#computer generated environment

[smbmatters]

Virtual Reality: The Future of Immersive Experiences

Virtual Reality (VR) is not just a concept from science fiction; it has become a tangible and trans-formative technology shaping our present and future. This immersive technology is revolutionizing various sectors, from entertainment and education to healthcare and business. By creating simulated environments that can be similar to or completely different from the real world, VR offers users unique and engaging experiences.

What is Virtual Reality?

Virtual Reality is a technology that uses computer-generated environments to simulate physical presence in real or imagined worlds. Users interact with these environments using VR headsets and other devices, such as gloves with sensors. This technology creates a sense of immersion, making users feel as though they are truly part of the virtual world.

The Evolution of Virtual Reality

The concept of VR dates back to the mid-20th century, but it wasn't until the 1990s that significant advancements were made. Early VR systems were bulky and expensive, limiting their accessibility. However, with technological progress, VR has become more affordable and sophisticated. Today's VR headsets are more compact, providing higher resolution and more interactive experiences.

VR in Entertainment and Gaming

One of the most popular applications of VR is in entertainment, particularly gaming. VR gaming takes player engagement to a whole new level by allowing them to step into the game world. Players can explore virtual environments, interact with game elements, and experience a more immersive game-play.

Beyond gaming, VR is also transforming the movie and music industries. VR movies and concerts provide audiences with a 360-degree view, making them feel like they are part of the scene or performance.

VR in Education and Training

Virtual Reality is proving to be a powerful tool in education and training. By creating interactive and immersive learning experiences, VR helps students understand complex subjects better. For example, medical students can practice surgeries in a risk-free virtual environment, gaining hands-on experience without the potential consequences of real-life practice.

In addition to education, VR is being used for employee training across various industries. From manufacturing to customer service, VR simulations allow employees to practice skills and procedures in a controlled setting, improving their proficiency and confidence.

VR in Healthcare

The healthcare industry is also benefiting from VR technology. VR is used in medical training, therapy, and rehabilitation. For instance, VR can help patients with anxiety disorders by exposing them to controlled virtual environments, aiding in gradual desensitization. In physical therapy, VR exercises can make rehabilitation more engaging for patients, improving their recovery outcomes.

VR in Business and Marketing

Businesses are leveraging VR to create immersive marketing experiences. Virtual showrooms and product demonstrations allow customers to explore and interact with products in a virtual space, enhancing their shopping experience. Real estate companies use VR to offer virtual property tours, enabling potential buyers to explore homes without physical visits.

The Future of Virtual Reality

The future of Virtual Reality looks promising, with continuous advancements in technology and increasing adoption across various sectors. Upcoming developments include more affordable VR headsets, improved graphics, and enhanced interactivity. As VR technology becomes more accessible, its applications will expand, further integrating VR into our daily lives.

Challenges and Considerations

Despite its potential, Virtual Reality faces challenges that need to be addressed. One major concern is the potential for motion sickness, which some users experience when using VR headsets. Additionally, the development of high-quality VR content requires significant resources and expertise.

Privacy and security are also important considerations. As VR systems collect and process user data, ensuring the privacy and protection of this information is crucial.

Conclusion

Virtual Reality is transforming the way we interact with digital content, offering immersive and engaging experiences across various fields. From gaming and entertainment to education and healthcare, the potential applications of VR are vast and varied. As technology continues to advance, VR will undoubtedly play an increasingly important role in shaping our future. Embracing and addressing the challenges of VR will pave the way for a more immersive and interactive world.

[mostlysignssomeportents]

Autoenshittification

Forget F1: the only car race that matters now is the race to turn your car into a digital extraction machine, a high-speed inkjet printer on wheels, stealing your private data as it picks your pocket. Your car’s digital infrastructure is a costly, dangerous nightmare — but for automakers in pursuit of postcapitalist utopia, it’s a dream they can’t give up on.

Your car is stuffed full of microchips, a fact the world came to appreciate after the pandemic struck and auto production ground to a halt due to chip shortages. Of course, that wasn’t the whole story: when the pandemic started, the automakers panicked and canceled their chip orders, only to immediately regret that decision and place new orders.

But it was too late: semiconductor production had taken a serious body-blow, and when Big Car placed its new chip orders, it went to the back of a long, slow-moving line. It was a catastrophic bungle: microchips are so integral to car production that a car is basically a computer network on wheels that you stick your fragile human body into and pray.

