12 ms·

I'm having a hard time getting past the assumption in the article that energy use is tied directly to GDP per capita and that by not following the 7% growth of

by rtkwe 2y ago

I'm having a hard time getting past the assumption in the article that energy use is tied directly to GDP per capita and that by not following the 7% growth of the Henry Adams Curve we're somehow below where we should/could be as a country. That embeds so many assumptions about the economy and where GDP comes from, the decoupling seems more likely to be from the transition from manufacturing and other energy heavy sectors to more services based economic activity..

rconti 2y ago
Yeah, I had to re-read it 3 times to understand why the article was saying that was a _bad_ thing. Now I understand what he's saying but I'm pretty lost because it assumes a set of priors I simply never had.
- rtkwe 2y ago
  Even the graph that I think is supposed to support the assertion by gesturing towards a link between GDP/c and energy consumption/c shows there's a huge range of per capita energy consumption for countries in the same GDP per capita band hidden because the graph uses dual logarithmic axes. It's always a little hard to read logarithmic values other than those explicitly labelled but it looks like up to a 5-7x difference in energy consumption can have basically no effect on the GDP per capita!
  - the8472 2y ago
    I don't think the range matters so much as the floor. The huge blank in the lower-right corner. No such thing as a rich, energy-poor country.
    - rtkwe 2y ago
      The articles argument is about the US moving inside the range band we see though not about the US having a 100 fold reduction in the power generated that would be required to drop us into the lower right. I did the math on the data elsewhere in this thread [0] and in an outcome that should surprise no one there's a transition around 1970 where the ratio between power used and gdp created per capita changes drastically, in 1970 we produced .69 units of GDP per unit of energy and in 2014 we were producing 7.94 using inflation adjusted dollars and oil kg equivalent per capita. We just moved into a different type of economy and there's no data in the graph or article to back up the assertion that falling off the HA curve and consuming ~5x less power per capita our GDP is somehow 5x smaller. [0] https://news.ycombinator.com/item?id=41392222 https://news.ycombinator.com/item?id=41392222
      - the8472 2y ago
        There's some wiggle room, but if you want another OOM of advancement then it appears that you also need to move an OOM up, unless you started out at the upper end of the band. And even then the next one would definitely need more. Of course this only tells us about the current state of things, but there are lots of things that would seem ludicrous today that would be feasible if we had a lot more cheap energy. Mass desalation, carbon drawdown, synthetic fuels, electric arc furnaces etc. will all needs loads of energy. The signum is wrong. In the long term you want it to be one, not zero.
toss1 2y ago
>> That embeds so many assumptions about the economy and where GDP comes from YUP. Of course there is a strong correlation between energy use and GDP growth; it takes more energy to produce more stuff. But ultimately, what produces more stuff is harder to measure. To light the factories, more energy used to correlate with more light, until we swap out the incandescent/halogen lighting sources for LEDs. Then, we get more light for something like 16% of the energy usage. Or, getting to the stage of "lights-out" automation, and the same production for zero lighting energy. Same for more efficient motors, swapping ovens for inductive heating, more efficient processes, etc. Seems that measuring GPD growth by energy consumption is like Bill Gates' famous example of saying that "measuring software progress by lines of code is like measuring progress in aircraft by the weight of the planes". Obviously, in specific cases, all things being equal, more is more, but in reality, fewer LOC and lighter airplanes generally produce more results.
- jvanderbot 2y ago
  > It takes more energy to produce more stuff Energy consumption is also about using that stuff. And in rich countries like the USA, a lot of energy goes to using things or just moving people. So it's possible to build more and use less energy, even if you don't reduce the energy cost of building things. They can simply become more efficient to use and move. In USA, 30% is just homes and commercial sector. 40% is transportation. 30% is "industry". By a sort-of inverse-amdahl's law, it's possible to get lower energy use with more throughput even if we don't make "industry" more efficient.
  - macspoofing 2y ago
    >So it's possible to build more and use less energy, even if you don't reduce the energy cost of building things. In principle I can imagine that being true - but that's not represented in our world, so it is an open question if you can 'build more and use less energy' in the real world.
    - ForHackernews 2y ago
      It's represented everywhere that efficiency counts. I haven't fact-checked it, but one source[0] claims that refinements in the design of soda cans (thinner walls, different shapes) since 1960 saves "at least 90 million kilograms of aluminum annually." [0] https://www.youtube.com/watch?v=hUhisi2FBuw https://www.youtube.com/watch?v=hUhisi2FBuw
macspoofing 2y ago
>I'm having a hard time getting past the assumption in the article that energy use is tied directly to GDP per capita Why? They are correlated and so are reasonable metric to gauge progress when starting from a subsistence economy (as all economies in the world began). At some point, this may be less true when you hit a energy generation ceiling and you start 'optimizing' and trying to do more with the same amount .. but again, we're not there yet so it's a good metric today, and especially for developing economies. Put another way, you show me GDP per capita or per capita Energy use and I can get a reasonable ballpark for the other as well as a measure for the wealth of that nation.
- sobellian 2y ago
  We are there today: https://ourworldindata.org/energy-gdp-decoupling https://ourworldindata.org/energy-gdp-decoupling
  - Mistletoe 2y ago
