#its especially fucked when they use an expensive AI model too. like damn.
Explore tagged Tumblr posts
spearxwind · 7 months ago
Text
do you guys fucking remember when music artists actually made or commissioned people to make song/album art for them instead of resorting to ugly ass AI slop?
i swear, it's always the artists with more money too. bring me the horizon keeps using AI art even though they basically shit money. incredibly popular artists that can very well pay artists using AI images to reduce costs. actually fucking sucks so bad. it just makes your song stand out way less
83 notes · View notes
crazy-pages · 1 year ago
Text
Ehhhhh. These are kind of shite articles. This is really nothing like bitcoin mining.
*cracks my knuckles*
Alright kiddos, we're going to approach this like proper goddamn researchers. I'm the first damn person in line for criticizing AI misuse, but I want to make sure we're actually criticizing a genuine problem with AI and not a smokescreen. We've got limited time and energy in this world to freak out about environmental issues, let's put a little of that time and energy in making sure this is a real problem, yeah?
The first article is extremely misleading, it implies that the growth in Google's water usage is 5.6 billion gallons, when that's the total use of all of Google's data centers. For comparison, that's 0.02% of the US farmland irrigation usage. Gonna be real here, that's a pretty fucking fantastic exchange rate of resources for all the services Google provides (even if I'd rather some of their services be publicly operated and all of them less evil). And if we assume the discrepancy in Microsoft's water usage rate increase from 2021 to 2022 is due to AI (a wildly presumptive assumption), then AI can't be amounting to more than a tenth of a percent of a percent of the water usage our agriculture accounts for.
This isn't to say that we shouldn't be constantly pressuring all corporations to reduce usage of our common resources with all the regulation we can muster. Especially in places with scare groundwater resources and minimal fresh water, we should be heavily pressuring data centers to use more expensive but less wasteful cooling methods. But AI's role in this just isn't panic inducing to me.
And while ChatGPT is a useless piece of garbage and most other large language models are too, that's not actually what most AI is. Most AI is a bunch of very quiet programs for stuff like cancer diagnosis, industrial process optimization, weather modeling, etc. You know, actually useful shit it's actually worth spending water on.
So the first article throws around some really big numbers, but fails to contextualize them with the really big numbers they should be compared to - how much water we use for other stuff. And it looks like AI water use is not actually an existential problem. At worst it's a minor extension of existing data center cooling problems which could use some more regulation anyway.
Meanwhile the second article has a chart from 2019 which predates the AI explosion it claims to be worried about, so one strike there for bad practices.
Tumblr media
But if we take it at face value then this is just, not bad for a major industrial investment? Compared to the water cost of prototyping and developing almost any other industrial tool you could name, this is pretty fantastic actually. As for its other claims-
Automation fuelled by AI may result in greater consumption as well as increasing waste in certain sectors, such as the e-commerce industry, which has normalised the rapid and frequent delivery of goods. 
Citation needed
The rising use of AI in agriculture could result in the overuse of pesticides and fertilisers, contaminating the soil and water, and harming biodiversity. Implementing AI in agricultural practices to increase yields at the expense of maintaining ecosystem health could lead to monocultures and biodiversity loss. 
Citation needed that this is even happening
Using AI for environmental management also raises ethical questions. Decisions made by AI systems could be biased if they were presented with inaccurate or incomplete data. For instance, if an AI system received instructions to value economic growth over environmental protection, it might choose to put short-term financial gain ahead of environmental sustainability.
Citation needed that this is even happening
All told, the second article doesn't make a bad point when taken at its most abstract. Which is that it is good to consider the environmental impacts of new technologies and their implementations.
But it is not a credible source for the amount of environmental damage AI is actually doing. The single actual datapoint related to AI (and not e-waste in general) is too out of date to be relevant to the conversation and not actually all that bad. And nothing else it says is anything other than unsourced speculation.
Always remember: When you see sources, read them. And when you read them, do so critically. This ain't shit compared to the impact of crypto.
i wish more people were focusing on the environmental impact of ai
2K notes · View notes