The car manufacturers got so desperate for chips that they started buying up washing machines for the microchips in them, extracting the chips and discarding the washing machines like some absurdo-dystopian cyberpunk walnut-shelling machine:

https://www.autoevolution.com/news/desperate-times-companies-buy-washing-machines-just-to-rip-out-the-chips-187033.html

These digital systems are a huge problem for the car companies. They are the underlying cause of a precipitous decline in car quality. From touch-based digital door-locks to networked sensors and cameras, every digital system in your car is a source of endless repair nightmares, costly recalls and cybersecurity vulnerabilities:

https://www.reuters.com/business/autos-transportation/quality-new-vehicles-us-declining-more-tech-use-study-shows-2023-06-22/

What’s more, drivers hate all the digital bullshit, from the janky touchscreens to the shitty, wildly insecure apps. Digital systems are drivers’ most significant point of dissatisfaction with the automakers’ products:

https://www.theverge.com/23801545/car-infotainment-customer-satisifaction-survey-jd-power

Even the automakers sorta-kinda admit that this is a problem. Back in 2020 when Massachusetts was having a Right-to-Repair ballot initiative, Big Car ran these unfuckingbelievable scare ads that basically said, “Your car spies on you so comprehensively that giving anyone else access to its systems will let murderers stalk you to your home and kill you:

https://pluralistic.net/2020/09/03/rip-david-graeber/#rolling-surveillance-platforms

But even amid all the complaining about cars getting stuck in the Internet of Shit, there’s still not much discussion of why the car-makers are making their products less attractive, less reliable, less safe, and less resilient by stuffing them full of microchips. Are car execs just the latest generation of rubes who’ve been suckered by Silicon Valley bullshit and convinced that apps are a magic path to profitability?

Nope. Car execs are sophisticated businesspeople, and they’re surfing capitalism’s latest — and last — hot trend: dismantling capitalism itself.

Now, leftists have been predicting the death of capitalism since The Communist Manifesto, but even Marx and Engels warned us not to get too frisky: capitalism, they wrote, is endlessly creative, constantly reinventing itself, re-emerging from each crisis in a new form that is perfectly adapted to the post-crisis reality:

https://www.nytimes.com/2022/10/31/books/review/a-spectre-haunting-china-mieville.html

But capitalism has finally run out of gas. In his forthcoming book, Techno Feudalism: What Killed Capitalism, Yanis Varoufakis proposes that capitalism has died — but it wasn’t replaced by socialism. Rather, capitalism has given way to feudalism:

https://www.penguin.co.uk/books/451795/technofeudalism-by-varoufakis-yanis/9781847927279

Under capitalism, capital is the prime mover. The people who own and mobilize capital — the capitalists — organize the economy and take the lion’s share of its returns. But it wasn’t always this way: for hundreds of years, European civilization was dominated by rents, not markets.

A “rent” is income that you get from owning something that other people need to produce value. Think of renting out a house you own: not only do you get paid when someone pays you to live there, you also get the benefit of rising property values, which are the result of the work that all the other homeowners, business owners, and residents do to make the neighborhood more valuable.

The first capitalists hated rent. They wanted to replace the “passive income” that landowners got from taxing their serfs’ harvest with active income from enclosing those lands and grazing sheep in order to get wool to feed to the new textile mills. They wanted active income — and lots of it.

Capitalist philosophers railed against rent. The “free market” of Adam Smith wasn’t a market that was free from regulation — it was a market free from rents. The reason Smith railed against monopolists is because he (correctly) understood that once a monopoly emerged, it would become a chokepoint through which a rentier could cream off the profits he considered the capitalist’s due:

https://locusmag.com/2021/03/cory-doctorow-free-markets/

Today, we live in a rentier’s paradise. People don’t aspire to create value — they aspire to capture it. In Survival of the Richest, Doug Rushkoff calls this “going meta”: don’t provide a service, just figure out a way to interpose yourself between the provider and the customer:

https://pluralistic.net/2022/09/13/collapse-porn/#collapse-porn

Don’t drive a cab, create Uber and extract value from every driver and rider. Better still: don’t found Uber, invest in Uber options and extract value from the people who invest in Uber. Even better, invest in derivatives of Uber options and extract value from people extracting value from people investing in Uber, who extract value from drivers and riders. Go meta.

This is your brain on the four-hour-work-week, passive income mind-virus. In Techno Feudalism, Varoufakis deftly describes how the new “Cloud Capital” has created a new generation of rentiers, and how they have become the richest, most powerful people in human history.

