If you are old enough you remember posting to Usenet and the warning that would accompany each new submission:
This program posts news to thousands of machines throughout the entire civilized world. Your message will cost the net hundreds if not thousands of dollars to send everywhere. Please be sure you know what you are doing. Are you absolutely sure that you want to do this? [ny]
Maybe we meed something similar in LLM clients. Could be phrased in terms of how many pounds of atmospheric carbon the request will produce.
They're incredibly unpopular because even when they're made revenue-neutral (meaning, everyone gets a refund check) people don't realize most of them would make money if they actually reduced their carbon.
They're incredibly unpopular because the ultraweathly use massive amounts of fossil fuels and thus lobby very, very hard against them...and make sure the public is often told just how evil they are and how expensive they'd hurt Johnny Everyday Worker, even car ownership, especially in a city (where much of the US populative lives) is not affordable to a large segment of the population.
If memory serves Jet A is not taxed at all federally in the case of for-profit corporations (while non-commercial users DO pay a tax!) and many states also either do not tax it or tax it very litte.
It's completely insane that we do not tax fuel usage for probably the most energy-intensive way to move people and/or goods and often that movement of people is entirely frivelous.
Assuming the tax is high enough (or grows over time to become high enough) to offset the negative externalities, and that the money raised is used to offset negative externalities, they're better phrase not as "it's fine to burn up the world as long as you're rich", but rather as "it's fine to emit CO2 as long as you sufficiently offset the damage". Accounting for the damage could involve investments into green technologies, or paying ordinary people to make the tax popular, among other things.
Personally I like the idea of setting the price for emitting 1 ton of CO2 equivalent emissions to the realistic cost of capturing 1 ton of CO2. At least, that seems like a reasonable end goal for a carbon tax, since that could fully account for the negative externality. This would of course be obscenely expensive, which would be a strong incentive to lower carbon emissions where possible, and for people to consume less of products that require large emissions to make or use.
The carbon tax would also have to apply to imported goods to be effective, and tracking how much tax should apply to imports would be even more difficult than doing so for domestic sources of pollution.
I don’t think they are regressive, if you make it revenue neutral, because carbon footprint is heavily correlated with spending. Everything you buy has a good amount of embodied carbon. Revenue neutral is actually redistributive.
Believe me, the common man doesn't need the ultrawealthy to dislike arbitrary cost increases. Carbon taxes are incredibly unpopular. As is common sense and/or planning for a future 20 years ahead. Humans are, on average, selfish beings. All this climate change activism is NOT the norm, and does NOT resonate with most common people. Talk to some outside of your activist bubble, and you will learn a thing or two about humans.
Please see my comment again. Under a revenue-neutral carbon tax everyone gets money back. But they don't realize it. Costs only go up for people who emit more carbon than average.
That's quite condescending btw. Is it "activism" to try to avert a calamity that will increase the cost of living by a lot more 20 years from now? I think it's good fiscal sense. Long-term thinking and planning. Y'know adult shit.
> Humans are, on average, selfish beings
And easily swayed by stupid arguments. Exhibit B: Canada's recent repudiation of the carbon tax because fossil fuel industry propaganda convinced everyone that the tax was the cause of price increases. Now prices will stay the same (because the market will bear them) but no one will get any rebate money.
A revenue-neutral carbon tax redistributes the money collected from the tax equally. Poor people get back much more, as a fraction of their wages, than rich people.
> They're incredibly unpopular because the ultraweathly use massive amounts of fossil fuels and thus lobby very, very hard against them...and make sure the public is often told just how evil they are and how expensive they'd hurt Johnny Everyday Worker, even car ownership, especially in a city (where much of the US populative lives) is not affordable to a large segment of the population.
Eh. It's not Bill Gates and Alice Walton. Sometimes the obvious answer is the real one: It's the fossil fuel industry.
> It's completely insane that we do not tax fuel usage for probably the most energy-intensive way to move people and/or goods and often that movement of people is entirely frivelous.
That one's just the arbitrage problem. Planes move around. If there is an international flight to a country that doesn't tax jet fuel (or taxes it less) then the plane is going to fly into LAX with enough fuel still in the tank to get back to the other jurisdiction and fill up again. Which actually increases fuel consumption because fuel is heavy and they otherwise wouldn't want to do that.
