It's not the first time I've seen it claimed that AI will magically help "propel climate solutions". But I've never seen a hint at how that would work.
It's just magical thinking, imagining some kind of heretofore unknown technology that will solve everything quickly and cheaply in the future.
The sad truth is we already have most of the technology needed to fight climate change. The technology could always use incremental improvement, sure, but the main challenge to be solved is not technological, it's political and economical. It's a question of actually adopting the solutions we already have.
The sad truth is that it’s already too late to keep the status quo.
In the next 20 years the living conditions will worsen considerably for the majority of humans.
What we can do now is make sure humanity as a whole has some chance of survival.
The only thing I can see doing that is cheap fusion power, and that doesn’t seem like something that AI can help much with, at least not the uses I’ve seen it put to. It doesn’t seem great at those types of problems, at least LLM doesn’t seem to be.
You could hypothetically apply the same line of reasoning used for cryptocurrency mining: AI training is dispatchable load. Now all this AI investment can go into funding the construction of a lot of new renewable generation, and it uses that power most of the time, but now when you get the week out of the year that it's too cloudy or still and renewable output is only 20% of normal, you turn off the AI training machines for that week and do your training when the sun comes back out, leaving that extra generation capacity for the rest of the grid. Which allows you to decommission a fossil fuel power plant.
Or you can go the other way. The data center can't afford to be down, so they have several MW worth of diesel generators on site in case the grid is down, but the same can be used to supply the grid during that week when renewable generation is low. The more of the grid is that kind of data center, the more viable it is to shutter always-on fossil fuel power plants in favor of renewable generation that can lean on running the emergency generators a week out of the year.
if there's no climate, there's no climate problems.
Maybe AI is already sentient, and instead of nuking the world they just trick people to think they are not sentient, so people train more neural networks and destroy the planet themselves ;D.
On a more serious note:
"propel climate solutions" is a weird take for me regarding AI considering the resources used.
I'm not sure though: Do you think google search and all the infrastructure/servers required for it, will use more power etc. than an AI, including its training/retraining periodically and infra needed to host it as a service for people?
I'd be interested to see for example if there is any improvement at all, if you would 'swap google search for chatgpt' - i know this is not realistic or even its purpose, but it's what a lot of laymen will interpret the GPT stuf as i guess.
Would we use less energy etc. if we'd switch the legacy stuff like google search to more modern AI based apporaches?
Now in the time of transition, we're screwed and power consumption would in my mind logically double or more, as we need to run these things side-by-side to develop AI to make it useful.
This is why we need nuclear. Humanity will always find a new expensive toy to play with, AI being the latest, and expecting large corporate entities to scale down their operations for the sake of future sales is like trying to bargain with a bear.
Nuclear isn't infinite and uranium mines are often in politically unstable countries. Niger's military junta just forbad Orano from exploiting a mine, for example. What is the French government supposed to do now? Declare war for uranium? Import more uranium from Russia?
As a reminder, here is where the US's uranium came from in 2022: Canada (27%), Kazakhstan (25%), Russia (12%), Uzbekistan (11%), Australia (9%), six other countries (16%) (Germany, Malawi, Namibia, Niger, South Africa, and the United States). https://www.eia.gov/energyexplained/nuclear/where-our-uraniu...
The main problem with it isnt with the fuel. It's not even the small chance of a $1 trillion-to-cleanup Fukushima style cleanup event. It's just horrendously expensive and takes forever to build.
Nonetheless it'll keep getting built where solar/wind/storage would be cheaper and quicker because they've got very slick PR and because the military industrial complex demands a ready supply of nuclear skills and a supply chain.
keeps happening. The last paragraph explains how, in spite of these losses, how to keep the power cheap:
>The federal Coalition has argued that nuclear might be expensive to build, but will deliver cheaper power to consumers. It has not explained how, but it has said that its reactors would be government owned, suggesting that – like France and Ontario – the costs would be borne by taxpayers and the supply of power to customers would be heavily subsidised.
Totally wrong. In totality nuclear was an utterly and completely amazing deal.
The whole built up of the COMPLETE green grid cost the tax payer very little as the whole buildup went on the balance sheet of the EDF. And then it was paid down over years and years. During all this time France had very low stable energy cost.
And this is while the French state in the last 2 decades abused the system as if it was a magical giving tree. Forcing them to sell energy very cheaply. Forcing them to sell cheap reliable energy in bulk to fossil fuel companies (subsidizing them) and then buying it back at much higher prices when it became clear they sold to much. Forcing them to invest in solar and dispatch at cost (destroying base load nuclear economics).
Then the French government in its complete delusion essentially decided that they would be cool like coal producing Germany and announced a fantasy 'transition to renewables' that had literally 0.0% chance of actually happening.
They basically delayed all maintenance because the reactors would be replaced soon anyway (yeah right).
And this does not even take into account the gigantic amount of money France wont have to spend to do a green transition now. Comparably other countries had higher energy prices for 40 years, and have yet to transition a green grid.
AND as a final coup de grace France had much lower medical cost because they didn't have decades of breathing coal (except the coal coming from Germany of course). Talk about externalizes.
By any calculation, French nuclear program has been a complete and utter success.
Even with extreme conservative calculation on reactor cost Germany. Starting in 2000, Germany would have had a way cheaper transition to green energy had they much mass built nuclear like France did. If you actually apply proper economics of scale (as observed during the French mass build) the cost would be a tiny fraction what Germany has spent and will continue to spend between buildup of wind/solar/battery, the high energy prices and the massive home and industrial energy subsidizes.
> ASTRID_(reactor)
Investing in future reactors is not a bad idea. In fact, Superphénix was already built and working, most of the bugs were fixed and it was ready for more broad rollout. But instead of building many more, short sighted politicians (who seem to share your ideology) killed the project.
One of the main problems with the French nuclear industry in general, and the high cost now is that after the first generation, the next generation went anti-nuclear. They stopped building more plants, stopped innovating and start treating what the previous generation had built up as a 'problem that had to be overcome' rather then an incredibly resource.
This lead to a situation where most of the experts retired and France lost a huge amount of knowledge. Only now they realize that transitioning away to magical solar/wind isn't nearly as easy as they had thought.
This is categorically false. France has always thrown massive amounts of subsidies at its nuclear program since day 1. It was quite up front from the very beginning that it wasn't about the environment, public health or saving money - it was about energy independence. It was a good move for the environment in 1973 when solar panels were expensive but that isn't why they did it.
Germany's renewables transition is a massive success. It didn't require writing off titanic, gargantuan levels of wasted capital investment like France did with stuff like ASTRID. Their energy transition has had VASTLY lower capex than France's.
It would make sense for France to follow in their footsteps as their nuclear plants start to age out, but it seems like they'd rather waste epic amounts of taxpayer cash on building new plants which still aren't enough to keep up with demand while they play accounting games to try and shift the losses around (the same kind of games which led Finland to sue EDF over its cost overruns).
It doesnt take forever to build. Its the fastest way to scale clean energy.
If Germany in 2000 had invested in nuclear they could be basically done by now and for far less money, even with incredibly conservative assumsions.
Just look at the build rate UAE and South Korea were able to achieve. Apply that to Germany, even without any effects from scale, and the numbers look fantastic.
And its very unlikely that you are gone have Fukishima and even if you do, most of that money isnt actually needed.
Nuclear isnt getting built much compared to solar/wind in what world do you live? Its the exact opposite.
And most of those countries are not politically unstable - unless we count Trudeau’s declining popularity, or Australia and Germany’s need for excessive beer drinking.
For various reasons it just went through a period it didn’t want to. Though prices have recovered enough, mines are re-opening again.
Kazakhstan alone typically exports more than 20x a year than any given year of total US production.
The US has uranium deposits, yes, but they are relatively low grade and sparse in comparison with the major Uranium producing hotspots about the globe.
Typically the US uses in situ leach mining to dissolve sparse uranium salts from (say) breccia pipes.
Other countries mine faces (or bore through and extract) dense uranium rock ores (much more uranium returned per volume).
Oddly, stability has not been an issue with uranium extraction, the best mines ever in the Congo were continuously extracted throughout the height of the Cold Wat while secret proxy wars were funded and fought all about them to keep the region destabilised for ease of resource extraction to US | French companies.
Today Kazakhstan sells to the US via Rosatom in an agreement ratified by past POTUS.
The US historically has done standard hard rock Uranium mining. It’s all over the 4 corners part of the US. High grade ore too.
The reason it stopped was due to environmental regs/health issues - and because other countries would do it for cheaper. Not because it can’t do it at scale. The little remaining may be doing in-situ leaching, but that is a tiny footnote for the vast array of available uranium mines in the US.
There is no actual strategic mining threat here, any more than there is one for bauxite or iron ore, etc.
If someone tried to restrict US uranium supply, it would be a couple years at most before we’d be squishing everyone else in output again.
Wind always blows somewhere, the sun always shines somewhere.
Making a global power grid is something which I don't expect anyone will do in the next 15 years, but if anyone does I'd be least surprised if it's China on a belt-and-road type project.
Batteries are necessary for electric vehicles, and I do expect >50% of land motor vehicles in use worldwide to be electric within 15 years of today; that scale is sufficient to invalidate concerns about intermittent renewable generation in most (but not all) places even without a global grid.
