two can often be related. I've seen very few examples of anything approaching a just war, but the Iraq invasion did objectively eliminate a despot with war. Pretty sure it wasn't a net gain or as easy as you're making it sound.
So fuck it and let god sort them out? Because this massive world shattering event leading to the death of billions of humans and untold ecological damage will THIS TIME for some reason bring peace from the ashes? 2 world wars weren't enough? Humans don't exactly default to utopia when ripped from society into pure survival cultures. Quite the opposite if all the horrific local atrocities leading to people fighting to survive are anything to go by.
I've written the rest of this 3 times because I honestly find this view insulting to the cause you claim to care about. For all my cynicism I think its actually possible for humanity to progress without catastrophic annihilation, and bluntly think you're naive if you think that's going to lead to any sort of progress. It's not happening in my lifetime, and it's not happening easily, and it's sure as shit not happening if half the people who claim they want things to be better don't bother to learn about all the blockers and issues. Yes the hording of wealth is a problem. Pretending it's the only one, or like it's something new, solves nothing.
Iraq was never about removing Saddam and liberating the people. That was, at best, a personal win for W Bush since his dad couldn't do it. It was always about the oil. There was no other reason to lie about WMDs. The big oil companies (Exxon, BP, Chevron, Shell) spent more money to get W Bush elected than any other time in history, and by 2013 they owned almost all of the oil fields. GDP per capita is up in Iraq, but for some reason the standard of living is lower than it was before the war.
Heck, before the Iran-Iraq war in the 1980s, Iraq was believed to be one of the best education and health systems in the Middle East. It's amazing what decades of war (and the debts incurred from it) and economic sanctions will do to quality of life for an average person.
We didn't have to impose economic sanctions on Cuba. That was purely because we didn't like their politics, and were essentially trying to starve the population into revolution.
I 100% believe that we could, if willing to shed notions like nationalism and unifying behind the knowledge that if we don't, civilization will collapse. But it is more important to strengthen our borders. To keep the immigrants and refugees out. To impose trade deals that primarily benefit the wealthier country.
To be clear: I don't think we're going to make meaningful progress towards worldwide social parity in a short enough time window to preserve civilization as we know it. I know the unification is just that: a pipe dream. The prophesied nuclear war I spoke of is a Star Trek reference first and foremost. It's not a desirable thing by any means. But I do think that barring that level of destruction there will be no meaningful shedding of nationalism, which will only become stronger as wars for controlling scarce resources intensify.
Sorry to jump in, but now I'm a little confused. Are there enough resources for 8.5 billion to live comfortably with proper distribution, or are resources scarce? Seems like you're sort of arguing both ways.
When I say "scarce" I mean within the scope of any given nation being able to control the distribution of them and continue to distribute them as we do now.
If we're willing to more equitably distribute said resources (ie oil rations per person across the world), there is plenty to go around. Just Americans will have to live with say 1 car instead of 2 or 3. Or a cellphone that they keep for 7-10 years instead of 2-4.
When have humans ever been satisfied with "plenty enough to go around"?
I don't think rich, comfortable societies are in any way inherently better or more deserving than others. I don't subscribe to a Just World mindset. But I also recognize that animal populations tend to expand to use up available resources, making hay while the sun shines as it were. While our species isn't necessarily limited by the same thoughtless animal instinct that leads to the boom-bust cycles one learns about in biology 101, we are still animals, and our behavior is much more akin to the instinct-driven beasts than we care to acknowledge.
Supposing we convinced the genie to nod her head and make resource distribution equitable across the whole of humanity instantaneously, how many people would there be in 30 years time? Or 50? And more to the point, what would their quality of life be? What impact would 10+ billion comfortable people make on the rest of the world?
Again, that's not to say that the desire to improve the quality of life for the impoverished billions is dangerous or foolish–I don't believe in hell, but if there is one, I hope there's a special wing of it devoted to Malthusian demagogues–but I think it does suggest that the notion that simple redistribution of resources isn't sufficient in itself to lead to an equitable, sustainable world (and I'm firmly of the belief that you can't have one without the other). If that's true, then arguments like "we could feed, water and shelter 15 billion if we managed things right" don't really amount to a helluva lot, and the impulse to dismiss this observation as veiled fascism isn't really helpful.
The thing that makes me scratch my head about this topic is that some of the same crowd who criticize capitalism with "unlimited growth is the ideology of the cancer cell" refuse to apply it to human populations as well. I get why human overpopulation is a fraught subject, since it's a favorite hobby horse of so many with repellant ideologies, but it seems like the default response to the premise "maybe 10 billion people is just too damn many" is to shut down discussion and label the notion fascistic.
I think there's probably room to consider the implications of a growing world population that doesn't inevitably descend to a Thanos Snap or the extermination of the Global South as solutions, but I doubt we'll find it if we insist on equating the idea with Malthus and Hitler and nothing else.
How do you square that against the lowering birthrates as quality of life improves?
Animals don't have access to birth control methods (chemical or surgical) or detailed knowledge about the fertility cycle. I'd wager population would flatten out if the world's wealth was equitably distributed.
Outside of the obvious "correlation does not equal causation," I square it by noticing that our species "isn't necessarily limited by the same thoughtless animal instinct that leads to the boom-bust cycles one learns about in biology 101," as I did in my response to your other comment.
