• 0 Posts
  • 48 Comments
Joined 4 months ago
cake
Cake day: June 5th, 2026

help-circle

  • First I want to say thanks for answering to a few-days old comment, I was wondering whether anyone uses the fediverse more like a forum. That’s kind of nice, knowing that the comments are not just shouted into the void and forgotten :) Second I want to say sorry for my answer - you stirred up a hornets’ nest :D

    Even when ignoring ANN, I could list way too many arguments why data centers suck, and already sucked before “AI”. But before I get into that, let’s start from the top - buckle up.

    Water use is not really a problem of data centers - they can’t require more cooling than thermal power plants generating their electricity, and I have yet to hear someone complain about the water consumption of thermal power plants. (That’s the TL;DR of a way too long comment I put way too much time into, if you want more info and sources)

    And as for your proposal of requiring data centers to produce 110% of their own energy with renewables, I restate my earlier sentiment:

    Even if the data centers would cover their own energy usage with green energy, that’s energy that could go towards decarbonizing all the other stuff we “need” to do.

    As long as our energy requirements are not completely met by renewable energies it won’t help if you require data centers to produce 200% of their energy. It’s not like we have spare solar panels or wind turbines lying around that would not be put to use if there were no data centers. You basically already have that idea with

    Or they could just turn off the computers and sell the electricity.

    so yeah… building solar parks or wind turbines just without the data centers.

    when the bubble collapses, we get a bunch of cheap, green compute that hobbyists like Myself can use for our little websites.

    That would be nice indeed, but AI data center hardware is not really suited for hosting websites. Unless you can host websites on GPUs or TPUs you will have a lot of “green compute” e-waste. Even the bare-bone servers without the GPUs/TPUs are way too overpowered to host “little websites”. As far as I’m aware, serving websites does not need too much compute power. It’s more about storage and bandwidth, depending on the kind of website you host. For ideas on how to actually sustainably host websites, see e.g. https://solar.lowtechmagazine.com/about/the-solar-website - this is all just to say that “green compute” is a more complex topic. It’s not just about getting your energy from renewables. The 512KB Club also comes to mind. (This is a first step to respond to your “Most of the internet runs on [data centers].” - I don’t even want to know how much bandwidth and energy is wasted on ads and surveillance. That’s also part of the data centers that run the internet.)

    So using those AI servers to host websites would just be wasting energy and resources. Granted, those resources are already wasted, which would be my other criticism here - even if you repurpose the hardware used in the data centers, I would not dare call that “green”. Producing those chips requires a huge amount of energy, so they already have a lot of embodied CO2. Always getting the newest hardware just for the sake of it is not “green”. Keep your computers alive for as long as is reasonable!

    Now to finally get to

    Data centers are actually pretty cool when they’re not being used for ANN tech. Most of the internet runs on them.

    As I hinted at earlier, I’d say it strongly depends on what the data centers are used for. “Internet” is a broad term, and there are many data centers that aren’t used to host websites. The Oil and Gas industry used to (and actually still does) operate some of the biggest supercomputers just to find new oil fields. I wouldn’t call that useful. For some examples on how the tech industry sucks up to oil and gas companies, see one of my other comments. It mostly also works the other way around - search for the name of your favourite “multinational energy company” and a combination of “data center” or “supercomputer” and you’ll most likely find something.

    And it doesn’t stop there, that’s just the first example on my mind. There is a good reason some CPUs or GPUs have been classified as dual-use by the USA. Data centers are used for all kinds of atrocities. Even the ones that are used to run the internet, I’m wondering which of those are a benefit to society.

    Sorry this got so long, I can’t even blame you if you didn’t read it all. I think I just had to get all of this out somewhere.



  • [The guy working for Saudi Aramco] had declared the relationship on the IPCC’s conflict of interest form. “That didn’t seem to me a very good answer, but he was friendly and we had a very nice breakfast.”

    Oh well, then I’m relieved. They are nice!