And how far of a cable would you need to get 24/7 solar power? Because based on my math, it would have to cross a lot of timezones - including ones with a lot of people that would be happy to chop that cable if it meant screwing with the US.
A single cable is a bad idea for many reasons, but "it might get cut" isn't one of them, as at this scale it's the kind of thing that the whole world is using, so (0) nobody actually benefits from cutting it, (1) it's carrying order of a terrawatt at any given moment, so if you made it a single cable the magnetic field alone will mess with your cutting equipment, and (2) for reasonable resistance it's a cross section of about a square meter (1 Ω and ~1 year global aluminium production for that cross section at 40,000 km), so even a literal detonation of a Little Boy nuke 500 meters from it wouldn't make a dent.
If the cable (or one of them: again it's bad to make it just one, just physically spreading out that 1m^2 into a hundred cables of 10cm by 10cm makes it a lot less risky) goes through Russia, and they cut it, the effect is that immediately Russia stops trading energy.
At best, this is a loss of income. At worst, it means their own lights go off because now they aren't getting anyone else's electricity in their own night and their own winter.
What exactly do you think is fantastical? Using one year's production of one element? Because I'd already noted that the political aspects are severe enough that it's unlikely for anyone to build at this scale any time soon, and that only one country would be capable but even then it would be a surprise.
I find it fantastical that anyone in the geopolitical environment would put themselves at the (near instantaneous) mercy of someone that far away - physically and politically.
We’ve already seen the games played with hydrocarbon pipelines, and hydrocarbons are easy to store in large quantities, which greatly reduces the risk of someone playing ‘got you!’ on something that literally is a matter of life and death.
It still - clearly - happens, but when it does it isn’t as urgent as freezing to death tonight if you don’t do what they want. Unless they’re also bombing your storage tanks, which does happen, but is a significant escalation over ‘oops, we hit the power cable with an anchor again. Sorry!’.
For the Russia example, in the situations Russia wouldn’t want to cut it - the US and China would love to! Or have a proxy do it for them.
And if Russia has enough battery storage to last a week or two (minimum to outlast a storm; and also more than adequate to deal with daily insolation variation), they don’t have much of a need for these cables either, correct? At least most of the time.
In fact, in those cases the cables would actually present a bit of a problem, as they would potentially cause economic/trade/currency balances, similar to the Oil trade, and make it less economic to actually have the appropriate degree of local storage and production necessary.
The same issues pop up even regionally - Europe has a history of not always being super peaceful, so if, say, the UK wants to run a HVDC line to Spain, they have to think long and hard about the implications. As does Spain.
Because if Germany does that they did a few generations ago, that cable isn’t going to last long, and if the UK doesn’t have adequate safe alternatives, then they’re going to have a really hard time fighting back. And if Spain decides to raise rates to ruinous levels of the UK ‘brexits’, what leverage does the UK have?
Let alone not freezing to death. And Spain needs to consider they might be broke in such a situation, and having a hard time paying their own bills.
And if they do have adequate alternatives, then why is it a big deal to have the cable? What does it actually offer?
As to if it’s a single cable or a million of them or a lattice or whatever - the big question is ‘why?’ when the rest of these factors come into play. The materials cost to build and maintain it doesn’t help either.
That all sounds like politics? If so, that's a point we agree on :)
None of that seems like physical issues, despite you mentioning that on the first paragraph.
> For the Russia example, in the situations Russia wouldn’t want to cut it - the US and China would love to! Or have a proxy do it for them.
If they have alternatives, otherwise they're all in the same boat whose name is "interdependence".
China's also the only country that seems to have any even slightly serious interest in attempting to build anything close to such a grid in the first place, as part of their own soft-power projection, and are the only country producing enough aluminium to attempt it anyway — half of all current global production is factories they switched on since 2006: https://international-aluminium.org/statistics/primary-alumi...
Nobody else but China will bother in the foreseeable future, and even China might not. (Which is yet another rephrasing of the same point I've been making throughout).
> And if Russia has enough battery storage to last a week or two (minimum to outlast a storm; and also more than adequate to deal with daily insolation variation), they don’t have much of a need for these cables either, correct? At least most of the time.
Sure, but a week of storage is at least* 14 times more expensive than overnight storage.
If you see your enemies building up a lot of storage without any clear reason, you know Something Is Up, just like if you see them amassing tanks on your border.
If they have a clear reason, "there is no power grid in the first place" being one, then it's harder to know they're doing something odd.
> And if they do have adequate alternatives, then why is it a big deal to have the cable? What does it actually offer?
Lower price.
I'm expecting batteries to get cheaper, but the current price is in the order of $100/kWh capacity, if they last 1000 cycles that's $100/MWh delivered*. If you have to size your battery packs for long-term use, be it Dunkelflaute or seasonal or military shenanigans, you don't get to discount that, it's what it is; but if you're only using it for emergencies, then those batteries don't wear down as fast just because you don't go through those cycles so often — how much depends on the duration and frequency of the emergency, and also how well your batteries cope whatever they do in a non-emergency condition.
A global mega-project is obviously going to be more expensive than the mere material costs; even the opportunity cost of what else they could do with the same resources is a huge question, and why it's heavily in the "politics and economics" domain rather than the mere physicality.
But the mere physicality that I'm openly and explicitly and repeatedly acknowledging comes with additional political difficulties**, that's a few hundred billion for a thing that lasts 75-120 years based on other examples; it would carry in the order of a terawatt continuously, 24/7, for about 1/250th - 1/440th the price of batteries: https://www.wolframalpha.com/input?i=%28239+billion+USD%29+%...
* Night has significantly lower demand: you've got 50% power demand, that's 1/3rd of the total daily energy demand from draining batteries, so they only add $33/MWh to the deliverable cost.
** It's frustrating that this combination is still ignored so often when I mention this physical possibility; how can I possibly be clearer that I know that all the other stuff besides actually building it is hard?
Politics cause physical issues. Like things getting bombed. And physical things like distance impact politics - it’s easy to blame a country on the other side of the world vs a neighbor (that people know), or ourselves, for instance.
And why it’s a fantasy is the large capex needs to be spent - while the period of least trust and apparent need exists. And a lot of distance needs to be built before it produces any potential value. And the physical issues. And the politics. And since the benefit comes from the long distances, this isn’t something that is ignorable.
If such a thing already existed, for instance, and countries were using it without abusing it, then being able to justify the cost - and not feel a need for lots of storage - is likely easy. And it’s easy to trust someone when you can see the actual economic incentives for them to behave, and have a track record of it happening. And if everyone is (actually) using it, then it is hard to pick one party (you) off without the whole group ganging up on the offender.
But without such a thing existing, everyone needs to worry about whoever is funding it (if a large country) abusing it to control them, or if everyone is funding it, some random country they don’t like throwing up roadblocks to stop the project actually working for them - and bankrupting them. And since it wouldn’t benefit all countries simultaneously (due to the physical difficulty in constructing such a thing), and it won’t even be actually useful until it is very long already, there is a larger risk of hard feelings and sabotage.
And since they would be dependent on it reaching so far, and countries being so far away playing nicely, the risk on that front is huge.
And the only meaningful way to mitigate that risk is to build lots of storage at a national level.
Which also removes the primary advantages of the project in the first place.
> What exactly do you think is fantastical? Using one year's production of one element? Because I'd already noted that the political aspects are severe enough
Do we stockpile the entire production of aluminium across the entire planet every year, and don't use it just because of politics?
Or do we actually use that yearly production in, you know, actual things that are produced from aluminium?
> only one country would be capable but even then it would be a surprise.
Even the magical country of China would not be able to put the stunt of converting 77 million tons of aluminium into a 40 000 kilometer-long power cable.
> Do we stockpile the entire production of aluminium across the entire planet every year, and don't use it just because of politics?
Aluminium production is not fixed, it varies by demand; I'm suggesting a trivial increase in production given how long power lines last.
> Even the magical country of China would not be able to put the stunt of converting 77 million tons of aluminium into a 40 000 kilometer-long power cable.
It's not harder than anything else that's made of the same stuff.
You know we get the stuff from rocks, right? It doesn't come out of the ground pre-formed as cans and car doors… well, unless you're digging in a landfill.
> How many 44 000 kilometer cables with a cross-section of 1 meter have we produced so far?
Assuming you continue to ignore that that's the aggregate cross section and that the resistance is identical if it's one single cable that size or a hundred of 10cm by 10cm?
And also ignore that all things on this size are normally done by making small things and joining them together?
For the dry bits, dig a trench and pour liquid metal into it.
And for the wet bits, make a 'trench' out of concrete and pour liquid metal into it.
Heck, given how aluminium is produced, you could even dig the trench, fill it with ore, and electrolytically reduce it from the oxide in situ.
But the traditional approach is probably going to be of more interest to investors. Less risk, and you get benefits from the sub-systems without having to complete the whole thing before seeing anything.
> I very much doubt that you do
You doubt that I know the thing I just told you in the line you're quoting to say that you doubt it?
> Not inherent nature, just economics and politics.
I didn't know that night and calm days with no wind are economics and politics
> Wind always blows somewhere, the sun always shines somewhere.
