Because it takes work and capital to replace that generation with cleaner alternatives. It's already being done; have you considered working towards speeding up the transition?
> Just one hour of videoconferencing or streaming, for example, emits 150-1,000 grams of carbon dioxide (a gallon of gasoline burned from a car emits about 8,887 grams), requires 2-12 liters of water and demands a land area adding up to about the size of an iPad Mini.
Out of curiosity what is the water used for? Is it from hydroelectric power? Also, what is "land area adding up to the size of iPad Mini" referring to?
I think a lot of these claims can make sense on paper, but often don't really add up to meaningful impact if actually implemented. Hypothetically, turning off your webcam reduces overall network traffic and would reduce the resources required to deliver adequate connection speeds. But in reality, modest reductions in internet traffic won't impact the datacenters and network infrastructures that have already been built. A more likely outcome is that network traffic is used for other applications.
Similar situations exist when pointing to the environmental impact of meat production, for example. Not consuming meat would reduce transportation of feed, and transportation accounts for almost all of the carbon emissions from meat production. But that only actually reduces emissions if we assume that the total amount of transportation would go down, rather than the same transportation capacity being used for other goods.
It doesn’t get into the details of what “processing” means, but I suspect the processing of the video data itself adds significantly to the overhead. Cameras may (probably) don’t have hardware video encoders for the best codecs, so the processing on the CPU would use quite a bit of power. And on the servers, depending on what you’re doing, there’s a lot of mixing and reprocessing that may need to be done on the streams. My own laptop isn’t that old, and the fan spins up often enough on calls that you know the CPU is chugging more than usual.
Video conferencing is real time. You don't have the latency budget to do really intensive CPU based encoding (although ibviously there is still encoding). Hardware acceleration for video encoding is also quite common in many consumer products.
The assumption that freed capacity would be gobbled up by other consumption (100% of it) is a strong one that I wouldn't agree on. Reducing consumption (e.g. internet, meat) means that less would be produced in the long run, thus it does have a positive impact on the climate. We can talk about details or exceptions but the general idea that less consumption = less consumption [sic] = less climate warming is firm. I think the greater point is that we have to look at a long time horizon with long lasting changes to really see this, that's why we as humans have trouble implementing these changes.
Cars, roads, and parking have bumped up against hard limits. There's no way to expand a lane of traffic to more than 2000 cars per hour and people generally resist multipliers like buses and subways unless driving is already painful.
Broadband has expanded several orders of magnitude. There's still the ransition from copper to fiber, from 1G to 10G, and multiple wavelengths.
But in reality, modest reductions in internet traffic won't impact the datacenters and network infrastructures that have already been built.
It’s easy to measure. Camera and incoming video on is about 30 watts more power consumption on my Mac. I can see this on the LCD on the UPS. It’s not just about the datacentre.
Sounds like a performance issue that Zoom/RingCentral need to fix. My laptop takes off even with my camera off, when there are others with their camera on. Meanwhile
VLC can play x264 movies without the fan running.
> transportation accounts for almost all of the carbon emissions from meat production.
Not at all. Transportation accounts for only a tiny fraction of the emissions from the meat production supply chain. The type of food matters much more for total emissions (equivalent) than transportation too:
First of all, a large portion of that graph refers to methane not carbon dioxide. Methane has a half life of 10 years, as compared to carbon dioxide's thousands or tens of thousands.
Also take care to inspect, carefully, what these figures represent. This graph only counts transportation of final food product to stores under the "transportation" category: "Emissions from energy use in the transport of food items". In other words, the emissions from taking the packets of meat from a slaughterhouse to grocery store shelves. Not shipping loads of grain to ranches. The actual article [1] this image comes from explains this more clearly, "Food transport accounted for only 5% of this. This means that if we were to take the case where we assume a household sources all of their food locally, the maximum reduction in their footprint would be 5%". Transportation of feed to ranches is categorized under feed, not transportation.
This is why estimates for the amount of carbon emissions meat is responsible often vary by a large margin. Some place meat, alone, as responsible for 14-19% of emissions. Others place all food production (not just meat) as under 5%. It largely depends on whether you include things like transportation, nitrogen fixing via the haber process, and land conversion under the umbrella of food production or in their own separate categories.