This is the same reason the EU doesn't tax jet fuel.
> the plane is going to fly into LAX with enough fuel still in the tank to get back
Any reason that can't be treated as a fuel import and taxed accordingly? I understand current laws may not allow it but is that legislation impossible to write?
The irony is that carbon taxes don't really affect anyone that much.
Even flying would only cost about 10% more for example. And most other activities have carbon free alternatives they can shift to rather than just eat the cost. Which is kind of the point.
They've been implemented all over the world, because they're effective. They cover 31% of emissions in developed nations.
To whatever degree you could say they are unpopular, they're unpopular in regions where the government doing stuff about climate change (or just "the government doing stuff") is unpopular, which makes it odd to single out putting a price on carbon specifically
You mean, they expose the true cost of those things and make the user pay them. They’re already expensive, the cost is just diffused. That’s the whole problem.
Are solar panels convenient? All polysilicon today is made with fossil fuels, and the R&D to make it with renewable energy is still in-progress. Not to mention that we ship them across the ocean with fossil fuel.
The step before to reduce the silicon dioxide requires carbon, which is where the fossil fuels come in.
Same thing with steel – both are critical input materials and can be made without fossil fuels, but they aren’t today. Maybe a carbon tax would fix that!
The dangers of that web site are much more subtle and hard to defend against than those associated with "a motor vehicle", by which I guess you mean something like a car.
You can see traffic. It's easy to understand the dangers in a collision because when you drive into something unexpectedly your body takes a hit and you get frightened since you immediately realise that it might cost you a lot of money but you don't know for sure.
Being subtly manipulated by a disgustingly subservient fake conversationalist is another thing altogether.
Not sure about understanding, but anyone can use a LLM. That is the most intuitive way to interact with a computer and that's the entire point. It may even work on animals. There is serious research on how LLMs could interpret animal language, like with dolphins.
"Rolling coal" is the practice of modifying a diesel engine—usually in pickup trucks—to increase the amount of fuel entering the engine, which causes the vehicle to emit large, thick plumes of black smoke from the exhaust. This is often done by tampering with or removing emissions control devices.
You don’t need it for any pragmatic benefit because it won’t work. It doesn’t work for eating meat. It won’t work for AI.
The only purpose is to scapegoat the possible environmental or economic fallout. Might as well put it on individuals. Like what’s always done.
I’ve already seen it on the national broadcast. There some supposed experts were wagging their fingers about using AI for “fun”. Making silly images.
Meanwhile we’re gonna put AI to good use in arms races: more spam (automated applications, ads, ads, ads, abuse of services) and anti-spam. There’s gonna be so much economic activity. Disruptive.
AI will probably increase GDP. In the same way shooting the windows out of everyone's houses would increase GDP. Then we can claim is grew the economy.
Your quote is actually telling the opposite of your suggestion.
So they used to send this message, but then it stopped I assume. Costs lowered a lot or the benefits outweighed all associated costs. Same can happen here.
> Costs lowered a lot or the benefits outweighed all associated costs.
How is this even quantifiable?
How about this. Before using AI to make fake images and help kids cheat on their homework, we take it offline and use it to solve it's own problem of energy use.
You know what this does not happen? Because the goal is profit and the profit comes not from solving real important problem, but by making people think it is helping them solve made up problems.
I find that message very curious because the message itself clearly does not cost much but the machines it is send on do. So the more messages that are send the less the messages will cost.
But "asking ChatGPT" is something people do so casually without there being any apparent clues of the costs associated with that. I guess that is true of pretty much everything online though.
Even driving your car around you at least are somewhat aware of the gas you are burning.
We should make extraction expensive enough to capture the externalities.
The problem with fossil fuels isn’t that they pollute, but that most of the negative impact of that pollution is borne by others. This results in an artificially low price which distorts the market and results in economically inefficient overuse.
Capture that cost by making producers pay a tax of the corresponding amount, and market forces will use the “right” amount of fossil fuels in ways that are a net benefit.
I agree in principle. How do we set the price on emissions? Is there a market based approach? Maybe some form of credits that directly fund carbon capture?
I don’t think there’s a market based approach. Ultimately you need smart people to sit down and study the problem and come up with an approximate dollars per ton in costs.