Key word: somewhere. How much wind and sun needs to shine somewhere to cover the needs of, say Europe, on a quiet night? Note that that somewhere will need to cover that somewhere's needs, too.
I'm guessing your proposal is to pump energy from China to the US through the Pacific Ocean?
> Making a global power grid is something which I don't expect anyone will do in the next 15 years
Or ever. You underestimate the size of the Earth
> I'd be least surprised if it's China on a belt-and-road type project.
You know that Europe has a unified power grid, right? And that the US has three major grid regions?
> Batteries are necessary for electric vehicles, and I do expect >50% of land motor vehicles in use worldwide to be electric within 15 years of today; that scale is sufficient to invalidate concerns about intermittent renewable generation in most
wat?
Just because you have batteries in electric vehicles doesn't mean you have enough batteries to provide electricity for a country/continent.
> Note that that somewhere will need to cover that somewhere's needs, too.
And our day provides their night.
It's called "trade".
> you're also likely underestimating how large the area covered by nighttime
Your links are, bluntly, an insult. Everyone here knows the thing they show.
~50% of the planet technically, more than that in practice because local structure; average capacity factor (24 hour including night) of PV is 10%, so I'd be happy to approximate that as a 12 hour day-focused capacity factor of 20%.
> I'm guessing your proposal is to pump energy from China to the US through the Pacific Ocean?
China has an actual proposal for a demo project to go further across the Pacific and connect to Chile.
If that works, they might feel like making it thicker and longer.
> Or ever. You underestimate the size of the Earth
40,000 km, requires 1 m^2 cross section to keep the resistance down to 1 Ω when using Aluminium; this requires just over a year of current global production of aluminium and has a material cost of a few hundred billion dollars at current metal prices.
> You know that Europe has a unified power grid, right? And that the US has three major grid regions?
Which is why it's a political problem. There will always be a place like Texas is now that wants to be difficult and independent even at its own expense.
> Just because you have batteries in electric vehicles doesn't mean you have enough batteries to provide electricity for a country/continent.
There's approximately a billion cars on the road today.
Half a billion cars multiplied by let's say 60 kWh per car is 30 TWh; current global electricity use (which includes daytime) is about 2 TW, that's therefore 15 hours before I include trucks, bikes, busses, etc.
But night use is lower than day use, so in practice the capacity to build that many batteries is also the capacity to build enough batteries for overnight use too.
> I've seen EU night power draw is about 87-107 GW, so 85% of the area of Rhode Island but in Australia
And what's your proposal for getting that energy to the EU from Australia?
> And our day provides their night.
And who's providing your day if you're busy providing someone else's night?
> Everyone here knows the thing they show.
Oh, I doubt it. I keep seeing how people keep pretending that night doesn't exist or how it's easy to provide EU's electricity from Australia during the night.
> 40,000 km, requires 1 m^2 cross section to keep the resistance down to 1 Ω when using Aluminium; this requires just over a year of current global production of aluminium and has a material cost of a few hundred billion dollars at current metal prices.
For a single cable with no redundancy, unknown points of ingress/egress and how to handle that much power in those points of ingress/egress etc. And not pricing in the cost of actually making such a cable, and the cost to actually lay it securely, maintenance costs etc.
> Which is why it's a political problem.
I'm still waiting to hear an actual solution to getting continuous 107GW of power from Australia to the EU. Oh, at the same time keeping up with fluctuating energy demand, frequencies etc. And not a fantasy of redirecting Earth's entire aluminium output to building a single 40 000-kilometre cable.
> Half a billion cars multiplied by let's say 60 kWh per car is 30 TWh; current global electricity use (which includes daytime) is about 2 TW
Something tells me those cars need to be charged somehow, this is probably done by another 40 000 km cable from Australia.
> China has an actual proposal for a demo project to go further across the Pacific and connect to Chile.
Quick googling didn't reveal any such proposal for a ~12000-kilometre cable. The longest proposal is to connect Morocco to Britain with a 3800-kilometre cable, and even that will be an engineering feat of epic proportions if it's not scam/vaporware
> And what's your proposal for getting that energy to the EU from Australia?
Are you trolling? Obviously the same. Why would it possibly be anything other than the same?
> And who's providing your day if you're busy providing someone else's night?
Again, are you taking the piss? Because the only way you're not trolling here, is if you're guilty of what you accuse me of: You underestimate the size of the Earth.
There's plenty of room to do both.
As in, several thousand times over.
> Oh, I doubt it. I keep seeing how people keep pretending that night doesn't exist or how it's easy to provide EU's electricity from Australia during the night.
If you're constantly arguing with a whole host of other people, which is more likely: that they're all idiots who don't know about one of the more fundamental aspects of living on a planet, or that you're misunderstanding their points?
Hint: I did not say "easy", I said "I don't expect anyone will do in the next 15 years" and priced it at hundreds of billions of dollars and even then requiring political collaboration.
15 years is the minimum time in the best case, not the default.
> I'm still waiting to hear an actual solution to getting continuous 107GW of power from Australia to the EU.
You've already got one.
> Oh, at the same time keeping up with fluctuating energy demand, frequencies etc.
DC.
> And not a fantasy of redirecting Earth's entire aluminium output to building a single 40 000-kilometre cable.
Output for one year for something that, from other examples, lasts about a century.
> Something tells me those cars need to be charged somehow, this is probably done by another 40 000 km cable from Australia.
Again, are you trolling? Because that was an alternative to a global power grid.
If you've got those batteries then because the battery is itself the storage system it doesn't care when it gets charged, so it doesn't matter if you can only charge them in your own local daytime as they last through the night and to the next day (and for most people, days plural) anyway.
> This also assumes that all of those cars and busses and trucks are somewhat continuously and consistently plugged into energy grids, with no disruptions, which is not true:
It's not making any such assumption, it's the capacity to build that much battery in the first place, not where the batteries actually are.
> Quick googling didn't reveal any such proposal for a ~12000-kilometre cable.
> The longest proposal is to connect Morocco to Britain with a 3800-kilometre cable, and even that will be an engineering feat of epic proportions if it's not scam/vaporware
Epic, yes.
That's why I'm setting the minimum time for a global grid as pessimistically as 15 years, even if it's made by a government that has a history of using this kind of project as a soft-power projection policy, that already has a project of this kind in planning, and using (1 year/15 years = 6.666…%) ~7% of the aluminium that the world would otherwise be expected to produce in that timeframe.
I was going to write a long reply, but I decided against it.
Because I'm frankly tired of buffoons ignoring the basic realities like physics and geography and pretending their wishful thinking is "trivial to implement".
I mean, you're so detached from reality that you already claim that a vanity project announced by Chile is an actual leaving breathing existing project to which China has already committed.
I don't have the time arguing with people in denial about the basic realities of the world they live in.
Indeed, and likewise (well, the would not have guessed part, I'm not an EE).
It seems a really weird coincidence that the circumference of the planet happens to so closely match the resistivity of aluminium and the unit we use for resistance:
You cannot power up solar panels at night, or wind turbines when there's no wind.
As for nuclear, you don't need to stop them. Instead they provide a base load so you don't need to ridiculously overprovision other power sources. And they are flexible enough to go 50%-100% quickly: https://www.oecd-nea.org/upload/docs/application/pdf/2021-12...
It already takes longer to build new nuclear starting now than build enough renewable + batteries + transmission starting now. So good news! We don't have to "get there soon", we are there already! We have everything we need except the political will to take resources away from fossil fuels and invest them towards a clean energy future.
Nuclear will continue to be built and it should but it is not the thing we need to focus heavily on.
And how long from proposal to breaking ground?
And how many can we build at the same time?
Can we find enough sites or extend existing ones enough?
What problems will we have with water shortages for cooling?
And so on.
We are, thankfully, going to continue to build nuclear all over the world, perhaps most importantly in places like China where they may convert thermal coal plants to nuclear, but it's going to be a relatively small part of the energy mix overall and is not going to solve the doubt that the poster had and my original response was aimed at. Renewables + batteries + transmission, and I should add efficiency improvements, can do it now, at scale, and faster than we can imagine.
The United States, Canada, France, Germany, Japan and the United Kingdom have all reduced their per-capita energy usage since the peak (earliest peak usage was UK in 1973, latest was Canada in 2007), and I think we can argue that quality-of-life (by at least some metrics) has increased since then.
Let's use the UNCTAD definition of developing countries - there are a number of countries which exceed Canada's per-capita energy usage (the highest of those I mention above) - Qatar, UAE, Trinidad and Tobago and Kuwait.
But those are all small countries, the highest per-capita energy usage by a larger nation is Turkmenistan (exceeding all of the European countries mentioned in the parent - UK, France, Germany - and Japan).
But that's not really the point. Doubtless there is a correlation between energy usage and standard of living. But it's not a 1:1, and there are some huge benefits to be gained - e.g. the US (77,028 kWh / person / year) has triple the usage per capital of the UK (28,501 kWh / person / year). Even in Europe, France and Germany could reduce their usage by a quarter to bring it to the UK level.
