>Generating 1,000 images with a powerful AI model, such as Stable Diffusion XL, is responsible for roughly as much carbon dioxide as driving the equivalent of 4.1 miles in an average gasoline-powered car. In contrast, the least carbon-intensive text generation model they examined was responsible for as much CO2 as driving 0.0006 miles in a similar vehicle
But 3 billion pictures per week is equivalent to 12.4 million miles or 639.6 million miles extra per year.*
It's not about you. There are no 3 billion people each creating one picture per week. Its probably more like 10.000 assholes creating 2.5 billion pictures per week to satisfy some stupid online feed and the other users creating a picture per month on average.
It equals around 0.004% of actual driven miles if my estimates are right.
That is with an estimated ~270 billion miles vehicle miles every week (I put commercial in there.. about ~200 billion miles if you only include passenger vehicles.)
That study is from 2023 when the cost to produce 1000 images was 2.91 Wh per image. More recent numbers for Stable Diffusion's 2025 models that puts it at 1.3 Wh per image and that number continues to decrease.
For context that means a days worth of image generation emits about the same amount of carbon dioxide as a single transatlantic flight (New York to London). There are approximately 1500 transatlantic flights per day.
So every day the people of Vermont emit more CO2 just commuting than all image generation through OpenAI per week. Nice. That does put it in perspective. It’s really efficient.
yep, what our climate really needs is another carbon emitter similar to an entire state's worth of vehicles just so people can not pay artists or make dumb images of themselves ripping off some artistic style like Studio Ghibli. that is so much more of a social good than people being able to make it to work and earn a living
Oh I think paying people to make 3 billion images would result in far more carbon emissions. Just think about how much CO2 a human emits while painting. The paint, the food. Just for the sake of the climate I would never pay an artist for something like this.
for sure, because every one of those 3 billion images rendered by human hand and not automated with a tool would exist. we wouldn't have a magnitude smaller number of focused designs, of course, that's definitely not how art and design processes happen
Ours cars are atrocious. We can have the most powerful artificial minds imagine 1000 images from simple prompts, and that takes as much energy as moving a human being 4 miles.
And that will get more energy efficient. Gas cars have barely budged.
This is an embarrassingly idiotic article. They're comparing against the least intensive text model, ie some million parameter model nobody uses. Stable Diffusion is a 3.5 billion parameter model, while the GPT models a billion people are using for text generation are over 10 trillion parameters. To say nothing of the differences in average context size usage. The actual ratio of SDXL to text generation pollution is probably literally reversed from what the article claims by lying with statistics.
(Note, however, that OpenAI's image model is much much larger than SDXL; however, we don't have precise numbers for it. Nonetheless, misinformation is misinformation.)
in the communities where I am, there are small 'hacker' collectives where people are working on combating surveillance, providing alternative non-surveilled software solutions, and helping other movement folks set-up and adopt less-worse platforms (eg moving off Google to CryptPad)
I think its worthwhile for those of us who are involved in this kind of work to be aware of further descent into fascistic tendencies (ie no policing accountability for repressive violence). the stakes are high and always rising and, accordingly, our risk assessments and threat modeling should stay updated to match
and for people who may disagree with the current regime's politics, it's also healthy to stay apprised of the erosion of your own civil liberties, and the potential downstream effects it might have on your life
> Federal prosecutors in Minnesota had been preparing to charge Castro with deprivation of rights under color of law for the shooting of Sosa-Celis, according to several people familiar with the case. [But] late on Tuesday, Matthew Evans, the assistant U.S. attorney in Minnesota in charge of the case, told lawyers for Sosa-Celis and other victims to prepare for Castro to be charged “only with False Statements."
> “This is being directed by the Main Justice and the US Attorney,” Evans wrote in a remarkably candid account of internal deliberations. “I objected in the strongest possible terms and fought it as hard as I could. It wasn’t enough.”
>[...] DHS originally said Castro and the other agent were ambushed by the men and beaten with a shovel or broomstick, making Castro fear for his life and fire in self-defense. The next month, however, they dropped the charges against [the agents], and the DHS backed off its narrative in a rare admission of possible wrongdoing of its officers. “Video evidence has revealed the sworn testimony of the two agents appeared to be false,” then-acting ICE director Todd Lyons said in a statement.
Martha Stewart is probably the textbook example of someone getting screwed on these sorts of charges considering they were questions she answered in good faith and she wasn't even the subject of the investigation.
Never happens if you're working for the government though.
I test drove a new Subaru Forester a little while ago. I currently drive the 2015 version where the most advanced feature is the visual crash prevention system that brakes for you and a bluetooth sync that barely works
and yet I hated the new Subaru so much. I hated the fact that the entire console is now one giant touchscreen, requiring you to look away from the road in order to operate. I hated the huge bulky 'Starlink' hardware hidden in the roof. I hated the fact that the MPG was barely better even with ten years worth of supposed engineering advancement and the Forester was now wider and taller which sucks for being a city car
the last 15 years of 'innovation' in cars was a total mistake. call me a luddite but I really don't want my car to be able to be tracked by any satellite at any time. so it's either used cars for me from now or something like the Slate truck with the barest dashboard config (or a kei truck once it's legal in my state)
The Forester still has the awful infotainment system with the vertical screen, first introduced on the 2020 Outback. I actually like the vertical aspect ratio, and the screen is pretty bright and high resolution, but the underlying software is disastrous. It was slow and buggy and felt outdated when it was introduced 6 years ago!
