Jevons Paradox is conjecture, not immutable law. Consumption based energy usage has been on the trend downwards in western countries for some time. Much of this is explained by efficiency gains (e.g., LED lighting, OLED screens, smaller chips, heat pumps, etc.)
https://ourworldindata.org/grapher/consumption-based-energy-...
the jevons paradox certainly isn't an immutable law
It's hard to disentangle the rise of wealth in the world from increased consumption caused by cheaper energy. Certainly in the west capacity has become larger and...
> the link doesn't show that even that cherry-picked measure is falling
If you look again, they absolutely are.
Peak vs now in Mwh: US: 105 (2005) vs 97 Norway: 93 (2008) vs 78 Sweden: 69 (2001) vs 57 Australia: 80 (2011) vs 55 Germany: 63 (2006) vs 59 France: 61 (2005) vs 54
I could manually do this for the rest, but I shouldn't need to. You can see it for yourself. Here's the chart as line graphs (this option isn't obvious in the UI)
https://ourworldindata.org/grapher/consumption-based-energy-...
The only outlier I see here, in terms of developed countries, is South Korea, but that's perhaps as the country is still unevenly developed. In terms of developing countries, we of course see the uptake of HVAC, refrigeration, vehicles and other automation appliances to match the living standards of the developing world. But crucially, developed countries are showing that at the highest living standards, energy consumption is coming down.
that's the chart i was looking at. you're cherry-picking years as well to get that answer. looking at all the years instead, the usa starts out at 94, peaks at 104, dips down to 87, and then rises back up to 97, with lots of wiggles in between. that isn't a picture of a secular decline in energy use, it's a picture of random fluctuation around an average. the others are largely similar. the main outlier among 'developed' countries is not south korea but the people's republic of china, though slovenia deserves a mention for falling 40%
why? what holds the usa line so steady? probably partly the usa has lost the ability to build not only high-speed trains, subway tunnels, highways, and leading semiconductor fabs, but also electrical infrastructure, but two thirds of energy use is non-electric, and other countries are succeeding in building trains and whatnot
the price of energy has not been holding steady during this time but rather going up, although in the last decade that is starting to reverse. perhaps the reason consumption-oriented energy use has gone slightly up rather than way down is the efficiency improvements you mention
[the following paragraph is completely wrong, and i appreciate the correction from h0l0cube. therefore so is my claim that the measure itself is cherry-picked. i had just misunderstood it]
this measure, incidentally, doesn't count energy used for exports at all. so if the us and japan are producing an energy-intensive good independently at the beginning of the chart, and then in the middle the us starts importing it from japan, the usa's line will go down while japan's remains steady, because japan's added energy use is for export rather than for consumption. that's what i mean about it being a cherry-picked measure even for those cherry-picked countries
Literally, it does. It's consumption-based energy usage. It's trade adjusted
> the usa starts out at 94, peaks at 104, dips down to 87, and then rises back up to 97, with lots of wiggles in between. that isn't a picture of a secular decline in energy use, it's a picture of random fluctuation around an average
The trend line is at best a flat line. Here's a picture of energy usage per capita over a longer timespan (inc. prior to the 80s when offshoring of manufacturing and cheap shipping came into play)
https://ourworldindata.org/grapher/per-capita-energy-use?tab...
> the main outlier among 'developed' countries is not south korea but the people's republic of china
As I mentioned this is due to it's growing wealth, not the greater economy of power. Even though China has been growing phenomenally in GDP, it's GDP per capita vs consumption-based energy usage is going down (or yet to exceed 2011 levels):
https://ourworldindata.org/grapher/change-energy-gdp-per-cap...
thank you for the correction; i had misunderstood what the trade adjustment was
> The trend line is at best a flat line
i think that is an excellent description of it, as well as of the expanded and less debatable chart you link here, at least over the last 50 years, since the energy crisis began. before that we were seeing a much different trend
i agree about wealth being the primary driver of energy use. but that's precisely the story jevons tells: you improve your steam engine to use less coal, so now you can build a railroad, which is a form of wealth. previously a tonne of coal cost £100, say, and produced 30 megajoules of work, which had a value of £200 in pumping out a mine or £50 hauling goods on the railroad. with your new engine you improve efficiency to 0.3,% and get 100 megajoules, so pumping out the mine now costs £30 per £200 produced, but now the railroad is viable because it produces £50 - £30 = £20 net. the railroad consumes much more coal than the mine did, so you're using more energy more efficiently at the level of the machine, raising your gdp, but producing less value per tonne of coal
well, i got the numbers a bit wrong, but hopefully you can see what i mean
That's not what I meant. I was referring to the development of populations. Populations that didn't have what the global north would call 'the basics', electricity, household water/sanitation, lighting, adequate heating/cooling, refrigeration, as well as 'basic luxuries' like television/computing/internet and personal transport, increasingly now have access to these technologies (even in 'underdeveloped' countries). This is going to increase energy (not just electricity) usage. But once 'the basics' have been met, they will already be in line with the efficiency standards adopted by the west (maybe even moreso because energy is more expensive in the global south)
So, a counterargument might be that AI will become part of our 'basic' lifestyle, and that we will see a resurgence of demand again, but we're seeing that cloud based compute is acceptable for the vast majority. So even though, say, internet search was energy intensive at its inception, it has largely been amortized and itself energy optimized to not really raise the bar of energy usage. Custom silicon, tighter semiconductor nodes, newer algorithms, photonics, and eventually yet-to-be-discovered technologies like room temperature super conductors can again bring the energy usage down for compute, increasing efficiency not just for AI, but across the board.