Also, those high usage small countries (including Trinidad and Turkmenistan) are huge oil producing countries where energy is subsidized to an absurd degree to ‘buy’ population compliance. That energy is coming straight from burning hydrocarbons.
They still are way below the big developed countries, near as I can tell. Sorting by ‘most recent value’, the list is pretty much either ‘huge petrokingdom’, or ‘highly developed nation’ until Estonia/Slovenia/Netherlands at 6.something MWh/yr.
The data does seem to be old though!
I am impressed by Spain’s low usage, but they culturally also minimize things like HVAC - since they were ‘civilized’ long before AC was a thing. They still are about 2x the median.
I’m using „our world in data“ - the link in my first post.
A difference may be that my figures are not just electricity, but „primary energy“, defined as:
„Primary energy includes energy that the end user needs, in the form of electricity, transport and heating, plus inefficiencies and energy that is lost when raw resources are transformed into a usable form.”
Ah, that makes sense - it’s doing the ‘tons of oil energy equivalent’, which those countries are petrostates/petrokingdoms which typically have things like free (or nearly so) gasoline, natural gas, electricity, etc. as part of the gov’t popularity equation.
Which those states need to do that typically because they don’t really care about any actual quality of life improvements, or developing any other parts of the economy.
Because why bother, when you can do those other things more easily, it makes folks dependent on the gov’t, and it avoids things like people being more educated and independent and asking tough questions about those in power.
So I guess I should amend my comment - ‘except for countries where the gov’t is incentivizing free unlimited petro energy due to gov’t policy, quality of life still roughly tracks energy consumption and availability’.
Though then we’d need to argue about what quality of life really means - most Spaniards would top most Americans in physical and mental health in my experience, for example. But Americans definitely have more stuff.
This slightly mad take seems very popular with people who at the very same time seem to hate cheap solar and wind energy and love expensive (and/or nonexistent) nuclear and fossil fuels.
It's a really weird subculture that's grown out of climate denial.
Energy efficiency is a thing. The lowest hanging fruit is EVs and heat pumps that provide 4x more travel or heat than using fossil fuels directly. They're so much more efficient that you could burn the fossils in power plants to generate electricity and still come out ahead. That would be silly though since wind and solar are the cheapest available electricity.
And of course you can power these with nuclear, which would be relevant if they weren't just pretending to like nuclear and really just stalling progress for fossil fuel interests.
> What do you think will happen to the power grid when everyone switches to EVs?
Well the people who run grids keep suggesting that it'll help them make better use of their grid assets and so reduce the average cost of a kWh of electricity to consumers. It is after all a giant fleet of smart batteries.
But you, the person who thinks more energy use is better for society, are apparently making an exception for electricity use, which if it goes up will cause some unspecified calamity?
That doesn't seem very consistent? Well, I suppose it's consistently pro-fossil fuels.
> Well the people who run grids keep suggesting that it'll help them make better use of their grid assets
Do they? Or are they also concerned about the increasing strain on grid infra, especially due to the fact that so many energy sources are now intermittent?
> which if it goes up will cause some unspecified calamity?
It's not an unspecified calamity. Failure modes for grids are known.
> Well, I suppose it's consistently pro-fossil fuels.
Funny how you accuse others of making mad takes while keeping doing one mad take after another
> Yes, it is. There's a limit to how efficient you can go though
This is why reducing energy use needs to be in addition to everything else: solar, nuclear, wind, batteries, the works. Some of the R&D won't work out as planned; some solutions will only work in specific situations. That's okay, as long as we don't have all our eggs in one basket.
The orders of magnitude needed for storage are not competitive with nuclear in any reasonable scenario.
Nuclear exists and is proven technology. Battery storage at the capacity and price scale required are not. That remains true even if you want absolutely to believe we can't build nuclear reactors faster than in 10 years, and we almost certainly can by starting mass producing them again.
> However, the company warned in Tuesday’s report that the “termination” of some clean energy projects during 2023 had pushed down the amount of renewables it had access to.
It's true. If you build all your infrastructure n times over you could have solar panels powering (some of?) each infrastructure full infrastructure location. But would that work out better?
By the decade is gone that any new nuclear project has been commissioned and built, the rest of the world is running on solar + battery storage at a tenth of the price.
Your first graph shows diminishing returns for both battery storage and solar. It's also not clear whether that's the marginal cost or includes the capital cost to build the structures and all the infrastructure.
To compare nuclear against solar you have to compare it against the sum of the cost of solar plus battery storage.
Investing in solar/wind/nuclear energy sources, storage, and research for improving all of the above seems like the best strategic option for at least the next decade.
Yes we need nuclear but these statements often feels like attempt to sweeping the problem under the rug. Because nuclear takes time to build. And this is not just about nuclear. Exxon report from early 80s stated that to cover energy demand nuclear or renewables would need 50 years to catch up. I guess demand is even bigger now.
This proposition still fits inside a "line goes up" vision, and it can't go up forever.
More nuclear doesn't come without additional risk nor cost, nor waste. I don't claim to have a solution, but it feels like jumping into a volcano to escape from the bear. There sure must be some way to deal with the bear instead.
I just don't understand the implied argument here at all. We can do something that results in productivity gains, which people want or they wouldn't pay for it, but we don't want private companies to do this because we have decided an increase energy use is a bad thing. Isn't the obvious path to continue to grow towards more efficient energy sources rather than trying to suppress progress?
You either have growth or limited emissions. Progress to more sustainable or more efficient energy sources til now just led to covering part of ever increasing demand.
Coal is still rising as well.
Interesting considering that Google said it had a strategy that would ensure AI consumed less power. It was based on the 4Ms: model, machine, mechanization, and mapping.
I wonder to which extent data centers are built already or will be built close to where renewables are readily available, to relieve the burden of transporting energy over long distances. Is it cheaper to build fibre optics to a remote location than it is to build power lines?
does anyone know if this is due to fleet growth (ie the existing commitment to going to zero emissions is still happening, quickly, but growth is so high it's swamping that in absolute terms) or has Urs gone back on his promise and is using emission-y energy?
It's not the first time I've seen it claimed that AI will magically help "propel climate solutions". But I've never seen a hint at how that would work.
It's just magical thinking, imagining some kind of heretofore unknown technology that will solve everything quickly and cheaply in the future.
The sad truth is we already have most of the technology needed to fight climate change. The technology could always use incremental improvement, sure, but the main challenge to be solved is not technological, it's political and economical. It's a question of actually adopting the solutions we already have.
The sad truth is that it’s already too late to keep the status quo. In the next 20 years the living conditions will worsen considerably for the majority of humans.
What we can do now is make sure humanity as a whole has some chance of survival.
The only thing I can see doing that is cheap fusion power, and that doesn’t seem like something that AI can help much with, at least not the uses I’ve seen it put to. It doesn’t seem great at those types of problems, at least LLM doesn’t seem to be.
You could hypothetically apply the same line of reasoning used for cryptocurrency mining: AI training is dispatchable load. Now all this AI investment can go into funding the construction of a lot of new renewable generation, and it uses that power most of the time, but now when you get the week out of the year that it's too cloudy or still and renewable output is only 20% of normal, you turn off the AI training machines for that week and do your training when the sun comes back out, leaving that extra generation capacity for the rest of the grid. Which allows you to decommission a fossil fuel power plant.
Or you can go the other way. The data center can't afford to be down, so they have several MW worth of diesel generators on site in case the grid is down, but the same can be used to supply the grid during that week when renewable generation is low. The more of the grid is that kind of data center, the more viable it is to shutter always-on fossil fuel power plants in favor of renewable generation that can lean on running the emergency generators a week out of the year.
It's hype. Just a few short years ago various crypto scams claimed to solve climate issues, too (and to lead the revolution in renewables).
if there's no climate, there's no climate problems.
Maybe AI is already sentient, and instead of nuking the world they just trick people to think they are not sentient, so people train more neural networks and destroy the planet themselves ;D.
On a more serious note: "propel climate solutions" is a weird take for me regarding AI considering the resources used.
I'm not sure though: Do you think google search and all the infrastructure/servers required for it, will use more power etc. than an AI, including its training/retraining periodically and infra needed to host it as a service for people?
I'd be interested to see for example if there is any improvement at all, if you would 'swap google search for chatgpt' - i know this is not realistic or even its purpose, but it's what a lot of laymen will interpret the GPT stuf as i guess.
Would we use less energy etc. if we'd switch the legacy stuff like google search to more modern AI based apporaches?
Now in the time of transition, we're screwed and power consumption would in my mind logically double or more, as we need to run these things side-by-side to develop AI to make it useful.
15 hours ago. 107 comments: https://news.ycombinator.com/item?id=40859993
Better to be the summary or report directly: https://blog.google/outreach-initiatives/sustainability/2024...
This is why we need nuclear. Humanity will always find a new expensive toy to play with, AI being the latest, and expecting large corporate entities to scale down their operations for the sake of future sales is like trying to bargain with a bear.
Nuclear isn't infinite and uranium mines are often in politically unstable countries. Niger's military junta just forbad Orano from exploiting a mine, for example. What is the French government supposed to do now? Declare war for uranium? Import more uranium from Russia?
