this post was submitted on 31 Jul 2025
1426 points (98.2% liked)
Microblog Memes
8757 readers
2152 users here now
A place to share screenshots of Microblog posts, whether from Mastodon, tumblr, ~~Twitter~~ X, KBin, Threads or elsewhere.
Created as an evolution of White People Twitter and other tweet-capture subreddits.
Rules:
- Please put at least one word relevant to the post in the post title.
- Be nice.
- No advertising, brand promotion or guerilla marketing.
- Posters are encouraged to link to the toot or tweet etc in the description of posts.
Related communities:
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Sam Altman (OpenAI) is a Millennial. So is Zuckerberg. LLMs are one of the big energy sinks right now, reaching 1,000 terawatt-hours by 2026 and the current rate of use is doubling every year. For comparison, total global commercial (excluding industrial and transportation, so, office buildings - lights, AC, computers) energy use is 50,000 TWh.
It's still being ignored. Boomers are out of the work force (if not politics), and Gen X is just starting to retire. Between Millenials and Gen Z, they hold 32% of the voting power in the US, the same as Boomers. And Gen Z is only just entering voting age, at 8%.
Half the voting population is under 50 and global temperatures keep increasing. There's every indication sticking your head in the sand is a cross-generational behavior.
Edit: see reply. With correct numbers now I'm mad too. ~~By your own numbers that’s a tiny fraction of the world’s energy use. It seems strange to put such a disproportionate focus on such a small fraction. Where is this rage for the transportation sector?~~
Shh, we don't want to talk about that, car is comfy
On a more serious level, the type of AI that all the energy is being used for, generative AI, is not particularly necessary. Transportation often is. There are types of AI that are ridiculously useful, like the one that does protein folding, or a lot of machine learning algos that classify things for X or Y business reason... But LLMs and image generation are a fucking novelty.
Do you think these are just different technologies that happened to have been developed simultaneously? These are all from the same spark. Neural networks giving rise to emergent unexplainable phenomenon when prodded in very very specific ways. Ai research is almost all trying to understand how the fuck that happens and why it can do all these things.
Radiology is a good use case. Ai porn maybe not so much.
But as god help me. You are saying that cruising in a several ton metal missiles, often alone, back and fourth over the planet or to McDonalds is necessary! No transportation isn't often necessary! Americans often say this but DUH you didn't build sidewalks or trains! You astroturfed the shit out of oil. This is very embarrassing, I get it, but GOD DANN NO transportation with exploding dinosaurs that you frack out of gorgeous boreal forests ISNT NECESSARY AT ALL, we have invented train, bus, bicycle, electric cars, please for god sake stop working for astroturfing oil company proxies and get a fucking bike and then spit and be rude to all single drivers in all cities.
First off, I'm not American.
First of all, fuck off, I've flown back and fourth across the planet exactly once and that was to see my father before he died. I hadn't seen him in 25 years, because he left to the US to pursue something resembling income when I was 2, as our own country was only just getting started economically. Second of all, I said it's "often necessary", not strictly always necessary.
Who's been astroturfing oil company proxies? And anyway, when talking about the CO2 impact, trains, buses and electric cars are part of the number. Bicycles quite a bit less, because the CO2 there is production (once per bike and not comparable to a car or a train) and the extra food you need to eat. But trains, buses and electric cars absolutely do use energy - and therefore increase CO2 emissions, even if indirectly.
But the most important part
I realize that neural networks are the basis of all that, but I'm saying we don't need to be pushing everyone to use a super energy-expensive chatbot instead of a regular search. We don't need AI chatbots embedded into literally every software application we use daily. This doesn't benefit the research, it benefits stock values because AI is a buzzword and you CAN'T run a publicly traded company without saying you're harnessing the power of AI, shareholders will literally murder you.
That's why people are dunking on AI instead of cars. Because 99% of public-facing AI is useless shit people actively dislike and so is 99% of AI energy usage. With cars, I'm willing to bet at least 10% of trips are strictly necessary, and 40-50% of trips are deemed necessary because of stupid car-centric city design with no transit, so still necessary, but for the wrong reasons. I doubt more than about 50% of trips are just leisure altogether. But these are just numbers pulled out of my ass to illustrate a point: There is some car travel that is necessary, some car travel that could be avoided by political change, but is currently necessary for the people doing it. But very little AI usage that is necessary.
Google, Microsoft, etc, aren't building billions upon billions of dollars worth of data centers at a never-seen-before pace to run models that benefit humanity. They're doing it because right now all the money in the world is in building a better "Here are the tallest buildings in NYC to jump off after losing your job" machine than your competitor, and shoving it in more products nobody asked for.
And worst of all, just shoving more and more input data at larger and larger LLMs alone isn't likely to cause new breakthroughs in AI. For all we know, it might be a dead end in the search of AGI - and they're well into diminishing returns as far as investing more and more energy into training new models is concerned.
For sure cars are worse for the planet than AI. But cars DO something. They get you to places. AI tells you how to kill yourself, or how to make pizza with glue, etc. Its best use cases are for cheating at homework, and replacing human workers without even making sure AI CAN do their jobs (good luck hiring all your support staff back, Klarna). It's currently a completely new plague on the planet, and tech CEOs are doing everything to point it out more and more. You know when was the last time I heard anything from Gernot Döllner or Ola Källenius? Fucking never. They don't shove themselves everywhere to let you know what they're doing to destroy the planet. At best they'll tell you what they're doing to reduce their impact. But tech CEOs right now will outright tell you they're going to fire everyone they can, build as many energy-intensive data centers as they can, and drain desert towns of their last drinking water, just so you could see what it would be like if the chick from Avatar had 3 boobs.
THAT is why people are mad at the AI industry.
My numbers were mixed in the previous post; I was mixing total global and total annual use. I'm sorry about that; the numbers looked off but I didn't catch the time scale difference.
AI companies are projected to use 1kTWh in 2025. Transportation is projected to use 1.2TWh, industry, 1.1TWh. Bitcoin, everyone's favorite whipping-boy, is estimated to use only 173TWh globally, a mere 17% of AI. Residential is only 800TWh, 4.5x Bitcoin, but 80% of AI. Commercial is less, at 600TWh.
These all come from different sources: homeinst.org and Deloitte are the main ones, but the Bitcoin stat comes from Cambridge and the EIA (eia.gov), and the AI industry number comes from an MIT and backed by a different Deloitte report.
The industrial sector is the largest energy user, but AI is a close third just below transportation.
I was surprised that cryptocurrency energy use was so relatively small, given the hysteria. Bitcoin alone is 173TWh, far smaller than all of the sectors, and a fraction of AI; but even adding all of the other cryptocurrencies, the estimated consumption rises only to 215TWh. That pushes it past the smallest user, the agriculture sector sitting at 200TWh, but still well below everything else.
AI is the third largest energy consumer, annually, globally.
Transportation is moving in the right direction atm, even if it is slow. AI is going the wrong direction.
I think they're intent wasn't to take away from any other issues we have but to say that we have another burgeoning issue which if it continued to grow at scale could be as damaging as our other major contributors.
Now we have the the education and first-hand experience to understand the impact and scope of the issue and despite this have still showed no reluctance.