    This always makes me uncomfortable though. How would we tell the difference between rampant scamming and fudging numbers and an economy where we all pass around Monopoly money to do services for one another? I pay you to mow my lawn and you pay me to mow your lawn. Are we creating GDP?
    - sobellian 2y ago
      I'm not sure goods are quite immune from this objection. Industry turns out plenty of useless widgets. The Humane AI pin creates GDP!
      - rtkwe 2y ago
        Very true and you quickly get into a very command economy style argument about what should be produced. Ultimately we have the system we have and wasted or scammy products generally eventually die. Look at things like NFTs they were an extremely brief blip it turns out because people quickly saturated the ability of crypto early adopters to inflate values with their funny money. Some scams last longer like Thomas Kincade 'paintings' but trying to sort through the economic data to throw those out is just not possible.
jvanderbot 2y ago
Yes this confuses me too. In developed countries a huge amount of energy goes to using things, not just making them. Let alone efficiency gains from production. This is possibly why energy use has flattened while GDP clips along at its normal exponential curve - we're also more efficient.
- hadlock 2y ago
  Lighting going from 1% efficient (99% of the energy used by a traditional tungsten lightbulb is waste heat) to over 95% efficient (in many cases 99% efficient). Computers went from 200-400w continuous down to about 15w continuous. Lighting alone is responsible for a huge amount of energy drop-off.
  - obscurette 2y ago
    Sorry, but these numbers are not accurate. Luminous efficiency of tungsten lightbulbs is typically around 1-3% (depends on voltage and power), and luminous efficiency of LED bulbs is around 10-30%. (https://en.wikipedia.org/wiki/Luminous_efficacy https://en.wikipedia.org/wiki/Luminous_efficacy)
    - Iulioh 2y ago
      You forgot the part when we used 100W bulbs and now we use 10W led lights. Efficency in the use of energy ti generate the same amount of lumens, because it would be REALLY fun to have a 100W led lamp in my bathroom mirror (my parents still have 2 of these traditional lights ). I mean...energy conversion to lumen is fine but i think it's a little pedantic
      - MarkusQ 2y ago
        GP didn't "forget" it; that 10x improvement in efficiency is the reason we can use 10W bulbs to do the job that used to require a 100W bulb. Looking at energy conversion to lumens isn't "pedantic," it's "correct".
specialist 2y ago
I clocked that too. Perhaps author is nodding towards replacing fossil fuels with electricity. Decarbonizing steel uses a lot more energy. Ditto cement, plastics, fertilizers, HVAC, etc. Anyone care to guess how much more energy our glorious renewable energy future perfect economy will require? 4x? 6x?
- bryanlarsen 2y ago
  How about significantly less than 1x? Fossil fuels waste most of their energy as heat. Gasoline cars are ~30% efficient, EV's are ~90%. And heat pumps are often 200-500% efficient, unlike fossil fuel furnaces which cap out around 95%. Transportation and heating use far more fuel than industrial uses.
  - specialist 2y ago
    Ya, that's Saul Griffith's prediction too. I'll defer to you both. Tragically, I got wonksniped by Pueyo's Henry Adams Curve shout out. TIL He's referring the Roots of Progress thesis. The mythical "stagnation" phenomonen that some "rationalists" used to obsess over. From the hip: the mistake is measuring national vs global per capita energy use. As many, many have noted, we delegated our energy consumption by moving our mfg overseas. Mystery solved. Further, notice similarity between Roots of Progress' graph and https://wtfhappenedin1971.com/ https://wtfhappenedin1971.com/ Spoiler: Neoliberalism happened. aka globalization, austerity, supply-side economics -- FWIW: Omitting Pueyo's tangent about the Henry Adams Curve, I found this article to be a great overview of solar PV's current position on its cost-learning-curve. And I agree the cost of solar PV will decrease for some time. Even faster than the most optimistic projections, which has been the norm for years. Exciting times. I look forward to Pueyo's article explaining why price of electricity continues to rise despite decreasing production costs. Transmission? Utility monopolies? Financing?
    - bryanlarsen 2y ago
      > why price of electricity continues to rise despite decreasing production costs. That's economics 201. The price of a commodity is the cost of the marginal producer. So it doesn't matter how cheap some producers are, the price of a commodity is set by the most expensive producer that is meeting demand. So price of electricity won't drop until cheap producers can meet 100% of demand. Until that happens, cheap producers enjoy outsized profits, encouraging more cheap producers to join the market.
sharpshadow 2y ago
Look at Germany the next years. We gave up coal, nuclear and now gas. Have the highest prices for energy in the world and our industry is leaving. If we would have more energy we would get much richer and not more poor like right now.
- bryanlarsen 2y ago
  The price of a commodity is equal to the highest of the lowest cost producers that can satisfy demand. Germany cannot satisfy demand with cheap renewables, so the marginal producer is expensive gas & nuclear. German electricity is expensive because gas is expensive in Germany. Electricity will be expensive in Germany until Germany completely stops using gas shipped in via container ship.
  - sharpshadow 2y ago
    You got it right there Germany can’t satisfy demand with renewables and decided to cut itself of from coal and nuclear, brilliant right!? Cherry on top some powerful player blew up Nordstream forcing Germany to buy exorbitantly expensive LNG. Your argument to stop using LNG doesn’t work if you have many parts of industry build with gas infrastructure which can’t be replaced just like that.
    - kitkat_new 2y ago
      > Cherry on top some powerful player blew up Nordstream forcing Germany to buy exorbitantly expensive LNG. Germany was forced already before that, since Russia used the Gas to blackmail Germany
    - bryanlarsen 2y ago
      Only if Germany could replace all of their expensive natural gas usage with coal & nuclear would coal & nuclear be relevant. They can't, so it isn't.
    - derriz 2y ago