As a reminder, here is where the US's uranium came from in 2022: Canada (27%), Kazakhstan (25%), Russia (12%), Uzbekistan (11%), Australia (9%), six other countries (16%) (Germany, Malawi, Namibia, Niger, South Africa, and the United States). https://www.eia.gov/energyexplained/nuclear/where-our-uraniu...
The main problem with it isnt with the fuel. It's not even the small chance of a $1 trillion-to-cleanup Fukushima style cleanup event. It's just horrendously expensive and takes forever to build.
Nonetheless it'll keep getting built where solar/wind/storage would be cheaper and quicker because they've got very slick PR and because the military industrial complex demands a ready supply of nuclear skills and a supply chain.
> horrendously expensive
How did France manage to power their country with it, then?
Lavish subsidies. Even then they had to nationalize EDF because it was going bankrupt after posting historically unprecedented losses.
The problem is that stuff like this:
https://en.wikipedia.org/wiki/ASTRID_(reactor)
and this:
https://reneweconomy.com.au/french-nuclear-giant-scraps-smr-...
keeps happening. The last paragraph explains how, in spite of these losses, how to keep the power cheap:
>The federal Coalition has argued that nuclear might be expensive to build, but will deliver cheaper power to consumers. It has not explained how, but it has said that its reactors would be government owned, suggesting that – like France and Ontario – the costs would be borne by taxpayers and the supply of power to customers would be heavily subsidised.
Totally wrong. In totality nuclear was an utterly and completely amazing deal.
The whole built up of the COMPLETE green grid cost the tax payer very little as the whole buildup went on the balance sheet of the EDF. And then it was paid down over years and years. During all this time France had very low stable energy cost.
And this is while the French state in the last 2 decades abused the system as if it was a magical giving tree. Forcing them to sell energy very cheaply. Forcing them to sell cheap reliable energy in bulk to fossil fuel companies (subsidizing them) and then buying it back at much higher prices when it became clear they sold to much. Forcing them to invest in solar and dispatch at cost (destroying base load nuclear economics).
Then the French government in its complete delusion essentially decided that they would be cool like coal producing Germany and announced a fantasy 'transition to renewables' that had literally 0.0% chance of actually happening.
They basically delayed all maintenance because the reactors would be replaced soon anyway (yeah right).
And this does not even take into account the gigantic amount of money France wont have to spend to do a green transition now. Comparably other countries had higher energy prices for 40 years, and have yet to transition a green grid.
AND as a final coup de grace France had much lower medical cost because they didn't have decades of breathing coal (except the coal coming from Germany of course). Talk about externalizes.
By any calculation, French nuclear program has been a complete and utter success.
Even with extreme conservative calculation on reactor cost Germany. Starting in 2000, Germany would have had a way cheaper transition to green energy had they much mass built nuclear like France did. If you actually apply proper economics of scale (as observed during the French mass build) the cost would be a tiny fraction what Germany has spent and will continue to spend between buildup of wind/solar/battery, the high energy prices and the massive home and industrial energy subsidizes.
> ASTRID_(reactor)
Investing in future reactors is not a bad idea. In fact, Superphénix was already built and working, most of the bugs were fixed and it was ready for more broad rollout. But instead of building many more, short sighted politicians (who seem to share your ideology) killed the project.
One of the main problems with the French nuclear industry in general, and the high cost now is that after the first generation, the next generation went anti-nuclear. They stopped building more plants, stopped innovating and start treating what the previous generation had built up as a 'problem that had to be overcome' rather then an incredibly resource.
This lead to a situation where most of the experts retired and France lost a huge amount of knowledge. Only now they realize that transitioning away to magical solar/wind isn't nearly as easy as they had thought.
This is categorically false. France has always thrown massive amounts of subsidies at its nuclear program since day 1. It was quite up front from the very beginning that it wasn't about the environment, public health or saving money - it was about energy independence. It was a good move for the environment in 1973 when solar panels were expensive but that isn't why they did it.
Germany's renewables transition is a massive success. It didn't require writing off titanic, gargantuan levels of wasted capital investment like France did with stuff like ASTRID. Their energy transition has had VASTLY lower capex than France's.
It would make sense for France to follow in their footsteps as their nuclear plants start to age out, but it seems like they'd rather waste epic amounts of taxpayer cash on building new plants which still aren't enough to keep up with demand while they play accounting games to try and shift the losses around (the same kind of games which led Finland to sue EDF over its cost overruns).
It’s horrendously expensive because we insist on doing it in the shittiest, least efficient way possible. Like California’s high speed rail (frankly).
It doesnt take forever to build. Its the fastest way to scale clean energy.
If Germany in 2000 had invested in nuclear they could be basically done by now and for far less money, even with incredibly conservative assumsions.
Just look at the build rate UAE and South Korea were able to achieve. Apply that to Germany, even without any effects from scale, and the numbers look fantastic.
And its very unlikely that you are gone have Fukishima and even if you do, most of that money isnt actually needed.
Nuclear isnt getting built much compared to solar/wind in what world do you live? Its the exact opposite.
Uh, the US could easily mine more than all those other countries combined. [https://en.m.wikipedia.org/wiki/Uranium_mining_in_the_United...]
And most of those countries are not politically unstable - unless we count Trudeau’s declining popularity, or Australia and Germany’s need for excessive beer drinking.
For various reasons it just went through a period it didn’t want to. Though prices have recovered enough, mines are re-opening again.
Kazakhstan alone typically exports more than 20x a year than any given year of total US production.
The US has uranium deposits, yes, but they are relatively low grade and sparse in comparison with the major Uranium producing hotspots about the globe.
Typically the US uses in situ leach mining to dissolve sparse uranium salts from (say) breccia pipes.
Other countries mine faces (or bore through and extract) dense uranium rock ores (much more uranium returned per volume).
Oddly, stability has not been an issue with uranium extraction, the best mines ever in the Congo were continuously extracted throughout the height of the Cold Wat while secret proxy wars were funded and fought all about them to keep the region destabilised for ease of resource extraction to US | French companies.
Today Kazakhstan sells to the US via Rosatom in an agreement ratified by past POTUS.
Look at historic US production rates. [https://world-nuclear.org/information-library/nuclear-fuel-c...]
See those 60’s-80’s numbers?
The US historically has done standard hard rock Uranium mining. It’s all over the 4 corners part of the US. High grade ore too.
The reason it stopped was due to environmental regs/health issues - and because other countries would do it for cheaper. Not because it can’t do it at scale. The little remaining may be doing in-situ leaching, but that is a tiny footnote for the vast array of available uranium mines in the US.
There is no actual strategic mining threat here, any more than there is one for bauxite or iron ore, etc.
If someone tried to restrict US uranium supply, it would be a couple years at most before we’d be squishing everyone else in output again.
Complete nonsense. Practically speaking there is no limitation. Its only that we have minimally invested in using it well.
Even with the current suboptimal use we have plenty
We already have fusion and it's super safe and cheap.
We just need to install more panels to harvest the incoming radiation.
Solar and wind are by their very inherent nature intermittent sources of power
Not inherent nature, just economics and politics.
Wind always blows somewhere, the sun always shines somewhere.
Making a global power grid is something which I don't expect anyone will do in the next 15 years, but if anyone does I'd be least surprised if it's China on a belt-and-road type project.
Batteries are necessary for electric vehicles, and I do expect >50% of land motor vehicles in use worldwide to be electric within 15 years of today; that scale is sufficient to invalidate concerns about intermittent renewable generation in most (but not all) places even without a global grid.
And how far of a cable would you need to get 24/7 solar power? Because based on my math, it would have to cross a lot of timezones - including ones with a lot of people that would be happy to chop that cable if it meant screwing with the US.
40,000 km will always be enough.
A single cable is a bad idea for many reasons, but "it might get cut" isn't one of them, as at this scale it's the kind of thing that the whole world is using, so (0) nobody actually benefits from cutting it, (1) it's carrying order of a terrawatt at any given moment, so if you made it a single cable the magnetic field alone will mess with your cutting equipment, and (2) for reasonable resistance it's a cross section of about a square meter (1 Ω and ~1 year global aluminium production for that cross section at 40,000 km), so even a literal detonation of a Little Boy nuke 500 meters from it wouldn't make a dent.
A country like Russia will always benefit from cutting such a cable.
Anything else, like the rest of this post, is pure fantasy.
If the cable (or one of them: again it's bad to make it just one, just physically spreading out that 1m^2 into a hundred cables of 10cm by 10cm makes it a lot less risky) goes through Russia, and they cut it, the effect is that immediately Russia stops trading energy.
At best, this is a loss of income. At worst, it means their own lights go off because now they aren't getting anyone else's electricity in their own night and their own winter.
What exactly do you think is fantastical? Using one year's production of one element? Because I'd already noted that the political aspects are severe enough that it's unlikely for anyone to build at this scale any time soon, and that only one country would be capable but even then it would be a surprise.
I find it fantastical that anyone in the geopolitical environment would put themselves at the (near instantaneous) mercy of someone that far away - physically and politically.
We’ve already seen the games played with hydrocarbon pipelines, and hydrocarbons are easy to store in large quantities, which greatly reduces the risk of someone playing ‘got you!’ on something that literally is a matter of life and death.