Shopping at Amazon is like visiting a bustling city center full of stores — but each of those stores’ owners has to pay the majority of every sale to a feudal landlord, Emperor Jeff Bezos, who also decides which goods they can sell and where they must appear on the shelves. Amazon is full of capitalists, but it is not a capitalist enterprise. It’s a feudal one:

https://pluralistic.net/2022/11/28/enshittification/#relentless-payola

This is the reason that automakers are willing to enshittify their products so comprehensively: they were one of the first industries to decouple rents from profits. Recall that the reason that Big Car needed billions in bailouts in 2008 is that they’d reinvented themselves as loan-sharks who incidentally made cars, lending money to car-buyers and then “securitizing” the loans so they could be traded in the capital markets.

Even though this strategy brought the car companies to the brink of ruin, it paid off in the long run. The car makers got billions in public money, paid their execs massive bonuses, gave billions to shareholders in buybacks and dividends, smashed their unions, fucked their pensioned workers, and shipped jobs anywhere they could pollute and murder their workforce with impunity.

Car companies are on the forefront of postcapitalism, and they understand that digital is the key to rent-extraction. Remember when BMW announced that it was going to rent you the seatwarmer in your own fucking car?

https://pluralistic.net/2020/07/02/big-river/#beemers

Not to be outdone, Mercedes announced that they were going to rent you your car’s accelerator pedal, charging an extra $1200/year to unlock a fully functional acceleration curve:

https://www.theverge.com/2022/11/23/23474969/mercedes-car-subscription-faster-acceleration-feature-price

This is the urinary tract infection business model: without digitization, all your car’s value flowed in a healthy stream. But once the car-makers add semiconductors, each one of those features comes out in a painful, burning dribble, with every button on that fakakta touchscreen wired directly into your credit-card.

But it’s just for starters. Computers are malleable. The only computer we know how to make is the Turing Complete Von Neumann Machine, which can run every program we know how to write. Once they add networked computers to your car, the Car Lords can endlessly twiddle the knobs on the back end, finding new ways to extract value from you:

https://doctorow.medium.com/twiddler-1b5c9690cce6

That means that your car can track your every movement, and sell your location data to anyone and everyone, from marketers to bounty-hunters looking to collect fees for tracking down people who travel out of state for abortions to cops to foreign spies:

https://www.vice.com/en/article/n7enex/tool-shows-if-car-selling-data-privacy4cars-vehicle-privacy-report

Digitization supercharges financialization. It lets car-makers offer subprime auto-loans to desperate, poor people and then killswitch their cars if they miss a payment:

https://www.youtube.com/watch?v=4U2eDJnwz_s

Subprime lending for cars would be a terrible business without computers, but digitization makes it a great source of feudal rents. Car dealers can originate loans to people with teaser rates that quickly blow up into payments the dealer knows their customer can’t afford. Then they repo the car and sell it to another desperate person, and another, and another:

https://pluralistic.net/2022/07/27/boricua/#looking-for-the-joke-with-a-microscope

Digitization also opens up more exotic options. Some subprime cars have secondary control systems wired into their entertainment system: miss a payment and your car radio flips to full volume and bellows an unstoppable, unmutable stream of threats. Tesla does one better: your car will lock and immobilize itself, then blare its horn and back out of its parking spot when the repo man arrives:

https://tiremeetsroad.com/2021/03/18/tesla-allegedly-remotely-unlocks-model-3-owners-car-uses-smart-summon-to-help-repo-agent/

Digital feudalism hasn’t stopped innovating — it’s just stopped innovating good things. The digital device is an endless source of sadistic novelties, like the cellphones that disable your most-used app the first day you’re late on a payment, then work their way down the other apps you rely on for every day you’re late:

https://restofworld.org/2021/loans-that-hijack-your-phone-are-coming-to-india/

Usurers have always relied on this kind of imaginative intimidation. The loan-shark’s arm-breaker knows you’re never going to get off the hook; his goal is in intimidating you into paying his boss first, liquidating your house and your kid’s college fund and your wedding ring before you default and he throws you off a building.

Thanks to the malleability of computerized systems, digital arm-breakers have an endless array of options they can deploy to motivate you into paying them first, no matter what it costs you:

https://pluralistic.net/2021/04/02/innovation-unlocks-markets/#digital-arm-breakers

Car-makers are trailblazers in imaginative rent-extraction. Take VIN-locking: this is the practice of adding cheap microchips to engine components that communicate with the car’s overall network. After a new part is installed in your car, your car’s computer does a complex cryptographic handshake with the part that requires an unlock code provided by an authorized technician. If the code isn’t entered, the car refuses to use that part.