      Germany has been the biggest exporter of electricity in the world for 8 of the last 10 years[0]. It consistently generates more than it consumes. It's been this way since around 2004. German wholesale electricity prices are relatively low by European standards - so far this year about 8th cheapest - about 13% cheaper than of France, for example[1]. This reflects the blended cost of production. Household prices are higher than average - because domestic consumption of electricity is taxed more heavily in Germany than the average in Europe. [0] https://oec.world/en/profile/hs/electrical-energy https://oec.world/en/profile/hs/electrical-energy [1] https://ember-climate.org/data-catalogue/european-wholesale-electricity-price-data/ https://ember-climate.org/data-catalogue/european-wholesale-...
ZeroGravitas 2y ago
It also neatly corresponds with the mass rollout of nuclear power so it's possible it's just the classic measurement issue of "primary energy" vs "useful energy", sometimes called the primary energy fallacy, which makes fossil fuel based systems appear 4x (or probably more going back in time) better than they actually are.
- ZeroGravitas 2y ago
  I tried to track this down, found some blogs by the originator, but it only gets more confusing: https://wimflyc.blogspot.com/2021/01/the-henry-adams-curve-closer-look.html https://wimflyc.blogspot.com/2021/01/the-henry-adams-curve-c... Henry Adams points out that over 60 years you got 3-4x more power from a ton of coal. That combined with the extra coal dug up, he claims, doubled usable energy every ten years. But then the modern graph simply shows the coal energy, with (as far as I can tell) no attempt to account for the extra efficiency, even though the modern author of the graph makes explicit reference to the increasing efficiency of steam engines.
javiramos 2y ago
A fantastic deep dive into this topic is Vaclav Smil's book Energy and Civilization: https://mitpress.mit.edu/9780262536165/energy-and-civilization/ https://mitpress.mit.edu/9780262536165/energy-and-civilizati...
- kragen 2y ago
  thank you! this looks highly relevant to my interests
- pfdietz 2y ago
  Smil was wildly wrong about the trajectory of PV improvement. In his defense, so was almost everyone else. I found the following comment about Smil: "There is a way how to evaluate the quality of prophets, seers and visionaries. Find their 5,10,15 years old predictions. I own the Smil’s book: Energy Myths and Realities: Bringing Science to the Energy Policy Debate (2010). So have a look how good prophet he was 10 years in advance and focus on photovoltaics (I own Czech translation of the book, so I need to re-translate his text back to English, I hope I will not skew his ideas too much): * To get 1 PWh/year of electricity you need to install about 450 GW worth of solar panels. You need dozens of years to acomplish such task. Reality check: 3 years in current speed, in the future probably faster. * The cost of PV panels fell from 5 USD/W in 2000 to 4.5 USD/W in 2009. He don’t see much perspective of price plummeting as predicted f.e. by Al Gore (who cited the learning curve, Smil counted with 0.05 USD/W in 2020) or by PV industry (1.5 USD/W in 2020). Smil predicted that PV panels would be 25% cheaper in 2020 and 50% in 2030. Reality check: Current price of PV panels is ~0.2 W/USD. While Smil wrote the book manufacturers finally scaled their production of polysilicone and PV cells to cover the demand. Competition among them set the cost of PV panel on the freefall trajectory. PV panels cost less than he predicted for 2030 in 2011. How credible are such visionaries?" I will add that in 2023, 447 GW of PV was installed globally. So, we're at the point where Smil's "dozens of years" is being done each and every year.
  - ZeroGravitas 2y ago
    It's hard to tell if Smil is a grumpy old academic who's been catapulted to fame because his cranky opinions happened to line up with some political faction (like Jordan Peterson) or if he's been fundamentally warped by the money and attention available to those who take certain positions (like a slightly later Jordan Peterson). His doomerism and degrowth perspective points to the early JP, writing articles mocking Obama for lack of progress with EVs points to the latter. Either way he managed to convince Bill Gates to massively misinvest in climate solutions and while doing so loudly bad mouth all the actual solutions so that alone is a massive net negative for humanity (and, more parochially, for the USA, which is super ironic given his anti-communist stance--another weird JP parallel--as he's been super helpful to the Chinese by letting them solve all the problems he claimed were inpossible and dissuaded US companies from investing in).
countvonbalzac 2y ago
Same - take lightbulbs for example. Thanks to LEDs, the amount of energy you need to generate X amount of light has reduced considerably, but we still have as much if not more lighting than ever. And it's not like our GDP is suffering due to a lack of sufficient lighting, at a certain point there's no gain to productivity gained from having another lightbulb. Same thing can be said about cars, CPUs etc.
- Havoc 2y ago
  Think it’s meant more as a broad generalization than something that is always true. Many physical things take pretty fixed amounts of energy. Eg heating a liter of water.
  - HideousKojima 2y ago
    Or refining aluminum, which uses something like 1.5% of all US electricity generation.
    - kragen 2y ago
      it would be much more, but the us imports 130% of the aluminum it consumes from places with lower energy costs: https://pubs.usgs.gov/periodicals/mcs2022/mcs2022-aluminum.pdf https://pubs.usgs.gov/periodicals/mcs2022/mcs2022-aluminum.p...
      - bobthepanda 2y ago
        Bahrain and UAE being big aluminum smelters is a bit surprising.
        HPsquared 2y ago
        Iceland too. It takes a few workers and a lot of energy. Very sensitive to market conditions though. Peaky.
        cameldrv 2y ago
kragen 2y ago
presumably what's going to happen is that, as energy becomes dramatically cheaper, production will shift from less-energy-intensive processes toward more-energy-intensive processes. this is likely to happen at many levels: between different routes for producing the same good (for example, pidgeon vs. dow process), between alternative goods (for example, aluminum vs. steel), between subsectors (for example, heavy industry vs. high-precision manufacturing), and across sectors (for example, manufacturing vs. services) so, with the advent of innovations that dramatically drop the cost of energy, we should expect to see energy use grow faster than gdp. that's decoupling but in the opposite direction from the decoupling you're talking about, which has been driven by the 01973–02023 energy crisis