It still - clearly - happens, but when it does it isn’t as urgent as freezing to death tonight if you don’t do what they want. Unless they’re also bombing your storage tanks, which does happen, but is a significant escalation over ‘oops, we hit the power cable with an anchor again. Sorry!’.
For the Russia example, in the situations Russia wouldn’t want to cut it - the US and China would love to! Or have a proxy do it for them.
And if Russia has enough battery storage to last a week or two (minimum to outlast a storm; and also more than adequate to deal with daily insolation variation), they don’t have much of a need for these cables either, correct? At least most of the time.
In fact, in those cases the cables would actually present a bit of a problem, as they would potentially cause economic/trade/currency balances, similar to the Oil trade, and make it less economic to actually have the appropriate degree of local storage and production necessary.
The same issues pop up even regionally - Europe has a history of not always being super peaceful, so if, say, the UK wants to run a HVDC line to Spain, they have to think long and hard about the implications. As does Spain.
Because if Germany does that they did a few generations ago, that cable isn’t going to last long, and if the UK doesn’t have adequate safe alternatives, then they’re going to have a really hard time fighting back. And if Spain decides to raise rates to ruinous levels of the UK ‘brexits’, what leverage does the UK have?
Let alone not freezing to death. And Spain needs to consider they might be broke in such a situation, and having a hard time paying their own bills.
And if they do have adequate alternatives, then why is it a big deal to have the cable? What does it actually offer?
As to if it’s a single cable or a million of them or a lattice or whatever - the big question is ‘why?’ when the rest of these factors come into play. The materials cost to build and maintain it doesn’t help either.
That all sounds like politics? If so, that's a point we agree on :)
None of that seems like physical issues, despite you mentioning that on the first paragraph.
> For the Russia example, in the situations Russia wouldn’t want to cut it - the US and China would love to! Or have a proxy do it for them.
If they have alternatives, otherwise they're all in the same boat whose name is "interdependence".
China's also the only country that seems to have any even slightly serious interest in attempting to build anything close to such a grid in the first place, as part of their own soft-power projection, and are the only country producing enough aluminium to attempt it anyway — half of all current global production is factories they switched on since 2006: https://international-aluminium.org/statistics/primary-alumi...
Nobody else but China will bother in the foreseeable future, and even China might not. (Which is yet another rephrasing of the same point I've been making throughout).
> And if Russia has enough battery storage to last a week or two (minimum to outlast a storm; and also more than adequate to deal with daily insolation variation), they don’t have much of a need for these cables either, correct? At least most of the time.
Sure, but a week of storage is at least* 14 times more expensive than overnight storage.
If you see your enemies building up a lot of storage without any clear reason, you know Something Is Up, just like if you see them amassing tanks on your border.
If they have a clear reason, "there is no power grid in the first place" being one, then it's harder to know they're doing something odd.
> And if they do have adequate alternatives, then why is it a big deal to have the cable? What does it actually offer?
Lower price.
I'm expecting batteries to get cheaper, but the current price is in the order of $100/kWh capacity, if they last 1000 cycles that's $100/MWh delivered*. If you have to size your battery packs for long-term use, be it Dunkelflaute or seasonal or military shenanigans, you don't get to discount that, it's what it is; but if you're only using it for emergencies, then those batteries don't wear down as fast just because you don't go through those cycles so often — how much depends on the duration and frequency of the emergency, and also how well your batteries cope whatever they do in a non-emergency condition.
A global mega-project is obviously going to be more expensive than the mere material costs; even the opportunity cost of what else they could do with the same resources is a huge question, and why it's heavily in the "politics and economics" domain rather than the mere physicality.
But the mere physicality that I'm openly and explicitly and repeatedly acknowledging comes with additional political difficulties**, that's a few hundred billion for a thing that lasts 75-120 years based on other examples; it would carry in the order of a terawatt continuously, 24/7, for about 1/250th - 1/440th the price of batteries: https://www.wolframalpha.com/input?i=%28239+billion+USD%29+%...
* Night has significantly lower demand: you've got 50% power demand, that's 1/3rd of the total daily energy demand from draining batteries, so they only add $33/MWh to the deliverable cost.
** It's frustrating that this combination is still ignored so often when I mention this physical possibility; how can I possibly be clearer that I know that all the other stuff besides actually building it is hard?
?
Politics cause physical issues. Like things getting bombed. And physical things like distance impact politics - it’s easy to blame a country on the other side of the world vs a neighbor (that people know), or ourselves, for instance.
And why it’s a fantasy is the large capex needs to be spent - while the period of least trust and apparent need exists. And a lot of distance needs to be built before it produces any potential value. And the physical issues. And the politics. And since the benefit comes from the long distances, this isn’t something that is ignorable.
If such a thing already existed, for instance, and countries were using it without abusing it, then being able to justify the cost - and not feel a need for lots of storage - is likely easy. And it’s easy to trust someone when you can see the actual economic incentives for them to behave, and have a track record of it happening. And if everyone is (actually) using it, then it is hard to pick one party (you) off without the whole group ganging up on the offender.
But without such a thing existing, everyone needs to worry about whoever is funding it (if a large country) abusing it to control them, or if everyone is funding it, some random country they don’t like throwing up roadblocks to stop the project actually working for them - and bankrupting them. And since it wouldn’t benefit all countries simultaneously (due to the physical difficulty in constructing such a thing), and it won’t even be actually useful until it is very long already, there is a larger risk of hard feelings and sabotage.
And since they would be dependent on it reaching so far, and countries being so far away playing nicely, the risk on that front is huge.
And the only meaningful way to mitigate that risk is to build lots of storage at a national level.
Which also removes the primary advantages of the project in the first place.
> What exactly do you think is fantastical? Using one year's production of one element? Because I'd already noted that the political aspects are severe enough
Do we stockpile the entire production of aluminium across the entire planet every year, and don't use it just because of politics?
Or do we actually use that yearly production in, you know, actual things that are produced from aluminium?
> only one country would be capable but even then it would be a surprise.
Even the magical country of China would not be able to put the stunt of converting 77 million tons of aluminium into a 40 000 kilometer-long power cable.
> Do we stockpile the entire production of aluminium across the entire planet every year, and don't use it just because of politics?
Aluminium production is not fixed, it varies by demand; I'm suggesting a trivial increase in production given how long power lines last.
> Even the magical country of China would not be able to put the stunt of converting 77 million tons of aluminium into a 40 000 kilometer-long power cable.
It's not harder than anything else that's made of the same stuff.
You know we get the stuff from rocks, right? It doesn't come out of the ground pre-formed as cans and car doors… well, unless you're digging in a landfill.
> I'm suggesting a trivial increase in production
A "trivial" doubling of production
> It's not harder than anything else that's made of the same stuff.
How many 44 000 kilometer cables with a cross-section of 1 meter have we produced so far?
> You know we get the stuff from rocks, right? It doesn't come out of the ground pre-formed
I do know that. I very much doubt that you do
> A "trivial" doubling of production
More like 7%, if there's the political will to build it in only 15 years.
1 year production / 15 years = 6.666… %
And 50% of current global aluminium production comes from factories that specifically China switched on since 2006: https://international-aluminium.org/statistics/primary-alumi...
> How many 44 000 kilometer cables with a cross-section of 1 meter have we produced so far?
Assuming you continue to ignore that that's the aggregate cross section and that the resistance is identical if it's one single cable that size or a hundred of 10cm by 10cm?
And also ignore that all things on this size are normally done by making small things and joining them together?
For the dry bits, dig a trench and pour liquid metal into it.
And for the wet bits, make a 'trench' out of concrete and pour liquid metal into it.
Heck, given how aluminium is produced, you could even dig the trench, fill it with ore, and electrolytically reduce it from the oxide in situ.
But the traditional approach is probably going to be of more interest to investors. Less risk, and you get benefits from the sub-systems without having to complete the whole thing before seeing anything.
> I very much doubt that you do
You doubt that I know the thing I just told you in the line you're quoting to say that you doubt it?
> Not inherent nature, just economics and politics.
I didn't know that night and calm days with no wind are economics and politics
> Wind always blows somewhere, the sun always shines somewhere.
Key word: somewhere. How much wind and sun needs to shine somewhere to cover the needs of, say Europe, on a quiet night? Note that that somewhere will need to cover that somewhere's needs, too.
Note: you're also likely underestimating how large the area covered by nighttime is. Here are two examples: https://www.timeanddate.com/worldclock/sunearth.html?iso=202... and https://www.timeanddate.com/worldclock/sunearth.html?iso=202...
I'm guessing your proposal is to pump energy from China to the US through the Pacific Ocean?
> Making a global power grid is something which I don't expect anyone will do in the next 15 years
Or ever. You underestimate the size of the Earth
> I'd be least surprised if it's China on a belt-and-road type project.
You know that Europe has a unified power grid, right? And that the US has three major grid regions?
> Batteries are necessary for electric vehicles, and I do expect >50% of land motor vehicles in use worldwide to be electric within 15 years of today; that scale is sufficient to invalidate concerns about intermittent renewable generation in most
wat?
Just because you have batteries in electric vehicles doesn't mean you have enough batteries to provide electricity for a country/continent.