I think this is the right idea but I also think measuring the harm is a really hard problem. Polluters should pay for cleanup. But how much does that cost really? How do we decide? Sure we can put smart people on the problem but what solution will they choose? I don’t mean this dismissively. It’s one of the biggest questions of our time. This is what we need to solve.
Batteries and EVs are the production side. Reducing demand is e.g. requiring you to drive fewer miles. You get a car that doesn't run on petroleum and CO2 goes down while vehicle miles can stay the same or go up.
I would bet most ppl drive around with very little awareness of how much it’s costing, either in money or environmental impact. Many people I’ve met seem to measure efficiency by how much it costs to fill up the tank.
Even more fundamentally, think about how much carbon running a cup of water from your faucet produces. No matter where you live, this is more carbon than an LLM prompt generates.
Or, even worse God forbid, think about how much carbon is produced to create a single bottle or carton of water. Then consider how casually people down bottles of water.
Can you source that? A lot of places have gravity fed reservoirs that are energy positive/neutral (LA, San Francisco and New York all rely on gravity fed reservoirs that were built before energy intensive pumping was practical). There are some costs but they are pretty small per gallon.
I’m assuming that calculation is amortizing the cost of running the water system (including wastewater treatment) adding the cost pumping it to the point that gravity can push it through the pipes. It’s never free.
There seems to be a vast gap between "free" and "more expensive than running LLMs". Also, the water seems...more necessary. Going without LLMs for three days will not threaten your life.
I'm pretty sure this is false. Here's my reasoning.
A google search claims water use uses 12.7% of US energy.
Another search gave 11.7% US energy goes to powering AI (projected to increase to roughly 25% by 2030).
Taking into account hydropower power provides 6.2% of US energy, I feel comfortable saying your statement isn't true.
To further strengthen my statement, I would like to point out another statistic. NPR gives us an estimate of 300K gallons of water use/day to cool the average data center. That pretty much guarantees an LLM query produces more carbon than my filling a cup from a gravity fed water system filled by rain.
That could be solved by charging more for the service. That is the only reason you are aware of the gas burning after all, you aren't conducting your own aq test you are noticing you are filling up twice a week at $50 a tank.
They're aware of the price they pay for the gas, not the emissions. I would wager that the mass ignorance of the impact of fossil fuels (and rubber on roads) that the broader population has is a significant reason why we're in this climate mess today.
This is normal. Once you solve the biggest problem, something else becomes the new biggest problem.
The biggest problem with tailpipe emissions used to be horrendous smog. That was mostly solved in many places, and now the biggest problem is the impact on the global climate.
The biggest issue with childhood mortality used to be disease. Now we (correctly) focus more on accidental deaths.
EVs solved tailpipe emissions, but they’re not perfect. Their biggest problem is just something else.
I am 100% on the side of reducing pollutants — but this was never publicly seen as a major issue and I'm suspicious about the timing.
The oil industry is a conglomerate of degenerates spamming boomer logic all the way down to the workers. Their memes propagate throughout society and lead to the other boomer characteristic of rewriting personal and societal history.
The finger waggers now are being programmed to pretend they talked about tire particulates and the carheads are being programmed to pretend they never cared about 0-60. This another "We have always been at war with Eastasia", just like they all opposed the Iraq war from day 1 and didn't cancel the Dixie Chicks, et cetra.
This may have been discussed in specialist literature somewhere but even when I did ecology courses in university circa 2001ish, I never heard about tire particulates, while I did hear a lot about greenhouse gasses.
It's a concern but not a civilization ending concern like climate change. I low key resent these attempts to move goalposts to satisfy the writer's urge for negativity.
Consider that a bus has six to ten tires that each weigh around ten times more than a typical car tire. This is presented as the alternative to cars, is it even any different? Not implausible that it could actually be worse, especially if the bus isn't at full occupancy at all times.
Meanwhile the weight difference between EVs and petroleum cars is emphasized in the complaints, even though it isn't very large, while the much larger weight difference between any cars and buses is ignored. Because the point isn't to complain about tires, it's to complain about EVs.
And if the point actually was to complain about tires then you still wouldn't be talking about EVs, you would be talking about tires and how to make them shed less or construct them out of lower toxicity materials etc.
