You’re not gonna save the world by not using ChatGPT, just like you won’t save all those slaves in Zambia by not buying from Apple, and just like you didn’t destroy Twitter by joining Bluesky.
To have real effect requires systemic change, so if you want to actually make a difference you can do things like canvassing, running for local office positions and school boards, educating friends and family about politics, or try killing a few politicians and tech CEOs. You know, basic stuff.
Also I asked Gemini’s Deep Research to research this for me because why not UwU
Executive Summary
Estimates for the energy consumed by ChatGPT during its training and inference phases vary considerably across different studies, reflecting the complexity of the models and the proprietary nature of the data. Training a model like GPT-3 is estimated to require around 1.3 GWh of electricity1, while more advanced models such as GPT-4 may consume significantly more, with estimates ranging from 1.75 GWh to over 62 GWh.2 Models comparable to GPT-4o are estimated to consume between 43.2 GWh and 54 GWh during training.3 These figures represent substantial energy demands, with the training of GPT-4 potentially exceeding the annual electricity consumption of very small nations multiple times over. The energy used during ChatGPT inference, the process of generating responses to user queries, also presents a wide range of estimates, from 0.3 watt-hours to 2.9 watt-hours per query.4 This translates to an estimated annual energy consumption for inference ranging from approximately 0.23 TWh to 1.06 TWh. This level of energy demand can be comparable to the entire annual electricity consumption of smaller countries like Barbados. The lack of official data from OpenAI and the diverse methodologies employed by researchers contribute to the variability in these estimates, highlighting the challenges in precisely quantifying the energy footprint of these advanced AI systems.4
Also I demand that everyone who calls it AI instead of procedural generation gets tazed on the butthole
It probably doesn’t consume as much. Cars however… 👀
Projected to consume 25% of total domestic energy production by the time the build out is done
Don’t cars account for 30% of total energy consumption, and 70% of petroleum? Not defending AI here, the last thing we need is another unchecked, massive, resource hog.
Don’t cars account for 30% of total energy consumption, and 70% of petroleum?
Transportation does. But that includes planes, trains, and boats, not just cars.
Not defending AI here, the last thing we need is another unchecked, massive, resource hog.
I’ve seen the argument that AI will reduce travel demand. But I’ve also seen AI guys lobby aggressively to end Work From Home.
Arm CEO Rene Haas cautions that if AI continues to get more powerful without boosts in power efficiency, datacenters could consume extreme amounts of electricity.
Haas estimates that while US power consumption by AI datacenters sits at a modest four percent, he expects the industry to trend towards 20 to 25 percent usage of the US power grid by 2030, per a report from the Wall Street Journal.
There’s a misconception regarding the “consumption” of water, also a bit of a bias towards AI data centers whereas most used water is actually from energy production (via carbon, fuel or even hydroelectric) which is actually a factor to be considered when calculating the actual water use and consumption.
Regarding energy production and water “consumption” I read some papers and as far as I could understand numbers flactuate wildly. 5-40% of the water that runs through the system ends up being consumed via evaporation (so from potentially drinkable/usable for agriculture water to mostly water that ends up in the sea).
What I’m trying to say is that, yes, we should be very aware of the water that we consume in our big data centers but should also put a great focus on the water used by the energy that fuels the data center itself, much of the discourse ends up being “haha use water for email silly” when it should be a catalyst for a more informed approach to water consumption.
Basically I fear that the ai industry can make use of our ignorance and eappease with some “net zero” bs completely ignoring where most of the water is consumed and how.
And yes there are solutions to avoid using fresh water for energy production: solar/wind, using sea water, using polluted water, more sophisticated systems that actually “consume” as little water as possible. These methods have drawbacks that our governments and industry refuse to face and would rather consume and abuse our resources, I really want people to focus on that.
How is it consuming so much water? It is probably used for cooling the data centers, but why could it not be reused or even used as heating network?
Most power generation usually uses steam as a way of converting heat into work. So that power consumption is still a bigger issue than the data center heat management.
Wasted water as in readily usable water, if they build a plant somewhere it will cut down from the natural sources and create their own thermal energy that the world has to also dispel.
Which leads to the bigger issue, all of this contributes to global warming. Along with forcing rationing of water, like dude just live one year in Cali to get a sense of water fears. It’s a downward spiral, often called the singularity.
We are creating destruction of worlds for the sake of what? AI waifus? Not having to solve your own issues? To not think? We fought for the freedom to not be brainless slaves, and we hand back the reigns to the elite above us. The internet was already a beast untamed and unleashed upon children, please do not support this one.