> I didn't know that night and calm days with no wind are economics and politics
Planet is round. Your winter solstice midnight is someone else's midsummer midday.
Convincing everyone to join up, that's economics and politics.
> How much wind and sun needs to shine somewhere to cover the needs of, say Europe, on a quiet night?
I've seen EU night power draw is about 87-107 GW, so 85% of the area of Rhode Island but in Australia: http://www.wolframalpha.com/input/?i=107%20GW%20%2F%20%281%2...
> Note that that somewhere will need to cover that somewhere's needs, too.
And our day provides their night.
It's called "trade".
> you're also likely underestimating how large the area covered by nighttime
Your links are, bluntly, an insult. Everyone here knows the thing they show.
~50% of the planet technically, more than that in practice because local structure; average capacity factor (24 hour including night) of PV is 10%, so I'd be happy to approximate that as a 12 hour day-focused capacity factor of 20%.
> I'm guessing your proposal is to pump energy from China to the US through the Pacific Ocean?
China has an actual proposal for a demo project to go further across the Pacific and connect to Chile.
If that works, they might feel like making it thicker and longer.
> Or ever. You underestimate the size of the Earth
40,000 km, requires 1 m^2 cross section to keep the resistance down to 1 Ω when using Aluminium; this requires just over a year of current global production of aluminium and has a material cost of a few hundred billion dollars at current metal prices.
> You know that Europe has a unified power grid, right? And that the US has three major grid regions?
Which is why it's a political problem. There will always be a place like Texas is now that wants to be difficult and independent even at its own expense.
> Just because you have batteries in electric vehicles doesn't mean you have enough batteries to provide electricity for a country/continent.
There's approximately a billion cars on the road today.
Half a billion cars multiplied by let's say 60 kWh per car is 30 TWh; current global electricity use (which includes daytime) is about 2 TW, that's therefore 15 hours before I include trucks, bikes, busses, etc.
But night use is lower than day use, so in practice the capacity to build that many batteries is also the capacity to build enough batteries for overnight use too.
> I've seen EU night power draw is about 87-107 GW, so 85% of the area of Rhode Island but in Australia
And what's your proposal for getting that energy to the EU from Australia?
> And our day provides their night.
And who's providing your day if you're busy providing someone else's night?
> Everyone here knows the thing they show.
Oh, I doubt it. I keep seeing how people keep pretending that night doesn't exist or how it's easy to provide EU's electricity from Australia during the night.
> 40,000 km, requires 1 m^2 cross section to keep the resistance down to 1 Ω when using Aluminium; this requires just over a year of current global production of aluminium and has a material cost of a few hundred billion dollars at current metal prices.
For a single cable with no redundancy, unknown points of ingress/egress and how to handle that much power in those points of ingress/egress etc. And not pricing in the cost of actually making such a cable, and the cost to actually lay it securely, maintenance costs etc.
> Which is why it's a political problem.
I'm still waiting to hear an actual solution to getting continuous 107GW of power from Australia to the EU. Oh, at the same time keeping up with fluctuating energy demand, frequencies etc. And not a fantasy of redirecting Earth's entire aluminium output to building a single 40 000-kilometre cable.
> Half a billion cars multiplied by let's say 60 kWh per car is 30 TWh; current global electricity use (which includes daytime) is about 2 TW
Something tells me those cars need to be charged somehow, this is probably done by another 40 000 km cable from Australia.
This also assumes that all of those cars and busses and trucks are somewhat continuously and consistently plugged into energy grids, with no disruptions, which is not true: https://publications.lib.chalmers.se/records/fulltext/179114...
Edit:
> China has an actual proposal for a demo project to go further across the Pacific and connect to Chile.
Quick googling didn't reveal any such proposal for a ~12000-kilometre cable. The longest proposal is to connect Morocco to Britain with a 3800-kilometre cable, and even that will be an engineering feat of epic proportions if it's not scam/vaporware
> And what's your proposal for getting that energy to the EU from Australia?
Are you trolling? Obviously the same. Why would it possibly be anything other than the same?
> And who's providing your day if you're busy providing someone else's night?
Again, are you taking the piss? Because the only way you're not trolling here, is if you're guilty of what you accuse me of: You underestimate the size of the Earth.
There's plenty of room to do both.
As in, several thousand times over.
> Oh, I doubt it. I keep seeing how people keep pretending that night doesn't exist or how it's easy to provide EU's electricity from Australia during the night.
If you're constantly arguing with a whole host of other people, which is more likely: that they're all idiots who don't know about one of the more fundamental aspects of living on a planet, or that you're misunderstanding their points?
Hint: I did not say "easy", I said "I don't expect anyone will do in the next 15 years" and priced it at hundreds of billions of dollars and even then requiring political collaboration.
15 years is the minimum time in the best case, not the default.
> I'm still waiting to hear an actual solution to getting continuous 107GW of power from Australia to the EU.
You've already got one.
> Oh, at the same time keeping up with fluctuating energy demand, frequencies etc.
DC.
> And not a fantasy of redirecting Earth's entire aluminium output to building a single 40 000-kilometre cable.
Output for one year for something that, from other examples, lasts about a century.
> Something tells me those cars need to be charged somehow, this is probably done by another 40 000 km cable from Australia.
Again, are you trolling? Because that was an alternative to a global power grid.
If you've got those batteries then because the battery is itself the storage system it doesn't care when it gets charged, so it doesn't matter if you can only charge them in your own local daytime as they last through the night and to the next day (and for most people, days plural) anyway.
> This also assumes that all of those cars and busses and trucks are somewhat continuously and consistently plugged into energy grids, with no disruptions, which is not true:
It's not making any such assumption, it's the capacity to build that much battery in the first place, not where the batteries actually are.
> Quick googling didn't reveal any such proposal for a ~12000-kilometre cable.
15,000 km, not 12,000:
https://markets.businessinsider.com/news/stocks/building-a-2...
> The longest proposal is to connect Morocco to Britain with a 3800-kilometre cable, and even that will be an engineering feat of epic proportions if it's not scam/vaporware
Epic, yes.
That's why I'm setting the minimum time for a global grid as pessimistically as 15 years, even if it's made by a government that has a history of using this kind of project as a soft-power projection policy, that already has a project of this kind in planning, and using (1 year/15 years = 6.666…%) ~7% of the aluminium that the world would otherwise be expected to produce in that timeframe.
I was going to write a long reply, but I decided against it.
Because I'm frankly tired of buffoons ignoring the basic realities like physics and geography and pretending their wishful thinking is "trivial to implement".
I mean, you're so detached from reality that you already claim that a vanity project announced by Chile is an actual leaving breathing existing project to which China has already committed.
I don't have the time arguing with people in denial about the basic realities of the world they live in.
Adieu.
Too lazy to do the calculations my self: How long would a "wire" with 1 m² cross section have to be to have 1 Ohm resistance?
40,000 km, almost exactly
Wow, that's not at all what I would have guessed. I'm an EE engineer who didn't know that this is actually (kind of) feasible.
Indeed, and likewise (well, the would not have guessed part, I'm not an EE).
It seems a really weird coincidence that the circumference of the planet happens to so closely match the resistivity of aluminium and the unit we use for resistance:
https://www.wolframalpha.com/input?i=aluminium+resistivity+*...
It's not feasible. You don't need to be an EE engineer to know that.
You can not stop a nuclear reactor to fit the grid needs.
You cannot power up solar panels at night, or wind turbines when there's no wind.
As for nuclear, you don't need to stop them. Instead they provide a base load so you don't need to ridiculously overprovision other power sources. And they are flexible enough to go 50%-100% quickly: https://www.oecd-nea.org/upload/docs/application/pdf/2021-12...
"You cannot power up solar panels at night, or wind turbines when there's no wind."
Amazing insight.
This is why we need to accelerate renewables and energy storage.
If we ever make large enough batteries to satisfy our on-demand energy whims, then I'm 100% with you.
I just have doubt that we will get there soon with battery tech, and nuclear seems like a rational dropin solution until then.
It already takes longer to build new nuclear starting now than build enough renewable + batteries + transmission starting now. So good news! We don't have to "get there soon", we are there already! We have everything we need except the political will to take resources away from fossil fuels and invest them towards a clean energy future.
Nuclear will continue to be built and it should but it is not the thing we need to focus heavily on.
Japan's mean time to build nuclear is under 5 years, I believe.
And how long from proposal to breaking ground? And how many can we build at the same time? Can we find enough sites or extend existing ones enough? What problems will we have with water shortages for cooling? And so on.
We are, thankfully, going to continue to build nuclear all over the world, perhaps most importantly in places like China where they may convert thermal coal plants to nuclear, but it's going to be a relatively small part of the energy mix overall and is not going to solve the doubt that the poster had and my original response was aimed at. Renewables + batteries + transmission, and I should add efficiency improvements, can do it now, at scale, and faster than we can imagine.
https://archive.is/s9dQh
We need all of the above. Including reducing overall energy use. We cannot afford to leave any option on the table.
There are no rich countries that are also energy poor. You cannot reduce energy use without reducing quality of life.
The United States, Canada, France, Germany, Japan and the United Kingdom have all reduced their per-capita energy usage since the peak (earliest peak usage was UK in 1973, latest was Canada in 2007), and I think we can argue that quality-of-life (by at least some metrics) has increased since then.
https://ourworldindata.org/grapher/per-capita-energy-use?tab...