empiricus 2y ago
You need energy to make things. Houses, roads, cars, food, many services. Basically without energy you are poor. Having energy is not enough, but it is necessary to be rich.
- hinkley 2y ago
  But sometime in the 90’s if not sooner a lot of us started valuing items that had a lower embodied energy as a ratio of the usefulness of the product. We want things that are the same but less intense, or that are much better at a fractional increase in input. And we really need it if we want a planet worth living on fifty years from now. So to ignore this desire is dangerous.
  - tmnvix 2y ago
    A lot of us, maybe. But my impression is that a lot of us also choose to consume more stuff more frequently (larger vehicles and homes, less durable items replaced more frequently, etc). I would not be surprised to learn that our direct energy use - such as vehicle fuel and electricity for heating - has increased with longer commutes and larger homes. Compared to just 50 years ago, I would say that our lifestyles are on average vastly more consumptive, despite being more environmentally aware. We seem unwilling to make the sacrifices that really matter. I think that some time in the future, our time will be seen as one of massive entitlement. As technology makes things possible, we feel entitled to make use of it if we can afford it. How many people in the 50's were using trucks to drag boats and horse floats around in the suburbs? These are the kind of reasons people will give for their consumption. e.g. "this is the lifestyle I want, it's possible, and why am I not entitled to it if I can afford to pay for it?" (in the financial sense only of course).
    - GardenLetter27 2y ago
      Higher consumption = higher standard of living. We have nuclear power (and soon nuclear fusion) - we can completely decouple energy production from environmental damage. And trying to force degrowth on populations just makes matters worse, as their economy collapses (Europe), or they resort to even worse forms of energy production like deforestation and wood burning (Sub-Saharan Africa and remote South America).
acchow 2y ago
> the decoupling seems more likely to be from the transition from manufacturing and other energy heavy sectors to more services based economic activity Put another way, instead of producing the products we consume, we offshore production (and the associated energy consumption)
pfdietz 2y ago
It's a confusion of cause and effect, classic cargo cult thinking.
llsf 2y ago
Energy is the fuel for machines. And machines build more and faster than humans. To increase productivity, we need more machines, and so more energy. If energy was 10 times cheaper than currently and still available (no shortage) then we would automate more tasks and increase the global productivity. We could mine more, build more, to train/run more LLMs. GDP is tied to energy price and availability. It is likely the main reason why US enjoyed a growing economy thanks to its free oil reserves (and the reason why Texas alone is the 8th economy in the world). Now one could argue that after pumping oil for a century now, it might not be as cheap or plentiful to extract in the future, and we better be prepared for it. It is all about energy. No energy, and we will reduce to human/manual or animal (bring back the horses!) labor. Productivity will decrease significantly. One issue I have with the paper, is that it equates energy and electricity. My favorite chart about US energy: https://flowcharts.llnl.gov/sites/flowcharts/files/2023-10/US%20Energy%202022.png https://flowcharts.llnl.gov/sites/flowcharts/files/2023-10/U... shows that in 2022, electricity was just 13.3 Quads of the 76.07 Quads of the energy consumed in US by residential, commercial, industrial and transportation. So unless US "electrify" the whole economy having more electricity would only help a small part of the economy.
- zamadatix 2y ago
  > Energy is the fuel for machines. And machines build more and faster than humans Sure. > To increase productivity, we need more machines, That is just one way to increase productivity. You can also make more efficient machines, come up with more efficient production processes, find alternatives to existing products that are more useful for a given task per quantity, and similar kinds of concepts. This is the point GP was making (there are more ways production changes than increasing energy used), not that energy is completely unrelated to production as a whole.
  - MostlyStable 2y ago
    At any given efficiency, more energy means more productivity (in the same way: at any given amount of energy production, more efficiency means more productivity). Yes, we should be increasing efficiency where reasonable, but we should also be increasing energy production. There is a reason that the Kardashev scale is about energy utilization.
alliao 2y ago
https://ourworldindata.org/grapher/energy-use-per-person-vs-gdp-per-capita https://ourworldindata.org/grapher/energy-use-per-person-vs-... is probably where the inspiration comes from...
neolefty 2y ago
> Exactly I'm having a hard time getting past the assumption in the article that energy use is tied directly to GDP per capita and that by not following the 7% growth of the Henry Adams Curve Yes, this was a central simplifying fallacy in the article for me too. Plenty of phenomena follow a logistic curve [1] — which starts out exponential but then flattens out when it reaches constraints or fills a niche or fully satisfies a need or something. The initial exponential growth may have been a period where economic productivity was constrained in a major by energy availability. We still have some energy constraints, but they seem to be secondary — and in many areas energy requirements have fallen, as others have pointed out in this discussion. It's too complicated to simply flatten, but modern energy use does seem to broadly resemble a logistic curve. --- [1] https://en.wikipedia.org/wiki/Logistic_function https://en.wikipedia.org/wiki/Logistic_function