The city bus comparison is uneven, but if we consider peak travel times during the week, the density intuitively seems like it works out to less waste. City buses have their numbers and schedule dialed back when you're not in peak hours, and I suspect that it's peak hours where you see the bulk of waste from tires.
My city buses in peak travel hours have anywhere from 20 to 75 people on them. Even if we assume that every one of those folks would have carpooled (which rarely happens), we're looking at a lot of cars, and thus tires, on the road.
> The city bus comparison is uneven, but if we consider peak travel times during the week, the density intuitively seems like it works out to less waste. City buses have their numbers and schedule dialed back when you're not in peak hours, and I suspect that it's peak hours where you see the bulk of waste from tires.
This is really the problem with buses outside of extremely high density areas. (And extremely high density areas should have subways.)
You get off work at 5PM, you want to go to an entertainment venue and then go home at 10PM. You can find a full bus a 5:15PM that will take you there because it's rush hour, but then you can't get home on the bus because there is no bus service after 9PM. Which means you can't take the bus there during rush hour either, because you need your car to be there so you can get home.
Or, you can run mostly-empty buses in the darkness hours, but there goes your efficiency.
Last time I did the math, a Tesla Model Y only had 3x less tire emissions than a semi truck per distance traveled. City buses are on-par with a Tesla Model Y if you only care about mL/km tire wear.
The city bus uses tires with a harder rubber and dimensions such that the pressure at the road is less, plus its normal driving patterns have less wear than typical Tesla use.
To make those sorts of calculations easy, you can ignore all the pressure/usage/etc nonsense and just do basic math on tire dimensions (including min/max tread depth and width, not just radius, though I typically ignore siping and whatnot) and typical longevity. Volume lost per mile driven is basic high-school arithmetic, and the only real questions are regarding data quality and whether the self-imposed constraints (e.g., examining real-world wear rather than wear given optimal driving or something) are reasonable.
> The city bus uses tires with a harder rubber and dimensions such that the pressure at the road is less
Harder rubber seems like it could make a difference, but then you could also put tires with harder rubber on a car.
You can get a heavier vehicle to have the same pressure at the road by using more and bigger tires, but then the problem is that the tires are bigger and there are more of them.
> plus its normal driving patterns have less wear than typical Tesla use.
Isn't a city bus constantly starting and stopping, both as a result of city traffic and picking up and dropping off passengers?
> To make those sorts of calculations easy, you can ignore all the pressure/usage/etc nonsense and just do basic math on tire dimensions (including min/max tread depth and width, not just radius, though I typically ignore siping and whatnot) and typical longevity.
I tried plugging these in and it still comes out as a 6-wheel commercial bus has several times the tire wear as a 4-wheel light truck, rather than being the same.
And I expected the difference to be even more, but I guess that goes to show how much the weight argument is motivated reasoning if ~7x the weight is only ~3x the tire wear and then people are complaining about something which is only ~1.2x the weight.
>I tried plugging these in and it still comes out as a 6-wheel commercial bus has several times the tire wear as a 4-wheel light truck, rather than being the same.
Pardon me if I ask the obvious question, but did you divide your result by the average number of people moved? Because that's the actual utility of mass vs. individual transport. I would find it rather surprising if tire wear was the one measure were buses didn't win out.
A typical city bus has something like 2500 cubic inches of tread that it burns through, compared to 650 for a Model Y. Tires typically last 500k miles, vs 50k, generously, for a Model Y. I'd said "comparable," but that was just to avoid argument. From a tire wear perspective, you're better driving a bus even if you're the only person on it.
I saw this one and figured out where it came from. Google's AI thing says bus tires can last up to 500,000 miles. You follow the link and it says that buses can last up to 500,000 miles, with no implication that they do so on a single set of tires.
Ehh you can't really just put harder tires on a car and leave it at that. Harder tires means less grip, and that is a serious setback and much less safe in a car than the typical bus that runs city routes at lower speeds and less adverse road conditions.
Tire temperature also will play a big roll in tire wear, and I wouldn't expect bus tires to get very hot only rolling half the time and at a lower speed than the typical car.
And of course you also gotta factor in passenger count. Buses generally have more than just 1 or 2 people, while the vast majority of cars will have 1 or 2 people most of the time. And even if a bus tires were to wear out twice as fast as a car's tire, that is still less wear per person than a car.