1. Land use for agriculture is going down in the US [1], as with most developed countries. Factory farming is more intensive per acre, leading to fewer. Likewise forests in Europe are expanding [2]. Industrialized agriculture leads to less land cultivated because of its greater production per unit of land. Much of the land used for agriculture today was deforested over 70 years ago.
2. NASA puts the estimate at 300-1000 years [3], but it looks like this includes carbon sequestration as one of the decay methods not decay in the atmosphere like methane. Absent sequestration, carbon dioxide lasts a lot longer.
3 & 4. While methane has greater short term effect, the total amount of methane being released is orders of magnitude less. About 2.5 ppm of carbon dioxide is added each year, as compared to methane's 5-7 ppb added each year. Note the difference in units, parts per million versus parts per billion. Sure it's 25x greater impact, but it's emitted 500x less and decays 30-100x faster. That's why methane's contribution to thermal imbalance is relatively static while carbon dioxide's is increasing [4]. Carbon dioxide is overwhelmingly the cause of climate change.
>Not consuming meat would reduce transportation of feed
Not consuming meat would also reduce its price. This would enable people who couldn't eat as much meat before to eat more meat. It might have a smaller impact on developed nations, but likely a larger one in developing nations.
Doesn’t this work out to just mean our electronics grid needs to become more carbon neutral, especially in places that house internet infrastructure and data centers?
It doesn’t seem like “use less YouTube and video chat” is the most impactful answer here.
Also, the comparison to 1 gallon of gas is a bit disingenuous when a car probably uses 2 or more gallons of gas when driving continuously for an hour (60mph, 30mpg)
Talk about missing the forest for the trees. Turning off ones camera will make business leaders less willing to permanently virtualize some meetings. Getting on a plane, riding a taxi, and using a hotel & convention center is a lot more emissions than video chat.
Exactly. I’d need to do some homework to prove it, but my theory is more power is used every second for Bitcoin than an hour’s worth of videoconferencing worldwide. That said, we have to start somewhere, but is videoconferencing really the thing that’ll have any positive impact on CO2 emissions?
Edit:
Per this source[0] (no claims to its accuracy), a single transaction on Bitcoin causes nearly 300kg of CO2 to be emitted. So compared even to the high point of the article’s range (aka 1kg per hour meeting, though the included chart shows ~1/10th that for Zoom), that’s multiple orders of magnitude higher, per transaction.
I did some further research and multiple seemingly solid sources give conservative estimates of Bitcoin electricity consumption in the range of 80-100TWh per annum. To put that number in perspective, that’s more electricity per year than all of Google’s servers worldwide combined since its founding over twenty years ago (estimates put that number around 75TWh). I’m not going to keep digging, but based on the numbers I saw, I have a feeling Bitcoin uses more electricity per year than AWS, Google, FB, Microsoft, and Apple’s data centers combined per year. I’ll leave it to you to decide if Bitcoin is worth it and you should just turn off your webcam next time you’re on a videoconference to save the environment!
Bitcoin's high electricity usage comes from finding the block (header), not from processing the transactions inside it, i.e. the same energy is used to find a block with 0 transactions or 10,000 transactions.
So, more transactions don't cause higher electricity usage.
All miners run similar hardware and software so the only way to be competitive is to find cheaper energy sources, which is why Bitcoin mining scoops up energy that is either very cheap, renewable or would be wasted otherwise (e.g. remote hydro power plants).
So, electricity consumed on the Bitcoin network has a much smaller CO2/Wh impact than what you would expect on average.
If any publication fails to acknowledge any of these key points in estimating Bitcoin's environmental impact, I would be very cautious in trusting their conclusions.
You really shouldn't trust articles that use metrics like CO2/tx or that claim to have "proven" Bitcoin's environmental impact. At the same time you shouldn't use such a "proof" as an excuse to dismiss conflicting information.
Bitcoin mining is a very competitive field. Information about the actual energy mix and location of miners are guarded secrets. Researchers are thus using models to estimate these holes.
Estimates on the electricity consumption of Bitcoin in 2018 are ranging from ~20 TWh/yr to >100 TWh/yr. Some papers assume coal fossil power only (0% renewables), global average energy mix (~10% renewables) or a regional energy mixes (~73% renewables [0]).