    The article is quite good, but the situation is depressing. If the IPCC is not trustworthy, then what is there even left? But that’s apparently not the first time that problem came up. According to the article, the conflict of interest policy only exists since 2011, when similar things happened.

    Here are some highlights from the article

    At least a dozen IPCC lead authors are employed by or linked to Saudi Aramco, Saudi institutions and other oil and gas interests, the investigation found. One IPCC vice-chair, part of the body’s leadership group, also represents the kingdom at global climate negotiations as a government employee.

    De Pryck sees the growing presence of Saudi-linked authors in the context of a broader shift in the kingdom’s climate diplomacy. Its longstanding opposition to cutting fossil fuel burning in UN climate negotiations is increasingly complemented by the promotion of technological climate solutions. “When it comes to the IPCC, the Saudi strategy now focuses on advocating technologies like carbon capture that would limit the impact on its fossil fuel economy,” says De Pryck.

    Hoteit, at King Abdullah University of Science and Technology, said: “There is no conflict of interest. My research and contributions related to the IPCC work were fully funded and supported by my employer, which is acknowledged in the referenced paper.” The multi-author paper also states: “We thank Saudi Aramco for conceiving and conducting this pilot project and for providing substantial financial support.”

    That’s an interesting understanding of a “conflict of interest”.

    And the last sentence hammers it home:

    The IPCC said there was no process currently under way to amend its conflict of interest policy.



  • Awesome! That recipe sounds really nice!

    One thing that has been going through my mind, but I didn’t mention in my earlier reply: another approach would be to look at what you eat in the morning and trying to get similar macronutrients, i.e. do you usually eat more protein, carbohydrates or fats in the morning. Then you could try to find a mix that fits for you. But seems like you have it figured out :)




  • True, that’s what they propose in the article, but it’s not like you get pure wood powder back, it’s a wood-epoxy powder mixture. To which you then add epoxy to make new stuff… at some point you just have a mass of epoxy with traces of wood. That’s downcycling at best.

    Not sure how serious/scientific the author of the article is, as they also propose using the powder as mulch, which sounds like an ecological nightmare when considering that it’s mixed with epoxy. But that might not have been the idea of the engineers who designed the wooden blades.

    I’m not saying this technology is bad, even if the wooden blades get burned in the end at least they are not left as landfill like the ones right now.



  • Not at all! However I’m not sure if I’m the best person for that, my breakfast eating habits aren’t the best, I’d say :D

    I’ve been raised with eating cake for breakfast, as my mother really likes to bake. Or, if there was no cake, bread (the german kind, I guess that might make a difference from a nutritional point of view) with marmalade. This caused me to also pick up baking and experiment with recipes to make vegan versions of them, which was a lot of fun and is now a new hobby.

    Other than that, I usually drink black tea during breakfast, as I don’t drink coffee.

    But most important is that you find your own “rhythm”. Eating habits and needs differ a lot between people. A friend of mine gets really hangry if they don’t eat a small snack every few hours in between meals. I can go for half a day without eating, but then also eat way too much afterwards.


  • Guess who is also arguing under this post? What a surprise…

    Sometimes I’m wondering why I’m even engaging in these discussions at all. By now it’s probably just because I’ve reiterated the statements so often, that I just have the sources easily available, but I always know the outcome before I even post an answer. Such people aren’t really interested in actually discussing veganism.






  • Yes and no. It is possible for someone earning 70k/year to emit below the average (let’s also keep in mind that the goal is not to be below the average, but to be at 0 emissions/impact), but everything you can buy has a CO2/environment-cost related to it.

    As soon as you spend more, you have a higher impact. Of course there are more and less impactful ways of consumption, so you do raise a good point, but we’re not going to consume our way out of this misery.

    A really nice book I shove down the throat of every single one of my friends who’s not running away fast enough is The Day the World Stops Shopping, in case you have way too much time at hand ^^