If you actually believe that software has the ability to change the world, this is it in action.
[0] https://www.eia.gov/totalenergy/data/browser/index.php?tbl=T...
Your own linked data from the EIA says that you're wrong.
Peak energy consumption per capita 1979: 359m btu
Energy consumption per capita 2000: 350m btu
Energy consumption per capita 2007: 335m btu
Energy consumption per capita 2018: 309b btu
That's an extremely clear downtrend over time.
If that is all the energy saved by moving all the heavy industry and manufacturing to the developing world, the overall up trend must be enormous.
Edit: since people are asking for examples, I would have to point to the insane cost of training modern ML models. For instance, according to [1], training a single big NLP model can burn up to 620,000 lbs of CO2.
I say a lot of energy consumption is hidden because it's generally moved to data centres, which by the way use up astronomical amounts of energy [2]
[1] https://arxiv.org/abs/1906.02243
[2] https://ec.europa.eu/jrc/en/publication/trends-data-centre-e...
Old Z80 based board I first worked on drew 1A at 5V, so 5W. Put your hand on it, it's nice a warm. Current board I'm dicking with draws 8ma at 3V or 0.024W. 200 times less power. But it's 50 to 100 times faster.
Price isn't good approximation either, as it's pretty much arbitrary - prices are determined by trade, fluctuate all the time. Regular people only see a smoothed-out, low frequency version of that.
--
[0] - 8.6 megapersons living in New York * 12 MWh US average electricity use per person per year.
Since energy suppliers meter their output to apply a price, I'm wondering how hidden costs could exist (unless the software is somehow circumventing or undermining the meters?).
I would think it'd be pretty straight-forward experiment to use software from the late 80s / early 90s for a couple weeks, measure machine energy consumption, then compare with the consumption from a machine running the latest version of everything for two weeks. Next month repeat but also use a processor from the 90s for the first two weeks. Rinse, repeat, extrapolate, apply statistics, &c. Maybe try adjusting for or comparing differences in appliances common then and now.
If software energy costs are somehow hidden, I can only think they're hidden because we're deliberately not choosing to look at them.
You're going to have to explain that one. We can pretty trivially measure energy usage...
Fortnite? Minecraft? AirBnB? Google? Android? iOS?
If it’s trivial, what are the numbers and what’s your confidence in them?
That's not 'an NLP model', it's one of the largest and baddest NLP models that has to be trained only once and can be reused 1000 times over by fine-tuning on different tasks. The paper was hyperbolising about energy use. The AI labs that can even attempt that kind of things can be counted on the fingers of one hand.
A regular NLP model trains in minutes up to a day on a single GPU, so it uses energy on the level of personal game play. And let's not forget that ML can solve difficult math problems with as little as 1/100 of the energy of classical computing by using smart approximation methods. Neural nets also have a nice degradation effect where you can shrink the net by 20x and lose just a few percentage points of accuracy.
That's why Google's speech recognition and synthesis can now run on a single cell phone instead of a datacenter. And we haven't really started building specialised AI chips that will reduce energy usage even further - we're still using mostly GPUs.
Can it? I thought it still sent out input to the cloud for processing?
https://yearbook.enerdata.net/total-energy/world-energy-inte...
Do 5 big-$ bank transactions and you contribute to GDP. What was the energy usage? What was the productive output?
Compare say a CRT from 1999 with an LED TV from 2019 and the vastly larger sizes are more than offset via increased efficiency.
You happen to have any links to attempts to disambiguate the two? I definitely don't.
By comparison in constant 2010 dollars global GDP went from 66T (2010) to 82.6T (2018). https://data.worldbank.org/indicator/ny.gdp.mktp.kd
Energy depends a little on the numbers you use BTU etc. However as the mix of services is not really changing very rapidly but the GDP PPP per oil equivalent has been growing quickly the trend is clear: https://data.worldbank.org/indicator/EG.GDP.PUSE.KO.PP?end=2...
I think the US economy probably had a bigger than 2% shift just into tech in the last 10 years, let alone if you include broader service-ization as a national goal for decades now.
I guess I'm saying, if our stated goal and the goal of every MBA in the country was to have more money in services.. I thought that happened.
EDIT: rate-limited, appreciate the follow-up on US numbers. Still not liking 2010 as a start-date as I think the relevant changes were before then. Would love to chat more, but hey.
But, I was using global GDP to avoid the outsourcing question.
PS: Using 2010 as the start date as these energy trends are fairly recent. Though I agree comparing say 2018 numbers to say the 1970’s would need to account for more systematic changes.
Hand wave about that all you want.
Hand wave about that all you want.
Plus "quality of life" is subjective. Some people (or some countries/cultures) can think about all kind of crap as essential to their "quality of life".
Hand wave about that all you want too.