by rtkwe 2y ago

rconti 2y ago

Yeah, I had to re-read it 3 times to understand why the article was saying that was a _bad_ thing. Now I understand what he's saying but I'm pretty lost because it assumes a set of priors I simply never had.

rtkwe 2y ago
Even the graph that I think is supposed to support the assertion by gesturing towards a link between GDP/c and energy consumption/c shows there's a huge range of per capita energy consumption for countries in the same GDP per capita band hidden because the graph uses dual logarithmic axes. It's always a little hard to read logarithmic values other than those explicitly labelled but it looks like up to a 5-7x difference in energy consumption can have basically no effect on the GDP per capita!
- the8472 2y ago
  I don't think the range matters so much as the floor. The huge blank in the lower-right corner. No such thing as a rich, energy-poor country.
  - rtkwe 2y ago
    The articles argument is about the US moving inside the range band we see though not about the US having a 100 fold reduction in the power generated that would be required to drop us into the lower right. I did the math on the data elsewhere in this thread [0] and in an outcome that should surprise no one there's a transition around 1970 where the ratio between power used and gdp created per capita changes drastically, in 1970 we produced .69 units of GDP per unit of energy and in 2014 we were producing 7.94 using inflation adjusted dollars and oil kg equivalent per capita. We just moved into a different type of economy and there's no data in the graph or article to back up the assertion that falling off the HA curve and consuming ~5x less power per capita our GDP is somehow 5x smaller. [0] https://news.ycombinator.com/item?id=41392222 https://news.ycombinator.com/item?id=41392222
    - the8472 2y ago
      There's some wiggle room, but if you want another OOM of advancement then it appears that you also need to move an OOM up, unless you started out at the upper end of the band. And even then the next one would definitely need more. Of course this only tells us about the current state of things, but there are lots of things that would seem ludicrous today that would be feasible if we had a lot more cheap energy. Mass desalation, carbon drawdown, synthetic fuels, electric arc furnaces etc. will all needs loads of energy. The signum is wrong. In the long term you want it to be one, not zero.

toss1 2y ago

>> That embeds so many assumptions about the economy and where GDP comes from YUP. Of course there is a strong correlation between energy use and GDP growth; it takes more energy to produce more stuff. But ultimately, what produces more stuff is harder to measure. To light the factories, more energy used to correlate with more light, until we swap out the incandescent/halogen lighting sources for LEDs. Then, we get more light for something like 16% of the energy usage. Or, getting to the stage of "lights-out" automation, and the same production for zero lighting energy. Same for more efficient motors, swapping ovens for inductive heating, more efficient processes, etc. Seems that measuring GPD growth by energy consumption is like Bill Gates' famous example of saying that "measuring software progress by lines of code is like measuring progress in aircraft by the weight of the planes". Obviously, in specific cases, all things being equal, more is more, but in reality, fewer LOC and lighter airplanes generally produce more results.

jvanderbot 2y ago
> It takes more energy to produce more stuff Energy consumption is also about using that stuff. And in rich countries like the USA, a lot of energy goes to using things or just moving people. So it's possible to build more and use less energy, even if you don't reduce the energy cost of building things. They can simply become more efficient to use and move. In USA, 30% is just homes and commercial sector. 40% is transportation. 30% is "industry". By a sort-of inverse-amdahl's law, it's possible to get lower energy use with more throughput even if we don't make "industry" more efficient.
- macspoofing 2y ago
  >So it's possible to build more and use less energy, even if you don't reduce the energy cost of building things. In principle I can imagine that being true - but that's not represented in our world, so it is an open question if you can 'build more and use less energy' in the real world.
  - ForHackernews 2y ago
    It's represented everywhere that efficiency counts. I haven't fact-checked it, but one source[0] claims that refinements in the design of soda cans (thinner walls, different shapes) since 1960 saves "at least 90 million kilograms of aluminum annually." [0] https://www.youtube.com/watch?v=hUhisi2FBuw https://www.youtube.com/watch?v=hUhisi2FBuw

macspoofing 2y ago

>I'm having a hard time getting past the assumption in the article that energy use is tied directly to GDP per capita Why? They are correlated and so are reasonable metric to gauge progress when starting from a subsistence economy (as all economies in the world began). At some point, this may be less true when you hit a energy generation ceiling and you start 'optimizing' and trying to do more with the same amount .. but again, we're not there yet so it's a good metric today, and especially for developing economies. Put another way, you show me GDP per capita or per capita Energy use and I can get a reasonable ballpark for the other as well as a measure for the wealth of that nation.

sobellian 2y ago
We are there today: https://ourworldindata.org/energy-gdp-decoupling https://ourworldindata.org/energy-gdp-decoupling
- Mistletoe 2y ago
  This always makes me uncomfortable though. How would we tell the difference between rampant scamming and fudging numbers and an economy where we all pass around Monopoly money to do services for one another? I pay you to mow my lawn and you pay me to mow your lawn. Are we creating GDP?
  - sobellian 2y ago
    I'm not sure goods are quite immune from this objection. Industry turns out plenty of useless widgets. The Humane AI pin creates GDP!
    - rtkwe 2y ago
      Very true and you quickly get into a very command economy style argument about what should be produced. Ultimately we have the system we have and wasted or scammy products generally eventually die. Look at things like NFTs they were an extremely brief blip it turns out because people quickly saturated the ability of crypto early adopters to inflate values with their funny money. Some scams last longer like Thomas Kincade 'paintings' but trying to sort through the economic data to throw those out is just not possible.