For what it’s worth, the new 2026 Outback has an excellent new infotainment system and tons of physical controls for car settings. But apparently everyone hates it for the external redesign (I think it looks fine). Hopefully that interior will roll out to the rest of the models too.
They are not going back to physical controls out of the goodness of their hearts. Physical controls are a requirement to get a European 5-star safety rating, and possibly in China as well.
Also everyone started screaming when Tesla was starting to rumor to start using the cabin cameras to recognize people when Subaru has a totally creepy facial recognition system on their current vehicles to do seating presets and the like. It was bad enough they use the cameras to imply I'm distracted or sleepy while I'm driving and have the things throw twenty million chimes and warnings, but actual driver ID is just terrible.
> Seat position profiles is a great feature, making it automatic is completely unnecessary.
Driving a Model 3, I disagree - having seats and basically everything else adjust as soon as you open the door is the biggest boon I can imagine. I totally wish the rearview mirror would adjust as well…
that's what a call a straw-man. NSPM-7 exists and is providing funding and direction for LEOs to target ideologically leftist organizations as tame as BLM or 50501 [0]
this, in combination with the rise of Flock, Clearview, Axon, and other mass-surveillance programs is leading us down an incredibly dark and Orwellian state - something that feels almost ludicrous given how anti-Big-Brother your average person is
They say "every accusation is a confession" yet I'm still continually surprised by accusations that I never thought would turn out to be confessions, turning out to be confessions. Like the Chinese surveillance state.
It seems to be a type of psychological projection - someone harbors an ideal like a surveillance state and assumes everyone else is harboring the same ideal so they accuse them.
I hope the Uyghur genocide accusation doesn't turn out to be a confession.
it's been around for quite a lot longer than that. 1984 was written in 1949 and wasn't strictly inspired by only fascist regimes - there were and still are a lot of similarities (eg ethnosupremaccy, nativism, etc) that have deep roots elsewhere
>With the end of World War II, Project SHAMROCK was established in 1945. The organization was created to accumulate telegraphic data entering and exiting from the United States.[4][22] Major communication companies such as Western Union, RCA Global and ITT World Communications actively aided the project, allowing American intelligence officials to gain access to international message traffic.[23] Under the project, and many subsequent programs, no precedent had been established for judicial authorization, and no warrants were issued for surveillance activities. The project was terminated in 1975.[4]
I was responding to the comment that said "Everyone who disagrees with the GOP is labeled antifa" with two of the many prominent Democrat politicians who haven't been. Should I have listed 50 of them or 100?
agreed. the comments in here feel reactionary, not taking into account the sheer amount of influence marketing has on society. beyond just hyperpalatability, the tobacco companies also leveraged all they learned from their extremely successful advertising campaigns and applied it to foods
that meme image of Garfield saying 'you are not immune to propaganda' comes to mind. it's always the people who are the most confident that they aren't influenced by marketing that also have some of the most irrationally strong brand loyalties
> Buying into the food industry also facilitated the tobacco companies in applying the knowledge, skills, and strategies developed from their tobacco businesses to another business domain in ultraprocessed foods and beverages. For example, Philip Morris applied its aggressive marketing tactics developed from its tobacco businesses to market ultraprocessed foods to the US public. Furthermore, US tobacco companies and their food companies were systematically involved with developing hyper-palatable foods, which have elevated addictive properties.
>Tobacco-owned foods were 29% more likely to be classified as fat and sodium hyper-palatable foods (HPF) and 80% more likely to be classified as carbohydrate and sodium HPF than foods that were not tobacco-owned between 1988 and 2001 (P-values = 0.005–0.009).
> Tobacco giant R.J. Reynolds led the transition to sweetened beverages in 1963 when it purchased Hawaiian Punch from Pacific Hawaiian Products Company, according to the documents. The beverage previously had been promoted to adults as a cocktail mixer, but R.J. Reynolds sought to beef up the drink’s “Punchy” mascot – a counterpart to the “Joe Camel” cartoon character the company used to promote cigarettes – and featured it on toys, schoolbook covers, comics, tumblers, clothing and TV commercials. Punchy became the “best salesman the beverage ever had,” according to tobacco industry documents.
> In the ’60s and ’70s, the company conducted taste tests with children and mothers to evaluate sweetness, colors and flavors for Hawaiian Punch product line extensions. The children’s preferences were prioritized, the authors noted.
> By 1983, R.J. Reynolds introduced the nation’s first juice box, marketed as a “handy little carton that comes with its very own straw.” This innovation was largely responsible for a 34 percent jump in sales, according to industry documents.