Based on my experience with papers in that area, is that more recent papers with more complex models tend to produce higher estimates of the shares of renewables. It seems reasonable to me that Bitcoin is greener than many other industries because energy costs are the dominating factor in mining profitability and Bitcoin allows you to convert excess energy into money - at any remote location on this planet as long as you have an internet connection.
> You really shouldn't trust articles that use metrics like CO2/tx or that claim to have "proven" Bitcoin's environmental impact.
So instead I should trust a random person on the internet whose account was created a short time ago, frequently shills Bitcoin, and gets easily proven facts wrong about BTC and needs to be corrected by others?
Please stick to the arguments at hand and don't attack the person/character.
Sure, you could rely on account age and post history to determine if you "believe" what somebody says.
The other option would be to actually read the sources/source code and come up with an informed opinion about falsifiable statements yourself. Although this might be more exhausting and has a risk of changing your opinion it might be the more rewarding one long term.
And since it has been some time since you created your account, I'd like to recommend to brush up on the HN guidelines. I'm out.
Why not? I really dislike people who don't turn on their video, which it was enforced. If you get the luxury of sitting at home, should at least be polite and show your face
>If you get the luxury of sitting at home, should at least be polite and show your face
So if I'm in the office I don't have to show my face? I work at a company that has offices in I honestly don't even know how many countries and pre-covid I had no need to see someones face from Germany, India, Japan, America, etc.
Working from home doesn't change that in my mind. If you're not in my house or my office, I have no need to see you.
I work in a remote-only team since mid-2019 and we've never used video in any meeting (even including the job interview). Why would we? On average there's one audio meeting per week; all the rest is done over chat, mail and issue tracker.
Video in meetings that don't require it is just a stressful distraction.
That is a wise move! my team is total remote, and we do audio only meetings, it works just right. Video calls are making everyone feel the way people on TV feel every slouchy every mumble every interruption is magnified. What do you use for audio calls? Zoom? Skype? Tappy is perfect for Audio-only meetings. Very simple and light, the audio quality is better than Zoom. https://tappy.so/
The original study says a car driving using 1 gallon of gas is 9 times worse than running video conferencing for an hour. My car burns 1.5 gallons in an hour (and that's with highway efficiency), so that means driving for an hour is 13.5 times worse than running a virtual meeting. Considering my commute used to be 45 minutes each way, videoconferencing seems pretty reasonably better than the alternative.
This is the kind of trash study, designed solely as PR fluff, that does so much damage to science. The fact that this was published almost instantly ( https://www.sciencedirect.com/science/article/abs/pii/S09213... ) in a journal with an IF of 8 makes me concerned for the entire field.
They are simply not true when you look at the data. Researchers are incentivized to look at the larger estimates along the way. They are almost always not as large as claimed.
Jobs and life is hard, working remotely takes an emotional toll in parts, putting this on top on people doesn't help their well being. There's no point saving CO2 emissions if it breaks you as a human or you have to go back to the office.
If universities really care, how about they start creating open codexes that allow more effect use of our resources rather than these studies. I get there are patent issues with things like H.266. How about universities theorise on how to fix the patent problem and how to speed up the process. We got a vaccine in a year, how do we do that in working groups?
Turning off the camera in meetings is often a good idea, see other articles on that.
This kind of analysis is pretty slippery. I live next to Lake Michigan. Power generation here doesn't consume any water, it adds a bit of heat to it and returns it to the environment in otherwise the same condition. There's likely some evaporative losses, but we have precipitation throughout the year (it gets moderately dry for a few weeks in May and sometimes in October).
The land footprints should be expressed in servings of rice or something (so that people can compare their internet usage to something they aren't going to stop doing).
Studies on how much carbon data transfer actual use tend to be all over the place, i dont think a single study should be treated all that highly until the various studies start agreeing with each other
56 comments
[ 2.5 ms ] story [ 122 ms ] threadhttp://h4labs.org/ive-got-another-stupid-idea-to-deal-with-c...
You might not have noticed but we are increasing the number of coal power plants globally.
The decrease in natural gas prices has accelerated the closing of coal in the US, ...
anyway, I would suggest that you look into this a little more.
