4 ms·
You can go millions of prompts before you use up as much water as it took to make a single beef burger. You can go tens of thousands of prompts to match the C0
by adjejmxbdjdn 6mo ago
You can go millions of prompts before you use up as much water as it took to make a single beef burger.
You can go tens of thousands of prompts to match the C02 emissions.
There are many legitimate concerns around AI. Water use/CO2 emissions isn’t currently one of them. Going vegan will make up your AI water consumption/CO2
Emissions many thousands of times over.
- SecretDreams 6mo agoWater I agree. C02 (which is really a tangential metric if energy consumption which will vary by energy mix) I'd want some citations. Also agree there are other ways we should pursue in parallel regarding emissions.
- LinXitoW 6mo agoObviously, there's different options and variables and bla bla bla, but considering how consolidated and highly industrialized and standardized meat production is, this data is very likely close enough to true for the wast majority of beef burgers eaten by the people complaining about AI resource consumption: https://ourworldindata.org/environmental-impacts-of-food https://ourworldindata.org/environmental-impacts-of-food
- SecretDreams 6mo agoI was moreso asking you about your data on how much AI is tied to CO2...
- Matticus_Rex 6mo agoAndy Masley does some plausible estimates here based on the data we have that puts 50 prompts per day at around 5kg CO2e/year: https://www.andymasley.com/writing/whats-the-full-hidden-climate-cost/ https://www.andymasley.com/writing/whats-the-full-hidden-cli... The difference between an average diet and a vegan diet via Scarborough et al. 2023/Poore & Nemecek 2018 is in the realm of 1450kg CO2e/year. Assuming those numbers, that difference is around 14,500 prompts per day, or ~5.3M prompts per year. So unless the prompt estimates are off by more than two orders of magnitude...
- SecretDreams 6mo agoThe premise of your link is founded on the energy associated to with a single prompt. The source in your link for that energy claim links to a blog post that then links back to an earlier blog post from the original author of the link you provided (it's basically a circular reference). Basically, there's a lot of words in your initial link, but they all hinge on the readers taken the stated energy assumption for a single (undefined) prompt at face value. If that initial assumption is wrong (at min, it's poorly defined in your link) all further conclusions are invalid.any a scientific publication have done this same trickery =]. They don't define what a query is when they are talking about AI power usage. If we want to get serious, we'd tie usage to tokens since we can actually track token usage.
- ToValueFunfetti 6mo ago>The source in your link for that energy claim links to a blog post that then links back to an earlier blog post from the original author of the link you provided (it's basically a circular reference). Huh? The latter blog post does link to the former's blog, but not as a source for that claim. It cites an Altman blog, an estimate from EpochAI, an article in the MIT Technology Review (albeit one that estimates 3x higher), and a paper put out by Google. It's really surprisingly well cited and I don't know how you came away from it thinking it was a circular reference. The google study is in the subheading!
- SecretDreams 6mo agoOrder of operations: 1) I click your link 2) I click the link associated with the 0.3 Wh of energy claim in the section "The full cost of a prompt". 3) The link from 2) takes me to a blog post from Hannah Ritchie. In Hannah's post, I click a link associated with the following excerpt: "Third, as a result, more recent estimates suggested that the assumptions I relied on (h/t to Andy Masley’s work on this) — that one standard query used 3 watt-hours (Wh) of electricity — were possibly an order of magnitude too high. In this case, I was happy to be conservative and overestimate the energy use." 4) This link takes me to the author of your original post, but earlier. None of this quantifies cost per token, which is really the much more relevant metric than whatever a "cost per text based query" means => which I think is both quite broad and quite model dependent.
- catlikesshrimp 6mo agoSource? Meat can be "produced" in a location where water is not as scarce. Rural areas. Datacenters "like" to grow in urban areas. This source says that a 100 prompt spends half a liter of water https://www.eesi.org/articles/view/data-centers-and-water-consumption https://www.eesi.org/articles/view/data-centers-and-water-co... I remember this year google reported one google search spend a drop of water (or 5 drops, around that)
- LinXitoW 6mo agohttps://ourworldindata.org/environmental-impacts-of-food https://ourworldindata.org/environmental-impacts-of-food
- 52-6F-62 6mo agoI think good faith would request that the source used for these kinds of questions is not one of the VC firms at the root of these questions. Doubly so when they use such innocuous and authoritative titling as "Our World in Data" which implies some collectivist, community-based outlook that this website is indeed not. To wit, this page is produced in part by the Global Change Data Lab which is a team of economists, and YCombinator.
- rick_dalton 6mo agoOurworldindata basically just uses data from published research papers and makes interactive graphs that are easy to understand. They also cite their source in every graph and every article. Trying to paint them as disingenious is pretty baseless, you would have to take it up with the authors of the source data and not owid.