things that already exist but could get much cheaper due to cheaper energy might include ai, as you suggest, but also personal computers, oocyte cryopreservation, ecm machining, space travel, weekly air travel, personal helicopters, atmospheric carbon capture, caribbean cruises, making things out of aluminum or titanium instead of plastic, cnc machining rather than casting or stamping, cars, buildings, photovoltaic panels themselves, etc. and presumably there are other things that haven't been invented yet because they'd be uneconomic
Anyone who remembers how pervasive CRTs (later rear projection, and plasma TVs) and incandescent bulbs were, can sense that things are far more efficient around the home. Better standards for HVAC (using heat pumps vs heating elements) are a big deal. A big driver of this has been large buildings where the cost reductions from efficiency can mean millions saved. Business errs towards efficiency. Another driver is that the more efficient stuff also seems to just be better too (like OLED screens)
We also have chips that run on about the same energy (or even less) and provide far more compute than 2 decades ago. Looking further back the ENIAC used 174kw of power for 0.005 MIPS (5000 additions). By comparison an M1 has 2.6 TFLOPS (2.6 trilling floating point operations) for 40-100 watts. My cores are mostly idling. In fact most of my computer usage happens on my phone which runs on far less energy still.
Then consider energy for transport. In the last 5 years, telecommuting has become so normal that fewer people are commuting to work, particularly cities.
not all consumer energy prices are rising, and the situation you describe where the producer captures the whole consumer surplus only happens in the absence of competition
jevons says people will compensate for more efficient tvs by buying more and bigger tvs, and for more efficient hvac by building bigger houses, living in hotter areas, and installing air conditioning in more spaces. this has been a major trend of the last 25 years you're talking about, in fact
similarly, it's true that the amount of energy per flop is going down, though not nearly as much as you might think; on the cpu it was about 2000 picojoules 30 years ago, and closer to 500 picojoules today. (the numbers you give for the m1 would work out to 16 picojoules, but the real numbers are even better. 2.6 teraflops is the graphics card, which uses 11.5 watts, which is 4 picojoules per flop. most computers lag far, far behind that in efficiency.) but the amount of energy spent on computers keeps going up and up
Collusion is the norm. Especially with electricity suppliers where they jack up the rates because most people won't be bothered to check for better deals and switch.
> jevons says people will compensate for more efficient tvs by buying more and bigger tvs
Doesn't matter if the efficiency gains outcompete the consumption increase (which is my whole point):
> As shown in the graph below, between 2006 and 2012 the average energy consumption of televisions dropped by 57%, while at the same time average screen size increased by 30% and average price decreased by almost two-thirds.
https://appliance-standards.org/blog/why-recent-progress-tel...
> for more efficient hvac by building bigger houses, living in hotter areas, and installing air conditioning in more spaces. this has been a major trend of the last 25 years you're talking about, in fact
McMansions became a thing leading up to the 2008 financial crisis, and was a feature of cheap finance, not energy cost. Most people want a house that's affordable, close to amenities and doesn't take an army of servants to clean – so a house in the suburbs and within distance of the city. The factors here are land price and construction costs.
> but the amount of energy spent on computers keeps going up and up
Might be true, but it would be nice to see some data to back that claim.
a 130% increase in tvs per person would cancel out the 57% reduction you cite in power per tv, and presumably if you measure over any other period of time, the energy consumption drop won't be that large, because that was the crt–lcd transition, but screens had been steadily getting bigger for decades and have continued to do so. if you include computer monitors with tvs, it seems likely that the increase over the last 25 years is a lot more than 130%. certainly it is in public places (dentists' offices, intercity buses, airports, restaurants)
people have been building continuously bigger houses for centuries; it's not a phenomenon limited to zero interest rate mcmansions. suburban sprawl is one well-known manifestation of it as for computers, unfortunately i don't have the data here at the moment. you see handwringing about it from time to time. i'd like to see it too
Is each person watching all these TVs at once Matrix style?
Even if you've got your laptop, iPhone and iPad all unlocked and playing videos while watching your TV, it's not really adding 30% of energy use compared to the 50 inch TV.
> certainly it is in public places (dentists' offices, intercity buses, airports, restaurants)
TVs in airports, bars, diners, buses, and dentists offices have been around since the 80s. The difference now is that all their CRTs are replaced — a clear energy drop. (Source: own experience, also you can watch movies from the time)
Lit billboards have been around for a long time, but instead of massive lamps, they are just LED screens. Tokyo used to be all fluorescent bulbs, now LED screens again. It might not be like for like, could maybe be an energy increase, but public-space advertising is the best argument you've got here, but I don't buy that it's a huge driver of energy growth.
> you see handwringing about it from time to time. i'd like to see it too
I've provided plenty of data to back myself, and to suggest otherwise is disingenuous. Given the amount of claims you're making without any supporting evidence, I can only think your argument is based on conviction, not fact. I'll let you have the last word here.
i've seen many more tvs in public places in recent decades, often replacing things like flip-dot displays or printed menuboards. mcdonalds now displays their menus on a wall of tvs instead of printed on plastic, for example
it's disappointing that you've chosen to attack my integrity, but it seems to be based on a misunderstanding. i have not claimed that you have not provided data. on the contrary, i have found your data highly informative and educational. i only meant that neither of us has convincing data about total computer power consumption, and that i would like to
the price of energy has started to fall again in the last ten years