Not sure what point you wanted to make, but this calculator is quite shocking.
GPT 5.5 pro, with "a long document" and 10 requests a day gives 25% of daily CO2 emissions!
Ten coding sessions a day with Opus is still 4.7%!
This feels enormous. I will definitely stop rolling my eyes when people complain about AI CO/water usage...
What are you on about? May be 1 out of 100,000 users are using 5.5 Pro to make 10 "Long Documents" as defined in that tool EVERY day. What a silly thing to harp on.
Six 100,000 token Claude coding sessions use less energy than a dryer load, and less water than making one egg. If you are truly concerned about energy and water usage, AI is not even in the top 100 things you should be concerned about in your daily life.
The real point is not "one session", it's the fact that people now do that routinely, that CICD are using those to check every commit, and each search engine query now does that too, so it multiplies
This very obtusely ommits the demand for new data centers and related infrastructure that using AI creates, the going "vegan for a year" option assumes less cows being born but somehow the "don't use AI" doesn't assume that the data center wasn't build in the first place.
The discrete number of cows being born is theoretically fine-grained enough to actually respond to 2–3 vegans yielding one fewer cow. It's unlikely on a one-year time scale, but one cow only goes so far.
Even a thousand AI objectors aren't going to limit the demand for a data center, in no small part because these investments are only partially driven by current demand and are significantly driven by expectation of future demand. And they're really not going to lead to smaller data centers either because if you're building a data center in the first place you're going to spec it out for future demand.
Regardless, I think in both cases it's important to be realistic about the actual impact that one person has. If that number is disappointingly small, that serves as signal that your conscientious objection isn't making the industry you're objecting to as uncomfortable as you would like to think. It may still be worth objecting for your own sense of self, or maybe it serves as an invitation to evangelize your position more, but either way there's not much value to measuring things in a way that gives you an illusion of greater impact than you actually have.
As someone who actually gives a shit about the environment and global warming and has been putting this into practice for more than a decade through daily personal sacrifices: no, I downvote it because if you properly look into it, AI is just completely insignificant compared to cars, air travel, clothing, food, needless junk and so on that it's a joke. It's always brought up by people who never cared, but now pretend to do so because they hate LLMs for other reasons. The irony is that some of those are actually _good_ reasons but they're too cowardly to admit them. There's nothing unmanly about admitting you're afraid of AI taking your job, becoming more intelligent, and ending up in a dystopia.
Go run the numbers and compare them vs. what it takes to produce a single hamburger or hoodie. Anyone who actually cares has already done this and drawn this conclusion.
Have you heard of "rebound effect"? Sure you can say, individually, one query is not that much... but then it becomes integrated in search engines, so suddenly when there was no queries at all, now there's 500 billions per day, and it gets included in your CICD at every commit, and soon enough in your OS, etc
"Run the numbers" means "run the numbers for using agentic coding for 2 hours per day on a frontier model" not "run the numbers for a single query". The former is the worst case scenario.
Google Search's "AI", which is what you're hinting at is such a good example. Let's say there's 10 billion Google searches per day. 10 billion completions on what is going to be a very tiny, ultra finetuned model with lots of caching (including outputs).
Check out how many queries an hour of agentic coding results in. And input/completion tokens. Estimate energy usage of Opus vs something like Gemma 4 E2B. Calculate how many developers using Opus for coding 1 hour a day would equate to those 10 billion search query originated LLM calls.
You could not have provided a better example to show that without running the numbers you'll end up with assumptions that oppose reality.
That's an interesting choice as a source. It doesn't mention climate change or human impacts at all and describes El Niño as a naturally occurring event.
> The El Nino is a phenomenon that occurs naturally
It was posted at 5am in New York... not sure that that was a US view, so the fact that the platform is US-owned doesn't seem so relevant, if there's a global audience.
That being said, I do agree it is a legit thought (and moreso, completely on point in the subthread discussing downsides), and that it shouldn't be downvoted.
Here's a handy calculator you can use to estimate how much CO2 and water I wasted with my coding agent session: https://www.andymasley.com/visuals/ai-prompt-footprint/
Not sure what point you wanted to make, but this calculator is quite shocking. GPT 5.5 pro, with "a long document" and 10 requests a day gives 25% of daily CO2 emissions!
Ten coding sessions a day with Opus is still 4.7%!