> At the end of 1988, Philip Morris Companies purchased Kraft for $12.9 billion. In 1989, Kraft merged with Philip Morris's General Foods unit—makers of Oscar Mayer meats, Maxwell House coffee, Jell-O gelatin, Budget Gourmet frozen dinners, Entenmann's baked goods, Kool-Aid, Crystal Light and Tang powdered beverage mixes, Post Cereals, Shake 'n Bake flavored coatings and numerous other packaged foods—as Kraft General Foods. Its aggressive product development was reversed after the merger, as it became slow in addressing issues on its product lines due to its size, and also company politics.[4]
> In 2000, Philip Morris (renamed Altria in 2003) acquired Nabisco Holdings for $18.9 billion and merged the company with Kraft Foods the same year.[212][4] Also in 2000, the company acquired Balance Bar.[213] In 2001, Philip Morris sold 280 million Kraft shares via the third-largest IPO of all time, retaining an 88.1% stake in the company.
>Livestock production accounts for 14-18 percent of global greenhouse gas emissions, including 32 percent of methane gas emissions worldwide.
14-18% seems like a fairly significant chunk, no? and what if we just started with not eating beef, all of us, especially the USA? that'd cut down the emissions by quite a large bit
not to mention the abandonment by state energy companies like Georgia Power of their solar/renewable goals and full buy-in to building yet more gas turbines in order to attract hyperscale data centers. I was part of some organizing against this in my city: https://atlpresscollective.com/2025/12/18/public-service-com...
just look at the funding for renewable R&D in the past few years - the reason we have the market we do now is that huge spike (thanks, Obama) in 2009 that drove down the cost of manufacturing and drove up conversion efficiency. that much investment into the renewables sector hasn't been the same since, especially now: https://www.theinvadingsea.com/2025/11/05/fossil-fuels-oil-c...
is any of this 'scientific'? does AA allow peer review of its processes? are these published in journals of at least medium impact? what are the sample sizes? how grounded is their mechanistic reasoning?
like they have words that are dressed in scientific language on their site like
"We estimate a 95% confidence interval for Artificial Analysis Intelligence Index of less than ±1% - based on experiments with >10 repeats on certain models for all evaluation datasets included in Artificial Analysis Intelligence Index v4.2."
but where's the outcome dataset justifying this? how did they get that probability? what was the specific methodology of the tests? what variables did they account for?
there's a major difference between scientific sounding and being truly empirically rigorous. the 'research' in AI intelligence feels somehow even less trustworthy than supplement-funded studies because those are at least subjected to scrutiny by peers without profit motives
> but where's the outcome dataset justifying this? how did they get that probability? what was the specific methodology of the tests? what variables did they account for?
why keeping journals as an argument on the tech site that was always less formal with mandatory institutions but more open source. Journals discredited themselves multiple times. tech is expanding boundaries of scientific methods and AI will push it more.
It's not about journals. If they want to be 100 trustworthy they should release end to end reproducible pipelines for the whole process with everything but the private datasets included, and all design decisions fully documented. And let third party labs audit to confirm that the holdout questions are equal difficulty and similar task types to the public ones.
1) open access is a rising trend in journal publishing. arxiv exists because of it but this also means that a lot of papers get 'published' before they are peer-reviewed which is often when significant issues are caught
2) there are definitely problems with journals and publishing but, like many similar absurdly reductive pronouncements, the argument that the current mode of scientific inquiry is bunk is both wrong and lacks nuance
the problem with modern publishing is that private equity is buying up publishers [0]. these publishers are then giving peer-reviewers no time and zero pay to do the necessary work of review [1] while also charging exorbitant rates for access. this is leading to worsening quality of the published research along with highly overburdened researchers who are stuck between shrinking funding [2] and their myriad other professional obligations
to just say that 'journals' are bunk is ignorant in a harmfully anti-empirical way. the process of empirical research and review is the entire reason why we see realworld results. foundations comprised of bullshit crumble fast but for some reason or another our economic system is highly driven to enshittifying everything it touches
to have defenders who claim AA is 'pushing the boundaries of the scientific method' sounds like the screeching refrain of anti-intellectual cargo cults, apeishly mimicking the features of rigor and methodology while avoiding any real accountability
Yes, journals are harmful to freedom of knowladge which might seem ignorant especially to people coming from academia and mainstream who doesn't know better and can't imagine anything else besided wallen gardens they grew up with.
Arxiv is dominated by CS [1] and nothing really changed over the years [2]. Saying about pushing the boundaries of scientific methods it was another rock into journals and the status quo on how it's done like p-hacking and other data manipulations. Being in open with community notes would would have a difference.
Empirical research can exist without journals, science can be for masses and journals are antithetical for sharing making the knowledge exclusive.
>Generating 1,000 images with a powerful AI model, such as Stable Diffusion XL, is responsible for roughly as much carbon dioxide as driving the equivalent of 4.1 miles in an average gasoline-powered car. In contrast, the least carbon-intensive text generation model they examined was responsible for as much CO2 as driving 0.0006 miles in a similar vehicle
reply