VIN-locking has exploded in popularity. It’s in your iPhone, preventing you from using refurb or third-party replacement parts:

https://doctorow.medium.com/apples-cement-overshoes-329856288d13

It’s in fuckin’ ventilators, which was a nightmare during lockdown as hospital techs nursed their precious ventilators along by swapping parts from dead systems into serviceable ones:

https://www.vice.com/en/article/3azv9b/why-repair-techs-are-hacking-ventilators-with-diy-dongles-from-poland

And of course, it’s in tractors, along with other forms of remote killswitch. Remember that feelgood story about John Deere bricking the looted Ukrainian tractors whose snitch-chips showed they’d been relocated to Russia?

https://doctorow.medium.com/about-those-kill-switched-ukrainian-tractors-bc93f471b9c8

That wasn’t a happy story — it was a cautionary tale. After all, John Deere now controls the majority of the world’s agricultural future, and they’ve boobytrapped those ubiquitous tractors with killswitches that can be activated by anyone who hacks, takes over, or suborns Deere or its dealerships.

Control over repair isn’t limited to gouging customers on parts and service. When a company gets to decide whether your device can be fixed, it can fuck you over in all kinds of ways. Back in 2019, Tim Apple told his shareholders to expect lower revenues because people were opting to fix their phones rather than replace them:

https://www.apple.com/newsroom/2019/01/letter-from-tim-cook-to-apple-investors/

By usurping your right to decide who fixes your phone, Apple gets to decide whether you can fix it, or whether you must replace it. Problem solved — and not just for Apple, but for car makers, tractor makers, ventilator makers and more. Apple leads on this, even ahead of Big Car, pioneering a “recycling” program that sees trade-in phones shredded so they can’t possibly be diverted from an e-waste dump and mined for parts:

https://www.vice.com/en/article/yp73jw/apple-recycling-iphones-macbooks

John Deere isn’t sleeping on this. They’ve come up with a valuable treasure they extract when they win the Right-to-Repair: Deere singles out farmers who complain about its policies and refuses to repair their tractors, stranding them with six-figure, two-ton paperweight:

https://pluralistic.net/2022/05/31/dealers-choice/#be-a-shame-if-something-were-to-happen-to-it

The repair wars are just a skirmish in a vast, invisible fight that’s been waged for decades: the War On General-Purpose Computing, where tech companies use the law to make it illegal for you to reconfigure your devices so they serve you, rather than their shareholders:

https://memex.craphound.com/2012/01/10/lockdown-the-coming-war-on-general-purpose-computing/

The force behind this army is vast and grows larger every day. General purpose computers are antithetical to technofeudalism — all the rents extracted by technofeudalists would go away if others (tinkereres, co-ops, even capitalists!) were allowed to reconfigure our devices so they serve us.

You’ve probably noticed the skirmishes with inkjet printer makers, who can only force you to buy their ink at 20,000% markups if they can stop you from deciding how your printer is configured:

https://pluralistic.net/2022/08/07/inky-wretches/#epson-salty But we’re also fighting against insulin pump makers, who want to turn people with diabetes into walking inkjet printers:

https://pluralistic.net/2022/06/10/loopers/#hp-ification

And companies that make powered wheelchairs:

https://pluralistic.net/2022/06/08/chair-ish/#r2r

These companies start with people who have the least agency and social power and wreck their lives, then work their way up the privilege gradient, coming for everyone else. It’s called the “shitty technology adoption curve”:

https://pluralistic.net/2022/08/21/great-taylors-ghost/#solidarity-or-bust

Technofeudalism is the public-private-partnership from hell, emerging from a combination of state and private action. On the one hand, bailing out bankers and big business (rather than workers) after the 2008 crash and the covid lockdown decoupled income from profits. Companies spent billions more than they earned were still wildly profitable, thanks to those public funds.

But there’s also a policy dimension here. Some of those rentiers’ billions were mobilized to both deconstruct antitrust law (allowing bigger and bigger companies and cartels) and to expand “IP” law, turning “IP” into a toolsuite for controlling the conduct of a firm’s competitors, critics and customers:

https://locusmag.com/2020/09/cory-doctorow-ip/

IP is key to understanding the rise of technofeudalism. The same malleability that allows companies to “twiddle” the knobs on their services and keep us on the hook as they reel us in would hypothetically allow us to countertwiddle, seizing the means of computation:

https://pluralistic.net/2023/04/12/algorithmic-wage-discrimination/#fishers-of-men

The thing that stands between you and an alternative app store, an interoperable social media network that you can escape to while continuing to message the friends you left behind, or a car that anyone can fix or unlock features for is IP, not technology. Under capitalism, that technology would already exist, because capitalists have no loyalty to one another and view each other’s margins as their own opportunities.

