just to set the record straight:

"For example, having DeepSeek R1 (70 billion parameters) answer 600,000 questions would create CO2 emissions equal to a round-trip flight from London to New York. Meanwhile, Qwen 2.5 (72 billion parameters) can answer more than three times as many questions (about 1.9 million) with similar accuracy rates while generating the same emissions."

i'd say that makes the average usage for personal reasons an non-issue.

That's a good question no one else seems to be asking. When you run local prompts, you're using more battery on your laptop or power from your PSU, but both of those are limited by their relatively small power consumption (when compared to server farms). Therefore, it could be likened to encoding heavy video or exporting complex blender models. Are you using more electricity? Yes. Is it city-eroding amounts more? Not even close.

And won't stop putting this slop in everything.

Sometimes i use ai to analyze spreadsheet data for me and spit out an excel sheet with the answers. It says "thinking" for like 10 mins, and i can just imagine some GPUs on fire in a datacenter somewhere...

Generally, heating and cooling are the main energy consumption for domestic purposes. next up is the car, and then electrical consumption. (from what i remember).

I suppose it depends on where you live. Our house consumes something over 20 000kWh per year as our heating is also electric (and rest of the consumption is pretty neglible compared to heating) and we also have a fireplace which consumes around 15m³ of firewood, depending on how cold winter happens to be. Electric grid here has a ton of renewables and nuclear, so co2 footprint should be on the smaller side compared to global average.

Also, as google and microsoft (among others) shoehorns AI "answers" to everything that adds up, but private use seems to be quite insignificant anyways.

FWIW, a short query to a typical sized LLM takes about 1Wh of energy, there lots of variance on how big the model you are using and how long the input and outputs are but thats the correct order of magnitude. 1Wh is the amount of energy consumed by a 1kW electric kettle in 3.6 seconds or a 2kW hairdryer in 1.8 seconds.

if you assume that energy was produced in a coal power plant (the worst for co2 emissions) then it makes around 0.3g of co2 emissions, which is the equivalent of burning about one droplet of gasoline.

I think most people have decided they like using AI more than they care about climate change

No they are magically free.

/s

Lemmy wants me to believe that AI is the danger for climate change. This is like drinking through paper straw while the pollution comes from mega-factories who dump their waste in the river.

So AI is the problem… huh… and not the thousands of missiles dropped in Ukraine, Gaza, Iran and the Occupied Terrorist Zionazi Land.

Cool I guess, AI Bad!

Also true: When you do 0 AI prompts, they all cause the same amount of emissions, which is 999999999 gigajoules.

Meh, whataboutism.

Shockingly, several things can be bad at the same time. What do you think will happen to places that already tend to struggle with heat (like quite a few places in the middle east), if they get even warmer? What will happen to the people that live there? Will they be able to move to colder places? Or will they simply run into borders? Borders, which are protected by more and more money and weapons (and drones controlled by AI, funny, huh?)?

Humans love to live in places that aren’t suitable for their life.

People in some parts of Europe, Russia, North America live in places where without heating, they would simply cease to exist. Same for places in the middle east, people in Qatar and UAE live in places where without AC, they would cease to exist.

Both ways of living is harmful for environment. Maybe we should reevaluate where we live.

I'm no fan of AI "answers", because if i search for something, i'd like to have access to the source or at least know if i depend on a random social media post as my answer. I'm also pretty sure that - if they are smart, and they (mostly) are - caching of questions and answers will cut down on the amount of total questions asked.

and then there are things like grok, which fuck with air quality because elon couldn't wait until the power grid was usable at his datacenter (or open a datacenter where you have access to the required amount of power) and uses dozens of gas turbines for power (without permits, because when the penalty is a fine, it's just the cost of doing business)

xAI Data Center Emits Plumes of Pollution, New Video Shows

A massive data center at xAI’s controversial site in Memphis, Tennessee is emitting huge plumes of pollution, according to an environmental watchdog group.

DeSmog (www.desmog.com)

Because if there is something that can be done in a responsible way, you can count on elon to do it in the most braindead way possible.

e: direct link to the video: https://www.youtube.com/watch?v=mSWgDOzfKRI

They should live in temperate climate instead, so irresponsible, they need to let go of the past and more forward, to greener places.

You can run a model locally on your phone and it will answer most prompts without breaking a sweet, it's actually way less energy than googling and loading the content from a website that's hosted 24/7 just waiting for you to access the content.

Training a model is expensive, using it isn't.

Beautiful. I propose to start questioning very silly reasoning prompts. This way we get more CO2 and better food production/growth.

Wow tokens exist.

See i fit that in 3 words.

I would like to learn about this a bit more, I keep hearing it in conversations here and there. Do you have links around studies/data on this?

Can you link me to what model you are talking about? I experimented with running some models on my server, but had a rather tough time without a GPU.