That's true, but it is all relative. 70k+ mile tires for cars and suvs are fairly common. They sacrifice some ride quality and performance, but not so much as to be unsafe.
Well many of my fellow Americans would only accept an EV if it's gigantic, and even though I can't leave the house without seeing a Prius or a RAV4 hybrid, the news acts like it's gas versus electric as if Toyota hadn't solved this twenty years ago
There’s been decades of lies about climate change. And once the truth got out society was already massively dependent on it. For cars specifically it was a deliberate policy to make e.g. the US car-dependent. And once the truth got undeniable the cope was switched to people’s “carbon footprint” (British Petroleum). In fact there are rumors that the cope echoes to this day.
Zoom out enough and it becomes obviously unproductive to make “mass ignorance” the locus of attention.
People spend hours a day mindlessly scrolling social media apps(streaming video calls to boot) that also take up water and energy usage per hour but are totally disconnected from it
The real thing here is that these tools are currently in the "subsidy" phase, so the pricing doesn't reflect actual costs. Only once they've been made indispensible and impossible to remove will the prices be jacked up and the product enshittified.
Obviously people have zero awareness of or interest in their true impact on the environment. This extends to every facet of life and is not at all limited to AI use.
Do you really think the average person could within 2 orders of magnitude when estimating their carbon footprint for a year?
Yeah we do, it's basic epistemic hygiene. If you don't freak out about running your shower or microwave for a couple seconds or driving a few hundred feet you shouldn't be concerned about prompting an AI.
Except we do care about those things. We used to get tons of PSAs for carbon footprint. Turn off the lights when you leave a room, turn off your computer overnight, turn off the faucet when you're washing your hands. That type of thing.
> Except we do care about those things. We used to get tons of PSAs for carbon footprint. Turn off the lights when you leave a room, turn off your computer overnight, turn off the faucet when you're washing your hands. That type of thing.
“We care”. Because British Petroleum told us to care.[1] Now the new scapegoat grift is how many prompts you use to make your “creative” wedding invitations. Nevermind industrial use though. We’ll just hammer every little API or just endpoint, slurp up the data, then do the same thing tomorrow because why cache? Keep it simple. It’s not our cost.[2] No one will scold us (that can get to us).
Public service announcement huh. Let’s just fall for it again.
Apologizing for AI boiling the oceans sounds like a lot of whataboutism.
I can picture an Elizabeth Holmesian cartoon clutching her diamond necklace.
"Oh, won't somebody think of the tech billionaires?!"
If you don't freak out about running your shower or microwave for a couple seconds or driving a few hundred feet
The basic premise of the modern tech industry is scale. It's not one person running a microwave for a couple of seconds, it's a few billion people running a microwave for the equivalent of decades.
In a way the whole AI hand wringing is already the ultimate whataboutism. We’ve had irrefutable evidence of human caused climate change for over a decade at least, with dire consequences. And? Very little action. In fact, we’re now going backwards. The right way to look at this is comprehensively, not looking only at one issue.
So yes let’s hand wring over AI and continue to do nothing about everything else. And we’ll probably do nothing about AI either, but the endless articles will no doubt keep people distracted.
Its just dumb to only care about energy when it comes to this very specific use case, while freely buying cheap plastic bullshit that was literally shipped from the other side of the planet.
fairness to one polluter over another isn't the real issue - look at prop 65 in california; or if you're not used to this in CA, think of any time you've been on-call. alert fatigue is real and diminishes the urgency of the underlying message.
Individual LLM requests are vanishingly small in terms of environmental impact; inference providers use a lot of batching to do lots of work at once. Furthermore, LLMs and diffusion models are not the only ML workload. While generative AI tickles investors, most of the ML actually being deployed is more mundane things, like recommendation systems, classifiers, and the like; much of which is used for adtech purposes adversarial to that of users. If LLMs and diffusers were the only thing companies used ML for, but efficiency gains from new hardware remained constant, we'd still be at the 2017 baseline for environmental impact of data centers.
Likewise, I doubt that USENET warning was ever true beyond the first few years of the networks' lifetime. Certainly if everything was connected via dial-up, yes, a single message could incur hundreds of dollars of cost when you added the few seconds of line time it took to send up across the whole world. But that's accounting for a lot of Ma Bell markup. Most connections between sites and ISPs on USENET were done through private lines that ran at far faster speeds than what you could shove down copper phone wiring back then.
