Comment by HendrikHensen
24 days ago
All I can think about is how much power this takes, how many un-renewable resources have been consumed to make this happen. Sure, we all need a funny thing here or there in our lives. But is this stuff really worth it?
Luckily we live in a society where its ok to use power for personal pleasure, such as running an A/C in the summer which accounts for much more electricity use than LLM inference.
https://www.eia.gov/tools/faqs/faq.php?id=1174&t=1
[flagged]
> U.S. data centers consumed 183 terawatt-hours (TWh) of electricity in 2024, according to IEA estimates. That works out to more than 4% of the country’s total electricity consumption last year – and is roughly equivalent to the annual electricity demand of the entire nation of Pakistan. By 2030, this figure is projected to grow by 133% to 426 TWh.
https://www.pewresearch.org/short-reads/2025/10/24/what-we-k...
There are ~10M cows nationally. The average energy consumption is ~1000 kWh/cow annually. Summing up, the entire dairy industry consumes ~10TWh. That is less than 10% of the national data center energy burn. [edit: was off by a factor of 10]
1 reply →
> One dairy operation uses more resources than all the datacenters in the united states
citation for this claim?
https://www.pewresearch.org/short-reads/2025/10/24/what-we-k...
> U.S. data centers consumed 183 terawatt-hours (TWh) of electricity in 2024, according to IEA estimates. That works out to more than 4% of the country’s total electricity consumption last year – and is roughly equivalent to the annual electricity demand of the entire nation of Pakistan. By 2030, this figure is projected to grow by 133% to 426 TWh.
lol what? Can you please cite some sources for this claim?
The actual energy usage is probably not a big deal comparatively. But the attention / economic energy is absolutely a big deal and an increasingly farcical one.
I think the market is just waiting for the next Big Think to come around (crypto, VR, etc.) and the attention obsession will move on.
Trivial in the grand scheme of things. There are much larger problems to attend to - if worrying about the cost and impact of AI tokens was a problem, we'd be living in a utopia.
Literally pick any of the top 100 most important problems you could have any impact on, none of them are going to be AI cost/impact related. Some might be "what do we do when jobs are gone" AI related. But this is trivial- you could run the site itself on a raspberry pi.
I think this is a strange, and honestly worrying, stance.
Just because there are worse problems, doesn't mean we shouldn't care about less-worse problems (this is a logical fallacy, I think it's called relative privation).
Further, there is an extremely limited number of problems that I, personally, can have any impact on. That doesn't mean that problems that I don't have any impact on, are not problems, and I couldn't worry about.
My country is being filled up with data centers. Since the rise of LLMs, the pace at which they are being built has increased tremendously. Everywhere I go, there are these huge, ugly, energy and water devouring behemoths of buildings. If we were using technology only (or primarily) for useful things, we would need maybe 1/10th of the data centers, and my immediate living environment would benefit from it.
Finally, the site could perhaps be run on a Raspberry Pi. But the site itself is not the interesting part, it's the LLMs using it.
I don't think it's odd at all- having taken a deep look at the potential impact and problems surrounding AI, including training and datacenters, I've come to the conclusion that they're about as trivial and low ranking a problem as deciding what color seatbelts should be in order to optimize driving safety. There are so many more important things to attend to - by all means, do the calculus yourself, and be honest about consumed resources and environmental impacts, but also include benefits and honest economics, and assess the cost/benefit ratio for yourself. Then look at the potential negatives, and even in a worst case scenario, these aren't problems that overwhelm nearly any other important thing to spend your time worrying about, or even better, attempting to fix.
It’s odd that people seem to be so against the AI slop in particular, because energy and water and whatnot. I’m fairly sure video games eat a lot more power than AI slop and are just as useless. So is traveling - do people truly need to fly 3000 miles just to see some mountains? Why do people demand food they like when you’d survive just fine off of legumes and water?
> Everywhere I go, there are these huge, ugly, energy and water devouring behemoths of buildings.
Everywhere you go? Really?
The water consumption is minor, btw. Electricity is more impactful but you’d achieve infinitely more advocating for renewables rather than preaching at people about how they’re supposed to live in mudhuts.
2 replies →
I'm under the impression LLMs don't generally work that well on an RPI, and I'm guessing that's what the GP is referring to.
Evidence of European or misinformation
You are consuming non-renewable resources by reading this on your device and posting a comment for your entertainment.
At least with Moltbook, it is an interesting study for inter-agent communications. Perhaps an internal Moltbook is what will pave the path towards curing cancer or other bleeding-edge research.
With your comment, you are just wasting non-renewable resources just for your brain to feel good.