jvanderbot 2y ago

Yes this confuses me too. In developed countries a huge amount of energy goes to using things, not just making them. Let alone efficiency gains from production. This is possibly why energy use has flattened while GDP clips along at its normal exponential curve - we're also more efficient.

hadlock 2y ago
Lighting going from 1% efficient (99% of the energy used by a traditional tungsten lightbulb is waste heat) to over 95% efficient (in many cases 99% efficient). Computers went from 200-400w continuous down to about 15w continuous. Lighting alone is responsible for a huge amount of energy drop-off.
- obscurette 2y ago
  Sorry, but these numbers are not accurate. Luminous efficiency of tungsten lightbulbs is typically around 1-3% (depends on voltage and power), and luminous efficiency of LED bulbs is around 10-30%. (https://en.wikipedia.org/wiki/Luminous_efficacy https://en.wikipedia.org/wiki/Luminous_efficacy)
  - Iulioh 2y ago
    You forgot the part when we used 100W bulbs and now we use 10W led lights. Efficency in the use of energy ti generate the same amount of lumens, because it would be REALLY fun to have a 100W led lamp in my bathroom mirror (my parents still have 2 of these traditional lights ). I mean...energy conversion to lumen is fine but i think it's a little pedantic
    - MarkusQ 2y ago
      GP didn't "forget" it; that 10x improvement in efficiency is the reason we can use 10W bulbs to do the job that used to require a 100W bulb. Looking at energy conversion to lumens isn't "pedantic," it's "correct".

specialist 2y ago

I clocked that too. Perhaps author is nodding towards replacing fossil fuels with electricity. Decarbonizing steel uses a lot more energy. Ditto cement, plastics, fertilizers, HVAC, etc. Anyone care to guess how much more energy our glorious renewable energy future perfect economy will require? 4x? 6x?

bryanlarsen 2y ago
How about significantly less than 1x? Fossil fuels waste most of their energy as heat. Gasoline cars are ~30% efficient, EV's are ~90%. And heat pumps are often 200-500% efficient, unlike fossil fuel furnaces which cap out around 95%. Transportation and heating use far more fuel than industrial uses.
- specialist 2y ago
  Ya, that's Saul Griffith's prediction too. I'll defer to you both. Tragically, I got wonksniped by Pueyo's Henry Adams Curve shout out. TIL He's referring the Roots of Progress thesis. The mythical "stagnation" phenomonen that some "rationalists" used to obsess over. From the hip: the mistake is measuring national vs global per capita energy use. As many, many have noted, we delegated our energy consumption by moving our mfg overseas. Mystery solved. Further, notice similarity between Roots of Progress' graph and https://wtfhappenedin1971.com/ https://wtfhappenedin1971.com/ Spoiler: Neoliberalism happened. aka globalization, austerity, supply-side economics -- FWIW: Omitting Pueyo's tangent about the Henry Adams Curve, I found this article to be a great overview of solar PV's current position on its cost-learning-curve. And I agree the cost of solar PV will decrease for some time. Even faster than the most optimistic projections, which has been the norm for years. Exciting times. I look forward to Pueyo's article explaining why price of electricity continues to rise despite decreasing production costs. Transmission? Utility monopolies? Financing?
  - bryanlarsen 2y ago
    > why price of electricity continues to rise despite decreasing production costs. That's economics 201. The price of a commodity is the cost of the marginal producer. So it doesn't matter how cheap some producers are, the price of a commodity is set by the most expensive producer that is meeting demand. So price of electricity won't drop until cheap producers can meet 100% of demand. Until that happens, cheap producers enjoy outsized profits, encouraging more cheap producers to join the market.

sharpshadow 2y ago

Look at Germany the next years. We gave up coal, nuclear and now gas. Have the highest prices for energy in the world and our industry is leaving. If we would have more energy we would get much richer and not more poor like right now.

bryanlarsen 2y ago
The price of a commodity is equal to the highest of the lowest cost producers that can satisfy demand. Germany cannot satisfy demand with cheap renewables, so the marginal producer is expensive gas & nuclear. German electricity is expensive because gas is expensive in Germany. Electricity will be expensive in Germany until Germany completely stops using gas shipped in via container ship.
- sharpshadow 2y ago
  You got it right there Germany can’t satisfy demand with renewables and decided to cut itself of from coal and nuclear, brilliant right!? Cherry on top some powerful player blew up Nordstream forcing Germany to buy exorbitantly expensive LNG. Your argument to stop using LNG doesn’t work if you have many parts of industry build with gas infrastructure which can’t be replaced just like that.
  - kitkat_new 2y ago
    > Cherry on top some powerful player blew up Nordstream forcing Germany to buy exorbitantly expensive LNG. Germany was forced already before that, since Russia used the Gas to blackmail Germany
  - bryanlarsen 2y ago
    Only if Germany could replace all of their expensive natural gas usage with coal & nuclear would coal & nuclear be relevant. They can't, so it isn't.
  - derriz 2y ago
    Germany has been the biggest exporter of electricity in the world for 8 of the last 10 years[0]. It consistently generates more than it consumes. It's been this way since around 2004. German wholesale electricity prices are relatively low by European standards - so far this year about 8th cheapest - about 13% cheaper than of France, for example[1]. This reflects the blended cost of production. Household prices are higher than average - because domestic consumption of electricity is taxed more heavily in Germany than the average in Europe. [0] https://oec.world/en/profile/hs/electrical-energy https://oec.world/en/profile/hs/electrical-energy [1] https://ember-climate.org/data-catalogue/european-wholesale-electricity-price-data/ https://ember-climate.org/data-catalogue/european-wholesale-...