And where do they stand worldwide? Still higher than all the third world countries eh?
Not really.
Let's use the UNCTAD definition of developing countries - there are a number of countries which exceed Canada's per-capita energy usage (the highest of those I mention above) - Qatar, UAE, Trinidad and Tobago and Kuwait.
But those are all small countries, the highest per-capita energy usage by a larger nation is Turkmenistan (exceeding all of the European countries mentioned in the parent - UK, France, Germany - and Japan).
But that's not really the point. Doubtless there is a correlation between energy usage and standard of living. But it's not a 1:1, and there are some huge benefits to be gained - e.g. the US (77,028 kWh / person / year) has triple the usage per capital of the UK (28,501 kWh / person / year). Even in Europe, France and Germany could reduce their usage by a quarter to bring it to the UK level.
Where are you getting your data?
For instance, I’m showing about 2.6 MWh/ye per capita for Turkmenistan [https://data.worldbank.org/indicator/EG.USE.ELEC.KH.PC?locat...], below the world median of 3.1MWh/yr.
Also, those high usage small countries (including Trinidad and Turkmenistan) are huge oil producing countries where energy is subsidized to an absurd degree to ‘buy’ population compliance. That energy is coming straight from burning hydrocarbons.
They still are way below the big developed countries, near as I can tell. Sorting by ‘most recent value’, the list is pretty much either ‘huge petrokingdom’, or ‘highly developed nation’ until Estonia/Slovenia/Netherlands at 6.something MWh/yr.
The data does seem to be old though!
I am impressed by Spain’s low usage, but they culturally also minimize things like HVAC - since they were ‘civilized’ long before AC was a thing. They still are about 2x the median.
I’m using „our world in data“ - the link in my first post.
A difference may be that my figures are not just electricity, but „primary energy“, defined as:
„Primary energy includes energy that the end user needs, in the form of electricity, transport and heating, plus inefficiencies and energy that is lost when raw resources are transformed into a usable form.”
Ah, that makes sense - it’s doing the ‘tons of oil energy equivalent’, which those countries are petrostates/petrokingdoms which typically have things like free (or nearly so) gasoline, natural gas, electricity, etc. as part of the gov’t popularity equation.
Which those states need to do that typically because they don’t really care about any actual quality of life improvements, or developing any other parts of the economy.
Because why bother, when you can do those other things more easily, it makes folks dependent on the gov’t, and it avoids things like people being more educated and independent and asking tough questions about those in power.
So I guess I should amend my comment - ‘except for countries where the gov’t is incentivizing free unlimited petro energy due to gov’t policy, quality of life still roughly tracks energy consumption and availability’.
Though then we’d need to argue about what quality of life really means - most Spaniards would top most Americans in physical and mental health in my experience, for example. But Americans definitely have more stuff.
This slightly mad take seems very popular with people who at the very same time seem to hate cheap solar and wind energy and love expensive (and/or nonexistent) nuclear and fossil fuels.
It's a really weird subculture that's grown out of climate denial.
Energy efficiency is a thing. The lowest hanging fruit is EVs and heat pumps that provide 4x more travel or heat than using fossil fuels directly. They're so much more efficient that you could burn the fossils in power plants to generate electricity and still come out ahead. That would be silly though since wind and solar are the cheapest available electricity.
And of course you can power these with nuclear, which would be relevant if they weren't just pretending to like nuclear and really just stalling progress for fossil fuel interests.
> ho at the very same time seem to hate cheap solar and wind energy and love expensive (and/or nonexistent) nuclear and fossil fuels.
Now this is a mad take
> Energy efficiency is a thing.
Yes, it is. There's a limit to how efficient you can go though
> The lowest hanging fruit is EVs and heat pumps
What do you think will happen to the power grid when everyone switches to EVs?
> That would be silly though since wind and solar are the cheapest available electricity.
They are also intermittent sources of electricity.
> if they weren't just pretending to like nuclear and really just stalling progress for fossil fuel interests.
Another mad take.
All in all you never addressed my post
> What do you think will happen to the power grid when everyone switches to EVs?
Well the people who run grids keep suggesting that it'll help them make better use of their grid assets and so reduce the average cost of a kWh of electricity to consumers. It is after all a giant fleet of smart batteries.
But you, the person who thinks more energy use is better for society, are apparently making an exception for electricity use, which if it goes up will cause some unspecified calamity?
That doesn't seem very consistent? Well, I suppose it's consistently pro-fossil fuels.
> Well the people who run grids keep suggesting that it'll help them make better use of their grid assets
Do they? Or are they also concerned about the increasing strain on grid infra, especially due to the fact that so many energy sources are now intermittent?
> which if it goes up will cause some unspecified calamity?
It's not an unspecified calamity. Failure modes for grids are known.
> Well, I suppose it's consistently pro-fossil fuels.
Funny how you accuse others of making mad takes while keeping doing one mad take after another
> Yes, it is. There's a limit to how efficient you can go though
This is why reducing energy use needs to be in addition to everything else: solar, nuclear, wind, batteries, the works. Some of the R&D won't work out as planned; some solutions will only work in specific situations. That's okay, as long as we don't have all our eggs in one basket.
We have the battery tech, it's called a hydroelectric dam.
Isn't the point of using renewables so we don't destroy the environment?
Have you made a comparative analysis of how much a dam vs lithium mines destroy the environment?
The orders of magnitude needed for storage are not competitive with nuclear in any reasonable scenario.
Nuclear exists and is proven technology. Battery storage at the capacity and price scale required are not. That remains true even if you want absolutely to believe we can't build nuclear reactors faster than in 10 years, and we almost certainly can by starting mass producing them again.
> However, the company warned in Tuesday’s report that the “termination” of some clean energy projects during 2023 had pushed down the amount of renewables it had access to.
For Google's use case the sun is always shining somewhere. It might not be practical or realistic yet, but if they really wanted to they could.
This assumes that all datacenters are created equal and it doesn't matter if you run stuff in the US or in Asia, or...
You can't transport electricity that far, though.
But you can transport search queries.
It's true. If you build all your infrastructure n times over you could have solar panels powering (some of?) each infrastructure full infrastructure location. But would that work out better?
These are the only two graphs you need to understand that your assumption is based on >10y old data:
https://www.powerengineeringint.com/wp-content/uploads/2023/... https://upload.wikimedia.org/wikipedia/commons/7/7c/3-Learni...
By the decade is gone that any new nuclear project has been commissioned and built, the rest of the world is running on solar + battery storage at a tenth of the price.
Your first graph shows diminishing returns for both battery storage and solar. It's also not clear whether that's the marginal cost or includes the capital cost to build the structures and all the infrastructure.
To compare nuclear against solar you have to compare it against the sum of the cost of solar plus battery storage.
Investing in solar/wind/nuclear energy sources, storage, and research for improving all of the above seems like the best strategic option for at least the next decade.
Yes we need nuclear but these statements often feels like attempt to sweeping the problem under the rug. Because nuclear takes time to build. And this is not just about nuclear. Exxon report from early 80s stated that to cover energy demand nuclear or renewables would need 50 years to catch up. I guess demand is even bigger now.
This proposition still fits inside a "line goes up" vision, and it can't go up forever.
More nuclear doesn't come without additional risk nor cost, nor waste. I don't claim to have a solution, but it feels like jumping into a volcano to escape from the bear. There sure must be some way to deal with the bear instead.
Guys we HAVE to support the businesses! Even if they’re destroying everything! Think of the money
Google's carbon emissions surge nearly 50% due to AI energy demand - https://news.ycombinator.com/item?id=40859993 - (103 comments)
I just don't understand the implied argument here at all. We can do something that results in productivity gains, which people want or they wouldn't pay for it, but we don't want private companies to do this because we have decided an increase energy use is a bad thing. Isn't the obvious path to continue to grow towards more efficient energy sources rather than trying to suppress progress?
It’s about externalities.
The private companies aren’t paying for the damage they are causing. It’s ultimately borne by the public.
Very similar to a factory that poisons a river if there’s no disincentive not to.
Better would be if we could factor in the cost to the environment into their revenue. This would better align incentives.
You either have growth or limited emissions. Progress to more sustainable or more efficient energy sources til now just led to covering part of ever increasing demand. Coal is still rising as well.
Interesting considering that Google said it had a strategy that would ensure AI consumed less power. It was based on the 4Ms: model, machine, mechanization, and mapping.
https://ieeexplore.ieee.org/document/9810097
I wonder to which extent data centers are built already or will be built close to where renewables are readily available, to relieve the burden of transporting energy over long distances. Is it cheaper to build fibre optics to a remote location than it is to build power lines?
does anyone know if this is due to fleet growth (ie the existing commitment to going to zero emissions is still happening, quickly, but growth is so high it's swamping that in absolute terms) or has Urs gone back on his promise and is using emission-y energy?
[dupe]
More discussion: https://news.ycombinator.com/item?id=40859993
I'm curious to know how are the gas emissions measured
That's probably from all the hot air coming out of Mountain View about how virtuous and AI-oriented they are.