> Individual LLM requests are vanishingly small in terms of environmental impact;
The article uses open source models to infer cost, because those are the only models you can measure since the organizations that manage them don't share that info. Here's what the article says:
> The largest of our text-generation cohort, Llama 3.1 405B, [...] needed 3,353 joules, or an estimated 6,706 joules total, for each response. That’s enough to carry a person about 400 feet on an e-bike or run the microwave for eight seconds.
I just looked at the last chat conversation I had with an LLM. I got nine responses, about the equivalent of melting the cheese on my burrito if I'm in a rush (ignoring that I'd be turning the microwave on and off over the course of a few hours, making an awful burrito).
How many burritos is that if you multiply it by the number of people who have a similar chat with an LLM every day?
Now that I'm hungry, I just want to agree that LLMs and other client-facing models aren't the only ML workload and aren't even the most relevant ones. As you say adtech has been using classifiers, vector engines, etc. since (anecdotally) as early as 2007. Investing algorithms are another huge one.
Regarding your USENET point, yeah. I remember in 2000 some famous Linux guy freaking out that members of Linuxcare's sales team had a 5 line signature in their emails instead of the RFC-recommended 3 lines because it was wasting the internet or something. It's hard for me to imagine what things were like back then.
If what you're saying is true, why are we hearing about AI companies wanting to build nuclear power plants to power new data centers they think they need to build?
Are you saying all of that new capacity is needed to power non-LLM stuff like classifiers, adtech, etc? That seems unlikely.
Had you said that inference costs are tiny compared to the upfront cost of training the base model, I might have believed it. But even that isn't accurate -- there's a big upfront energy cost to train a model, but once it becomes popular like GPT-4, the inference energy cost over time is dramatically higher than the upfront training cost.
You mentioned batch computing as well, but how does that fit into the picture? I don't see how batching would reduce energy use. Does "doing lots of work at once" somehow reduce the total work / total energy expended?
> If what you're saying is true, why are we hearing about AI companies wanting to build nuclear power plants to power new data centers they think they need to build?
Well, partly because they (all but X, IIRC) have commitments to shift to carbon-neutral energy.
But also, from the article:
> ChatGPT is now estimated to be the fifth-most visited website in the world
That's ChatGPT today. They're looking ahead to 100x-ing (or 1,000,000x-ing) the usage as AI replaces more and more existing work.
I can run Llama 3 on my laptop, and we can measure the energy usage of my laptop--it maxes out at around 0.1 toasters. o3 is presumably a bit more energy intensive, but the reason it's using a lot of power is the >100MM daily users, not that a single user uses a lot of energy for a simple chat.
> not that a single user uses a lot of energy for a simple chat.
This seems like a classic tragedy of the commons, no? An individual has a minor impact, but the rationale switching to LLM tools by the collective will likely have a massive impact.
>If what you're saying is true, why are we hearing about AI companies wanting to build nuclear power plants to power new data centers they think they need to build?
Something to temper this, lots of these AI datacenter projects are being cancelled or put on hiatus because the demand isnt there.
But if someone wants to build a nuke reactor to power their datacenter, awesome. No downsides? We are concerned about energy consumption only because of its impact on the earth in terms of carbon footprint. If its nuclear, the problem has already been solved.
AI seems like it is speedrunning all the phases of the hype cycle.
"TD Cowen analysts Michael Elias, Cooper Belanger, and Gregory Williams wrote in the latest research note: “We continue to believe the lease cancellations and deferrals of capacity points to data center oversupply relative to its current demand forecast.”"
Because training costs are sky-high, and handling an individual request still uses a decent amount of energy even if it isn't as horrifying as training. Plus the amount of requests, and content in them, is going up with stuff like vibe coding.
This program posts news to thousands of machines throughout the entire civilized world. Your message will cost the net hundreds if not thousands of dollars to send everywhere. Please be sure you know what you are doing. Are you absolutely sure that you want to do this? [ny]
Maybe we meed something similar in LLM clients. Could be phrased in terms of how many pounds of atmospheric carbon the request will produce.