To my knowledge, nobody really knows why birth rates have declined in many developed nations. It might very well be a universal trait of our species that we moderate our populations as a spontaneous response to sufficiency, but it could also be a quirk of history. I don't think the phenomenon is understood well enough that we can take it as a universal verity of humanity. I don't think it matters much at present anyway, because believe it or not, I agree that inequitable distribution of resources actually is our most pressing problem.
What I object to are arguments that if we were to somehow overcome the technological and logistical hurdles to equitable distribution, and if we somehow miraculously got 8+ billion people on the same page about how the world's resources and living space should be apportioned (with an extreme preference for non-violence), then we might feed, clothe, water and shelter X billions of people. I don't find such arguments persuasive or helpful in any meaningful way, and I don't think people should accept them uncritically and use them as justification to shut down discussion, as I have seen quite a few times in the past.
It took a while to find a paper with a nice "up and to the right" figure, but I found one in Will Energy-Hungry AI Create a Baseload
Power Demand Boom? This image doesn't really have historical data with it, at least not beyond 2020. A nice plot for the pre-2020 years is in the Semiconductor Research Corporation 2021 Decadal Plan for Semiconductors
Both of these look very much like exponential growth, and so, while some algorithms may become more efficient on a per-operation basis, the overall effect seems like it will be to increase total utilization, as happens with Jevons Paradox.
To be clear, it looks like we are currently generating 12,000 TWh in 2024 and doubling every 20 years here. We should be at about 500 TWh for data centers with a ~100-fold increase every 20 years. So this would predict that we will be using 50,000 TWh for computing out of a total available 24,000 TWh of energy produced in 2044. So, something's gotta give.
This is also a great spot to point out that at current growth rates civilization has had since about 1650, if we migrate all energy production to solar panels, we will consume all of Earth's solar budget within 400 years.
Math is fun
I have noticed that reasoning models actually provide good output. I find that they catch their hallucinations during their reasoning process and fix them before showing you the final output. It’s not required for every prompt, but I tend to just default to reasoning models because I’m lazy and don’t want to have to correct it when it makes mistakes.
One thing that is being researched are models that read your prompt and automatically determine which model will best handle it. This kind of pre-processing could decrease the amount of energy wasted on simple prompts that get sent to reasoning models.
I am deliberately chasing a sort of naive optimism that this surge in demand for power — if it lasts — will finally, FINALLY be the kick in the pants we need to transition to sustainable green energy.
I know the political tides are moving in the opposite direction. I know history has time and again showed that moments like these just enrich the entrenched usual suspects. I know the techno-utopian ethos has all but evaporated from Silicon Valley. But I gotta have hope.
Honestly I know nuclear is contentious but I’m a bit relieved (and simultaneously deeply unnerved) to see companies looking into constructing private reactors. It’s at least better than doubling down on more coal. I’ll take little victories where I can.
I do not want privately managed reactors unless they are held to an extremely high auditing process. Like USDA-level full-time government auditor in each one that can hit the panic button at any time if something is even a mouse-fart off.
I think they're far more likely to propagandize "clean coal" again.
The problem is also much like with crypto: Even if it is 100% green, we would have been better off collecting that expense in tax money to put it all on the public grid and ban building additional data centers.
Increasing green energy isn't good enough if it isn't eating into the backlog of dirty energy.
I'm not very informed, but I was under the impression a lot of the reactor designs (like China's thorium / molten salt reactor) were pretty resistant to things going wrong. Not that they shouldn't also have a lot of oversight.
With the current attack and subsequent erosion of governmental oversight, I'm not sure I want the government to control that either.
If you can't trust the government to do it, you sure as shit can't trust the assholes funding said attack.
I prefer my electric plants publicly owned by nonprofit NGOs, somewhat solated from politics, but also removed from the profit motives.
Personally, I'd like to see these companies throw shitloads of money at nuclear fusion. If we finally crack that nut, so many possibilities will open up. It's a long play that AI companies would likely only invest a little in because of time-delayed ROI.
If only we had the Apollo Program for nuclear fusion.
And then I saw Helion Energy raised half a billion USD with sizable investments from Altman.
Like the one at Boar's Head? That level of oversight and stringency?
No. More than that. Because the inspectors did their job: they repeatedly found and reported, but hey likely didn't have the authority to stop it themselves. They had to push it up the rung.
The transition is well underway. Solar and battery storage were over 80% of new capacity in the US.
As a resident of Virginia, it was "fun" to see us used as an example so much. heh. Especially the bit about maybe paying $37.50/mo to subsidize data centers.
I know we've had some plans for action around here. In the article, there's a map of existing and planned datacenters, and I'm aware of the plans to build one in Virginia Beach - they're hoping to make a nice hub for undersea cables there.
I didn't see — or maybe the map was just too dense/small — the datacenter I've been hearing about planned near Smithfield (where the ham of the same name originates). So I wonder how many other plans they're missing.
On the energy usage… it seems like we're going to have to keep working on renewables. I think even Republicans realize this to some degree.
From everything I've read, nuclear is pretty darn safe. The design of Chernobyl had some major loopholes in safety margins that they've learned from. I think the technolgy has matured to the point of being pretty safe. But people are so scared of it. But I think we need more nuclear, and we need more solar+battery.
Mirror: https://archive.is/mnHb8