We’ve know about this problem for quite some time
https://www.forbes.com/sites/williampentland/2014/11/30/why-...
None of these little ideas come close to the damage done by coal power
Out of curiosity what is the water used for? Is it from hydroelectric power? Also, what is "land area adding up to the size of iPad Mini" referring to?
Similar situations exist when pointing to the environmental impact of meat production, for example. Not consuming meat would reduce transportation of feed, and transportation accounts for almost all of the carbon emissions from meat production. But that only actually reduces emissions if we assume that the total amount of transportation would go down, rather than the same transportation capacity being used for other goods.
Increasing the number of lanes on a road does not reduce traffic, it adds more cars. It’s a common, unfortunate phenomenon.
Broadband has expanded several orders of magnitude. There's still the ransition from copper to fiber, from 1G to 10G, and multiple wavelengths.
The more that can be fit in that 250ms or whatever it is, the more that will be fit into it.
It’s easy to measure. Camera and incoming video on is about 30 watts more power consumption on my Mac. I can see this on the LCD on the UPS. It’s not just about the datacentre.
Not at all. Transportation accounts for only a tiny fraction of the emissions from the meat production supply chain. The type of food matters much more for total emissions (equivalent) than transportation too:
https://ourworldindata.org/uploads/2020/02/Environmental-imp...
Here is the article that this graph is taken from:
https://ourworldindata.org/food-choice-vs-eating-local
Also take care to inspect, carefully, what these figures represent. This graph only counts transportation of final food product to stores under the "transportation" category: "Emissions from energy use in the transport of food items". In other words, the emissions from taking the packets of meat from a slaughterhouse to grocery store shelves. Not shipping loads of grain to ranches. The actual article [1] this image comes from explains this more clearly, "Food transport accounted for only 5% of this. This means that if we were to take the case where we assume a household sources all of their food locally, the maximum reduction in their footprint would be 5%". Transportation of feed to ranches is categorized under feed, not transportation.
This is why estimates for the amount of carbon emissions meat is responsible often vary by a large margin. Some place meat, alone, as responsible for 14-19% of emissions. Others place all food production (not just meat) as under 5%. It largely depends on whether you include things like transportation, nitrogen fixing via the haber process, and land conversion under the umbrella of food production or in their own separate categories.
1. https://ourworldindata.org/food-choice-vs-eating-local
1. CO2 is not the sole measure of environmental impact. The land usage problem is also pretty serious. Deforestation chief among it.
2. I can't find any literature to back up the claim that CO2 has an atmospheric residency of thousands of years, seems to be 100 at best estimates [0]
3. Over that 100 year period, Methane has 25x greater effect than CO2 [1]
4. Methane oxidises into CO2 and H20, so transport is still not the greatest contributor
[0] https://scholars.unh.edu/cgi/viewcontent.cgi?httpsredir=1&ar...
[1] https://unece.org/challenge
2. NASA puts the estimate at 300-1000 years [3], but it looks like this includes carbon sequestration as one of the decay methods not decay in the atmosphere like methane. Absent sequestration, carbon dioxide lasts a lot longer.
3 & 4. While methane has greater short term effect, the total amount of methane being released is orders of magnitude less. About 2.5 ppm of carbon dioxide is added each year, as compared to methane's 5-7 ppb added each year. Note the difference in units, parts per million versus parts per billion. Sure it's 25x greater impact, but it's emitted 500x less and decays 30-100x faster. That's why methane's contribution to thermal imbalance is relatively static while carbon dioxide's is increasing [4]. Carbon dioxide is overwhelmingly the cause of climate change.
1. https://www.ers.usda.gov/amber-waves/2017/december/a-primer-...
2. https://www.weforum.org/agenda/2019/07/forest-europe-environ...
3. https://climate.nasa.gov/news/2915/the-atmosphere-getting-a-...
4. https://www.climate.gov/news-features/understanding-climate/...
Not consuming meat would also reduce its price. This would enable people who couldn't eat as much meat before to eat more meat. It might have a smaller impact on developed nations, but likely a larger one in developing nations.
It doesn’t seem like “use less YouTube and video chat” is the most impactful answer here.