This feels enormous. I will definitely stop rolling my eyes when people complain about AI CO/water usage...
GPT-5.5 Pro is a notoriously expensive model, it's 6x the price of GPT-5.5. Not something to use as a daily driver!
That ten coding sessions a day with Opus number feels more credible to me.
What are you on about? May be 1 out of 100,000 users are using 5.5 Pro to make 10 "Long Documents" as defined in that tool EVERY day. What a silly thing to harp on.
Six 100,000 token Claude coding sessions use less energy than a dryer load, and less water than making one egg. If you are truly concerned about energy and water usage, AI is not even in the top 100 things you should be concerned about in your daily life.
[dead]
The real point is not "one session", it's the fact that people now do that routinely, that CICD are using those to check every commit, and each search engine query now does that too, so it multiplies
This very obtusely ommits the demand for new data centers and related infrastructure that using AI creates, the going "vegan for a year" option assumes less cows being born but somehow the "don't use AI" doesn't assume that the data center wasn't build in the first place.
The discrete number of cows being born is theoretically fine-grained enough to actually respond to 2–3 vegans yielding one fewer cow. It's unlikely on a one-year time scale, but one cow only goes so far.
Even a thousand AI objectors aren't going to limit the demand for a data center, in no small part because these investments are only partially driven by current demand and are significantly driven by expectation of future demand. And they're really not going to lead to smaller data centers either because if you're building a data center in the first place you're going to spec it out for future demand.
Regardless, I think in both cases it's important to be realistic about the actual impact that one person has. If that number is disappointingly small, that serves as signal that your conscientious objection isn't making the industry you're objecting to as uncomfortable as you would like to think. It may still be worth objecting for your own sense of self, or maybe it serves as an invitation to evangelize your position more, but either way there's not much value to measuring things in a way that gives you an illusion of greater impact than you actually have.
[flagged]
As someone who actually gives a shit about the environment and global warming and has been putting this into practice for more than a decade through daily personal sacrifices: no, I downvote it because if you properly look into it, AI is just completely insignificant compared to cars, air travel, clothing, food, needless junk and so on that it's a joke. It's always brought up by people who never cared, but now pretend to do so because they hate LLMs for other reasons. The irony is that some of those are actually _good_ reasons but they're too cowardly to admit them. There's nothing unmanly about admitting you're afraid of AI taking your job, becoming more intelligent, and ending up in a dystopia.
Go run the numbers and compare them vs. what it takes to produce a single hamburger or hoodie. Anyone who actually cares has already done this and drawn this conclusion.
Have you heard of "rebound effect"? Sure you can say, individually, one query is not that much... but then it becomes integrated in search engines, so suddenly when there was no queries at all, now there's 500 billions per day, and it gets included in your CICD at every commit, and soon enough in your OS, etc
"Run the numbers" means "run the numbers for using agentic coding for 2 hours per day on a frontier model" not "run the numbers for a single query". The former is the worst case scenario.
Google Search's "AI", which is what you're hinting at is such a good example. Let's say there's 10 billion Google searches per day. 10 billion completions on what is going to be a very tiny, ultra finetuned model with lots of caching (including outputs).
Check out how many queries an hour of agentic coding results in. And input/completion tokens. Estimate energy usage of Opus vs something like Gemma 4 E2B. Calculate how many developers using Opus for coding 1 hour a day would equate to those 10 billion search query originated LLM calls.
You could not have provided a better example to show that without running the numbers you'll end up with assumptions that oppose reality.
That's an interesting choice as a source. It doesn't mention climate change or human impacts at all and describes El Niño as a naturally occurring event.
> The El Nino is a phenomenon that occurs naturally
El Niño has been occurring naturally for more than 10,000 years. https://en.wikipedia.org/wiki/El_Ni%C3%B1o%E2%80%93Southern_...
The frequency and magnitude of the event is directly related to the warming up of climate
El Niño is a naturally occurring event
While one can raise environmental concerns about the AI datacenter buildout, I don't think it is fair to say that it "ruins the planet".
I don't think it is a good contribution to the discussion around Simon's LLM use to fix a CSS bug.
It was posted at 5am in New York... not sure that that was a US view, so the fact that the platform is US-owned doesn't seem so relevant, if there's a global audience.
That being said, I do agree it is a legit thought (and moreso, completely on point in the subthread discussing downsides), and that it shouldn't be downvoted.
[flagged]