ZeroGravitas 2y ago

It also neatly corresponds with the mass rollout of nuclear power so it's possible it's just the classic measurement issue of "primary energy" vs "useful energy", sometimes called the primary energy fallacy, which makes fossil fuel based systems appear 4x (or probably more going back in time) better than they actually are.

ZeroGravitas 2y ago
I tried to track this down, found some blogs by the originator, but it only gets more confusing: https://wimflyc.blogspot.com/2021/01/the-henry-adams-curve-closer-look.html https://wimflyc.blogspot.com/2021/01/the-henry-adams-curve-c... Henry Adams points out that over 60 years you got 3-4x more power from a ton of coal. That combined with the extra coal dug up, he claims, doubled usable energy every ten years. But then the modern graph simply shows the coal energy, with (as far as I can tell) no attempt to account for the extra efficiency, even though the modern author of the graph makes explicit reference to the increasing efficiency of steam engines.

javiramos 2y ago

A fantastic deep dive into this topic is Vaclav Smil's book Energy and Civilization: https://mitpress.mit.edu/9780262536165/energy-and-civilization/ https://mitpress.mit.edu/9780262536165/energy-and-civilizati...

kragen 2y ago
thank you! this looks highly relevant to my interests
pfdietz 2y ago
Smil was wildly wrong about the trajectory of PV improvement. In his defense, so was almost everyone else. I found the following comment about Smil: "There is a way how to evaluate the quality of prophets, seers and visionaries. Find their 5,10,15 years old predictions. I own the Smil’s book: Energy Myths and Realities: Bringing Science to the Energy Policy Debate (2010). So have a look how good prophet he was 10 years in advance and focus on photovoltaics (I own Czech translation of the book, so I need to re-translate his text back to English, I hope I will not skew his ideas too much): * To get 1 PWh/year of electricity you need to install about 450 GW worth of solar panels. You need dozens of years to acomplish such task. Reality check: 3 years in current speed, in the future probably faster. * The cost of PV panels fell from 5 USD/W in 2000 to 4.5 USD/W in 2009. He don’t see much perspective of price plummeting as predicted f.e. by Al Gore (who cited the learning curve, Smil counted with 0.05 USD/W in 2020) or by PV industry (1.5 USD/W in 2020). Smil predicted that PV panels would be 25% cheaper in 2020 and 50% in 2030. Reality check: Current price of PV panels is ~0.2 W/USD. While Smil wrote the book manufacturers finally scaled their production of polysilicone and PV cells to cover the demand. Competition among them set the cost of PV panel on the freefall trajectory. PV panels cost less than he predicted for 2030 in 2011. How credible are such visionaries?" I will add that in 2023, 447 GW of PV was installed globally. So, we're at the point where Smil's "dozens of years" is being done each and every year.
- ZeroGravitas 2y ago
  It's hard to tell if Smil is a grumpy old academic who's been catapulted to fame because his cranky opinions happened to line up with some political faction (like Jordan Peterson) or if he's been fundamentally warped by the money and attention available to those who take certain positions (like a slightly later Jordan Peterson). His doomerism and degrowth perspective points to the early JP, writing articles mocking Obama for lack of progress with EVs points to the latter. Either way he managed to convince Bill Gates to massively misinvest in climate solutions and while doing so loudly bad mouth all the actual solutions so that alone is a massive net negative for humanity (and, more parochially, for the USA, which is super ironic given his anti-communist stance--another weird JP parallel--as he's been super helpful to the Chinese by letting them solve all the problems he claimed were inpossible and dissuaded US companies from investing in).

countvonbalzac 2y ago

Same - take lightbulbs for example. Thanks to LEDs, the amount of energy you need to generate X amount of light has reduced considerably, but we still have as much if not more lighting than ever. And it's not like our GDP is suffering due to a lack of sufficient lighting, at a certain point there's no gain to productivity gained from having another lightbulb. Same thing can be said about cars, CPUs etc.

Havoc 2y ago
Think it’s meant more as a broad generalization than something that is always true. Many physical things take pretty fixed amounts of energy. Eg heating a liter of water.
- HideousKojima 2y ago
  Or refining aluminum, which uses something like 1.5% of all US electricity generation.
  - kragen 2y ago
    it would be much more, but the us imports 130% of the aluminum it consumes from places with lower energy costs: https://pubs.usgs.gov/periodicals/mcs2022/mcs2022-aluminum.pdf https://pubs.usgs.gov/periodicals/mcs2022/mcs2022-aluminum.p...
    - bobthepanda 2y ago
      Bahrain and UAE being big aluminum smelters is a bit surprising.
      - HPsquared 2y ago
        Iceland too. It takes a few workers and a lot of energy. Very sensitive to market conditions though. Peaky.
      - cameldrv 2y ago

kragen 2y ago

presumably what's going to happen is that, as energy becomes dramatically cheaper, production will shift from less-energy-intensive processes toward more-energy-intensive processes. this is likely to happen at many levels: between different routes for producing the same good (for example, pidgeon vs. dow process), between alternative goods (for example, aluminum vs. steel), between subsectors (for example, heavy industry vs. high-precision manufacturing), and across sectors (for example, manufacturing vs. services) so, with the advent of innovations that dramatically drop the cost of energy, we should expect to see energy use grow faster than gdp. that's decoupling but in the opposite direction from the decoupling you're talking about, which has been driven by the 01973–02023 energy crisis

empiricus 2y ago

You need energy to make things. Houses, roads, cars, food, many services. Basically without energy you are poor. Having energy is not enough, but it is necessary to be rich.