A lot of GDP growth is in box-moving, paper shuffling, services, sales and so on, which can be made with "less stuff" (e.g some even completely online).
OTOH, tons of industrial production for US companies, the dirty kind, has been moved to China and co.
Even if the US "manufacture's more units than ever" (and what those units are? If it manufactures more e.g. packaged food than ever, it wouldn't really be as polluting as manufacturing more cars or computers/mobiles, or several other kinds of products, than ever). I'd wager that the more pollutant and messy production has been moved to China / the developing world / etc.
The same way as most of the western "recycled" stuff just goes to the developing world to be pillages for parts and then dumped into landfills and rivers...
But even taking that into account it actually happened. Electronics, lighting, and appliances got vastly more efficient. Aircraft, AC, and automobiles improved though to a lesser extent. Even when you look a fossil fuel use per lb of food grown you see a huge impact. Further, many services like medicine are simply not major energy consumers.
PS: This is really a continuation of a long term trend. Humanity has always harnessed most of it’s energy extremely inefficiency via plant photosynthesis. However, the amount of food per acre has steadily grown across thousands of years.
A more interesting metric would be calories-harvested over calories-input.
Food per acre has certainly increased. A lot of the increase is through selective breeding, GM crops, and other technological efficiencies, but a lot of it is from increased fertilizer, which is just calories/energy in a different form (fertilizer is just a store of energy -- that's literally why you buy too much of it in bulk and you're put on a fertilizer-bomb watchlist).
You can grow more food per acre by throwing more energy at the problem, but that doesn't necessarily mean you're doing it more efficiently or sustainably. One way of looking at all the fertilizer runoff, for example, is it is quite literally leaking input energy out of your harvest, decreasing your efficiency.
IMO, a more interesting metric is nutrients like phosphorous which would be more difficult to acquire once current sources have mined out.
Source: recently read The Quest, book which discusses energy in the modern world
The upside is that some of the children of peasants are putting a serious dent in world climate problems like Wang Chuanfu who is the founder of BYD.
How dare you question woke capitalism, Mr. 'api'. We want to be told we're building a better world, and we'll hang anyone who says otherwise. Didn't you see Silicon Valley?
Yet it took the form of a question. It doesn't even try to support its implied criticism with any data or argument.
As such, it's a cheap shot that doesn't add anything. If you believe someone's work, published in a somewhat reputable outlet (and a book), to be wrong, find something that supports that criticism. Don't just throw out some lame counterargument that anybody reading the headline can come up.
Specifically, the article uses examples that are unlikely to be outsourced; Paper and cardboard, especially, aren't major inputs to big-hairy-manufacturing.
Fertilizer is an even better case, because it's impossible to outsource its use without also outsourcing agriculture. Agricultural output has obviously been growing steadily. And those statistics are rather reliable because the products are standardised, often need to be reported for various subsidies or regulatory purposes, and are traded on public markets.
Also strongly agreed about timezone karma - I would love to see some analysis of how upvotes in discussions like this fluctuate based on which part of the planet is exposed to sun at a given time.
For most Americans life expectancy and qualify of life are dropping measurably. By your logic we should increase energy consumption to reverse these losses.
Half of that is easily demonstrably false. Quality of life for the majority of Americans continues to improve.
The US welfare state has improved considerably in the last 30 to 50 years.
- Homelessness has been slashed by 20-25% in two decades thanks to housing programs the Bush and Obama administrations pursued. Homelessness nationally is sitting near the lowest levels in US history.
- Childhood poverty has collapsed by roughly 45-50% in 40 years.
- The overall poverty rate at 11.8% is near an all-time record low. That figure is lower than in either Canada or France today. In 1960 the poverty rate was 100% higher than it is now.
- The poorest 25% of Americans now get free healthcare, that was not true at all at the supposed peak of US well-being and financial per capita outcomes in the ~1965-1980 time frame. The fact is, the bottom 1/3 are far better off today than they were in the 1960s and 1970s. The homelessness and poverty rates demonstrate that.
- The middle class is expanding at the fastest rate since the 1960s. More income as a distribution share is finally beginning to shift from corporate profit to labor.
- For the last several years wages for the bottom 2/3 of labor have been rising dramatically faster than wages for the top 1/3.
- The US savings rate is strong at 8%. Household balance sheets are similarly strong. The share of household income going to debt service costs is near 40-50 year lows.
Several of those are entirely due to a continued resource shift via the rapidly expanding US welfare state and increasing progressive taxation.
You can't just flat out lie if you want to be effective at propaganda.
As per Wikipedia [0]: "Medicaid is the largest source of funding for medical and health-related services for people with low income in the United States, providing free health insurance to 74 million low-income and disabled people (23% of Americans) as of 2017"
So flat out lie?