Also, the comparison to 1 gallon of gas is a bit disingenuous when a car probably uses 2 or more gallons of gas when driving continuously for an hour (60mph, 30mpg)
Exactly. I’d need to do some homework to prove it, but my theory is more power is used every second for Bitcoin than an hour’s worth of videoconferencing worldwide. That said, we have to start somewhere, but is videoconferencing really the thing that’ll have any positive impact on CO2 emissions?
Edit:
Per this source[0] (no claims to its accuracy), a single transaction on Bitcoin causes nearly 300kg of CO2 to be emitted. So compared even to the high point of the article’s range (aka 1kg per hour meeting, though the included chart shows ~1/10th that for Zoom), that’s multiple orders of magnitude higher, per transaction.
[0] https://digiconomist.net/bitcoin-energy-consumption
Edit 2:
I did some further research and multiple seemingly solid sources give conservative estimates of Bitcoin electricity consumption in the range of 80-100TWh per annum. To put that number in perspective, that’s more electricity per year than all of Google’s servers worldwide combined since its founding over twenty years ago (estimates put that number around 75TWh). I’m not going to keep digging, but based on the numbers I saw, I have a feeling Bitcoin uses more electricity per year than AWS, Google, FB, Microsoft, and Apple’s data centers combined per year. I’ll leave it to you to decide if Bitcoin is worth it and you should just turn off your webcam next time you’re on a videoconference to save the environment!
So, more transactions don't cause higher electricity usage.
All miners run similar hardware and software so the only way to be competitive is to find cheaper energy sources, which is why Bitcoin mining scoops up energy that is either very cheap, renewable or would be wasted otherwise (e.g. remote hydro power plants).
So, electricity consumed on the Bitcoin network has a much smaller CO2/Wh impact than what you would expect on average.
If any publication fails to acknowledge any of these key points in estimating Bitcoin's environmental impact, I would be very cautious in trusting their conclusions.
You should’ve stopped at cheap, because the second too are conclusively wrong and well proven.
Bitcoin mining is a very competitive field. Information about the actual energy mix and location of miners are guarded secrets. Researchers are thus using models to estimate these holes.
Estimates on the electricity consumption of Bitcoin in 2018 are ranging from ~20 TWh/yr to >100 TWh/yr. Some papers assume coal fossil power only (0% renewables), global average energy mix (~10% renewables) or a regional energy mixes (~73% renewables [0]).
Based on my experience with papers in that area, is that more recent papers with more complex models tend to produce higher estimates of the shares of renewables. It seems reasonable to me that Bitcoin is greener than many other industries because energy costs are the dominating factor in mining profitability and Bitcoin allows you to convert excess energy into money - at any remote location on this planet as long as you have an internet connection.
[0]: https://coinshares.com/assets/resources/Research/bitcoin-min...
So instead I should trust a random person on the internet whose account was created a short time ago, frequently shills Bitcoin, and gets easily proven facts wrong about BTC and needs to be corrected by others?
Sure, you could rely on account age and post history to determine if you "believe" what somebody says.
The other option would be to actually read the sources/source code and come up with an informed opinion about falsifiable statements yourself. Although this might be more exhausting and has a risk of changing your opinion it might be the more rewarding one long term.
And since it has been some time since you created your account, I'd like to recommend to brush up on the HN guidelines. I'm out.
So if I'm in the office I don't have to show my face? I work at a company that has offices in I honestly don't even know how many countries and pre-covid I had no need to see someones face from Germany, India, Japan, America, etc.
Working from home doesn't change that in my mind. If you're not in my house or my office, I have no need to see you.
Video in meetings that don't require it is just a stressful distraction.
They are simply not true when you look at the data. Researchers are incentivized to look at the larger estimates along the way. They are almost always not as large as claimed.
Jobs and life is hard, working remotely takes an emotional toll in parts, putting this on top on people doesn't help their well being. There's no point saving CO2 emissions if it breaks you as a human or you have to go back to the office.
If universities really care, how about they start creating open codexes that allow more effect use of our resources rather than these studies. I get there are patent issues with things like H.266. How about universities theorise on how to fix the patent problem and how to speed up the process. We got a vaccine in a year, how do we do that in working groups?
Turning off the camera in meetings is often a good idea, see other articles on that.
The land footprints should be expressed in servings of rice or something (so that people can compare their internet usage to something they aren't going to stop doing).