hinkley 2y ago
But sometime in the 90’s if not sooner a lot of us started valuing items that had a lower embodied energy as a ratio of the usefulness of the product. We want things that are the same but less intense, or that are much better at a fractional increase in input. And we really need it if we want a planet worth living on fifty years from now. So to ignore this desire is dangerous.
- tmnvix 2y ago
  A lot of us, maybe. But my impression is that a lot of us also choose to consume more stuff more frequently (larger vehicles and homes, less durable items replaced more frequently, etc). I would not be surprised to learn that our direct energy use - such as vehicle fuel and electricity for heating - has increased with longer commutes and larger homes. Compared to just 50 years ago, I would say that our lifestyles are on average vastly more consumptive, despite being more environmentally aware. We seem unwilling to make the sacrifices that really matter. I think that some time in the future, our time will be seen as one of massive entitlement. As technology makes things possible, we feel entitled to make use of it if we can afford it. How many people in the 50's were using trucks to drag boats and horse floats around in the suburbs? These are the kind of reasons people will give for their consumption. e.g. "this is the lifestyle I want, it's possible, and why am I not entitled to it if I can afford to pay for it?" (in the financial sense only of course).
  - GardenLetter27 2y ago
    Higher consumption = higher standard of living. We have nuclear power (and soon nuclear fusion) - we can completely decouple energy production from environmental damage. And trying to force degrowth on populations just makes matters worse, as their economy collapses (Europe), or they resort to even worse forms of energy production like deforestation and wood burning (Sub-Saharan Africa and remote South America).

acchow 2y ago

> the decoupling seems more likely to be from the transition from manufacturing and other energy heavy sectors to more services based economic activity Put another way, instead of producing the products we consume, we offshore production (and the associated energy consumption)

pfdietz 2y ago

It's a confusion of cause and effect, classic cargo cult thinking.

llsf 2y ago

Energy is the fuel for machines. And machines build more and faster than humans. To increase productivity, we need more machines, and so more energy. If energy was 10 times cheaper than currently and still available (no shortage) then we would automate more tasks and increase the global productivity. We could mine more, build more, to train/run more LLMs. GDP is tied to energy price and availability. It is likely the main reason why US enjoyed a growing economy thanks to its free oil reserves (and the reason why Texas alone is the 8th economy in the world). Now one could argue that after pumping oil for a century now, it might not be as cheap or plentiful to extract in the future, and we better be prepared for it. It is all about energy. No energy, and we will reduce to human/manual or animal (bring back the horses!) labor. Productivity will decrease significantly. One issue I have with the paper, is that it equates energy and electricity. My favorite chart about US energy: https://flowcharts.llnl.gov/sites/flowcharts/files/2023-10/US%20Energy%202022.png https://flowcharts.llnl.gov/sites/flowcharts/files/2023-10/U... shows that in 2022, electricity was just 13.3 Quads of the 76.07 Quads of the energy consumed in US by residential, commercial, industrial and transportation. So unless US "electrify" the whole economy having more electricity would only help a small part of the economy.

zamadatix 2y ago
> Energy is the fuel for machines. And machines build more and faster than humans Sure. > To increase productivity, we need more machines, That is just one way to increase productivity. You can also make more efficient machines, come up with more efficient production processes, find alternatives to existing products that are more useful for a given task per quantity, and similar kinds of concepts. This is the point GP was making (there are more ways production changes than increasing energy used), not that energy is completely unrelated to production as a whole.
- MostlyStable 2y ago
  At any given efficiency, more energy means more productivity (in the same way: at any given amount of energy production, more efficiency means more productivity). Yes, we should be increasing efficiency where reasonable, but we should also be increasing energy production. There is a reason that the Kardashev scale is about energy utilization.

alliao 2y ago

https://ourworldindata.org/grapher/energy-use-per-person-vs-gdp-per-capita https://ourworldindata.org/grapher/energy-use-per-person-vs-... is probably where the inspiration comes from...

neolefty 2y ago

> Exactly I'm having a hard time getting past the assumption in the article that energy use is tied directly to GDP per capita and that by not following the 7% growth of the Henry Adams Curve Yes, this was a central simplifying fallacy in the article for me too. Plenty of phenomena follow a logistic curve [1] — which starts out exponential but then flattens out when it reaches constraints or fills a niche or fully satisfies a need or something. The initial exponential growth may have been a period where economic productivity was constrained in a major by energy availability. We still have some energy constraints, but they seem to be secondary — and in many areas energy requirements have fallen, as others have pointed out in this discussion. It's too complicated to simply flatten, but modern energy use does seem to broadly resemble a logistic curve. --- [1] https://en.wikipedia.org/wiki/Logistic_function https://en.wikipedia.org/wiki/Logistic_function