- 52-6F-62 6mo agoIt's literally funded and published by the VCs backing so many of the AI companies here. Literally the fountain of money that produced Sam Altman. I'm going to have to ask for a little more rigour than "they use data". Data can be wielded in all manner of manipulative fashion to suit a preferred narrative. And when it's wielded by engineers defending large AI (or any) companies, they should at least acknowledge the fact that the data reporting is "coming from inside the house", so to speak.
- trollbridge 6mo agoI have a few cows and rarely ever give them water. In the winter they get enough from snow and when it’s rainy we have a small pond that forms with a stream. They also prefer either of those to drinking well water from a cattle waterer. They are grass fed and rarely get fed stuff like corn. For for thousands of pounds of beef, I’ve barely used any water at all. Don’t notice the extra consumption on my well at all, and I have a very low producing spring fed well (1 gallon per minute). “Vegan” crops on the other hand line corn which are irrigated in many parts of the country use a great deal of water and often very inefficiently so.
- LinXitoW 6mo agoWell, I've got a small server rack and roof top solar, therefore data centers don't actually use water. In other words, bringing up some anecdotal, hyper specific (how many meat eaters just "have a few cows"?) information says absolutely nothing about the truth of the matter, but a lot about what you believe constitutes an argument.
- alphawhisky 6mo agoA third perspective here, but maybe small ownership of these things allows for best practices (i.e. small farmers are greener and care about passing arable land to the next generation, small server owners care more about total system ownership which necessitates alternative energy production and making use of hardware that would otherwise be trashed). I think you're both onto something, now kiss!
- 0xbadcafebee 6mo ago[dead]
- adrr 6mo agoDepends on the prompt. Do a video prompt and one 30 second video will use as much electricity as running your microwave on high for 15 minutes.
- pembrook 6mo agoMy guess is that's off by a bit, but sure let's assume that's true. Now measure the amount of electricity the same prompt will use in 6 years when both algorithmic efficiency and 3-4 generations of silicon lower that by 95% (or more). Will your microwave become 95% more efficient over the next 6 years? No. Also how many video prompts will the average person run in a given year? Almost certainly 0. I heavily use AI daily and have probably played with AI video less than 4 times, ever. Yet certainly the average person will use 20,000-100,000 microwave minutes over their lifetime. I use my microwave for 2-3 minutes every day at lunch for example. From first principles, the idea that electricity use = bad is wrong. If your electricity comes from burning coal or lignite, then obviously yes using that electricity has bad externalities. But a french person running their microwave on Nuclear powered grids? This is good. Dirty energy sources is the problem. Not energy use itself.
- adrr 6mo agoAre these companies going toss a $500b+ infrastructure investment away in next 6 years? Whats the average lifespan of a AI compute node?
- pembrook 6mo agoObviously no. AI is nowhere near as ubiquitous as the microwave so adoption is still scaling. But as chips improve and the algorithms improve (eg. a paper just came out about getting the same results with 90% less inference using a few algorithmic techniques...on top of the fact we've already had multiple 90% efficiency jumps in AI already) the energy use per prompt will drop over time. Meanwhile energy use per microwave minute will not meaningfully improve over time. So to make the comparison is silly. And to pretend like the efficiency of AI will never improve given it runs on compute which by definition constantly becomes more efficient, is dumb.
- greekrich92 6mo agoOk great let's get rid of non-renewable powered AI _and_ stop eating animals.
- taejavu 6mo agoHow many gallons of water are you assuming are in that burger patty? Because, while cows do drink water that falls from the sky and lands on fields where it was going to land anyway, they also urinate most of it back out. Very little water that would have been used for any other purpose, or isn’t naturally returned to the water cycle, ends up being consumed in the production of the burger patty. To be clear I think your point about AI not consuming all that much water relative to other things is valid, but comparing it to the water consumed by eating meat weakens your point for anyone that hasn’t bought into the bunk “cows are driving climate change” narrative.
- YetAnotherNick 6mo agoCows drinking isn't the major pie anyways. Cow feeding on Alfalfa which has been grown and irrigated with huge amount of water is the problem.
- taejavu 6mo agoThe vast majority of cattle are raised in a pasture and eat grass, even those labeled “grain fed”. Grain fed just means they spend some time (e.g. 30 days) in a feed lot. I hadn’t heard about cows eating alfalfa though, where is that happening? Wouldn’t it be more valuable to sell it to people instead of using it as feed?
- YetAnotherNick 6mo agoGrass takes lots of water to grow.
- taejavu 6mo agoThat water is free. It’s called rain.
- velvetfoxtrot 6mo agoSort of. It has the opportunity cost of the forests we cleared for it.