But under technofeudalism, control comes from rents (owning things), not profits (selling things). The capitalist who wants to participate in your iPhone’s “ecosystem” has to make apps and submit them to Apple, along with 30% of their lifetime revenues — they don’t get to sell you jailbreaking kit that lets you choose their app store.

Rent-seeking technology has a holy grail: control over “ring zero” — the ability to compel you to configure your computer to a feudalist’s specifications, and to verify that you haven’t altered your computer after it came into your possession:

https://pluralistic.net/2022/01/30/ring-minus-one/#drm-political-economy

For more than two decades, various would-be feudal lords and their court sorcerers have been pitching ways of doing this, of varying degrees of outlandishness.

At core, here’s what they envision: inside your computer, they will nest another computer, one that is designed to run a very simple set of programs, none of which can be altered once it leaves the factory. This computer — either a whole separate chip called a “Trusted Platform Module” or a region of your main processor called a secure enclave — can tally observations about your computer: which operating system, modules and programs it’s running.

Then it can cryptographically “sign” these observations, proving that they were made by a secure chip and not by something you could have modified. Then you can send this signed “attestation” to someone else, who can use it to determine how your computer is configured and thus whether to trust it. This is called “remote attestation.”

There are some cool things you can do with remote attestation: for example, two strangers playing a networked video game together can use attestations to make sure neither is running any cheat modules. Or you could require your cloud computing provider to use attestations that they aren’t stealing your data from the server you’re renting. Or if you suspect that your computer has been infected with malware, you can connect to someone else and send them an attestation that they can use to figure out whether you should trust it.

Today, there’s a cool remote attestation technology called “PrivacyPass” that replaces CAPTCHAs by having you prove to your own device that you are a human. When a server wants to make sure you’re a person, it sends a random number to your device, which signs that number along with its promise that it is acting on behalf of a human being, and sends it back. CAPTCHAs are all kinds of bad — bad for accessibility and privacy — and this is really great.

But the billions that have been thrown at remote attestation over the decades is only incidentally about solving CAPTCHAs or verifying your cloud server. The holy grail here is being able to make sure that you’re not running an ad-blocker. It’s being able to remotely verify that you haven’t disabled the bossware your employer requires. It’s the power to block someone from opening an Office365 doc with LibreOffice. It’s your boss’s ability to ensure that you haven’t modified your messaging client to disable disappearing messages before he sends you an auto-destructing memo ordering you to break the law.

And there’s a new remote attestation technology making the rounds: Google’s Web Environment Integrity, which will leverage Google’s dominance over browsers to allow websites to block users who run ad-blockers:

https://github.com/RupertBenWiser/Web-Environment-Integrity

There’s plenty else WEI can do (it would make detecting ad-fraud much easier), but for every legitimate use, there are a hundred ways this could be abused. It’s a technology purpose-built to allow rent extraction by stripping us of our right to technological self-determination.

Releasing a technology like this into a world where companies are willing to make their products less reliable, less attractive, less safe and less resilient in pursuit of rents is incredibly reckless and shortsighted. You want unauthorized bread? This is how you get Unauthorized Bread:

https://arstechnica.com/gaming/2020/01/unauthorized-bread-a-near-future-tale-of-refugees-and-sinister-iot-appliances/amp/

If you'd like an essay-formatted version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:

https://pluralistic.net/2023/07/24/rent-to-pwn/#kitt-is-a-demon

[Image ID: The interior of a luxury car. There is a dagger protruding from the steering wheel. The entertainment console has been replaced by the text 'You wouldn't download a car,' in MPAA scare-ad font. Outside of the windscreen looms the Matrix waterfall effect. Visible in the rear- and side-view mirror is the driver: the figure from Munch's 'Scream.' The screen behind the steering-wheel has been replaced by the menacing red eye of HAL9000 from Stanley Kubrick's '2001: A Space Odyssey.']

Image: Cryteria (modified) https://commons.wikimedia.org/wiki/File:HAL9000.svg

CC BY 3.0 https://creativecommons.org/licenses/by/3.0/deed.en

[wackywatchdotcom]

i will say im not super invested in the idea of the show giving us too much information on c&a. like itll be super cool if we find out about it! and im sure itll be relevant information if we do but i think the information we get is going to be not super indepth, because im not sure itd match the tone of the show? its less about mystery and like The Lore and more about the characters and examining how different characters respond and have been responding to the environment set up by the show

(it WOULD be fun to learn more details about c&a and i AM curious i just dont think itd actually enhance much abt the show itself, and as such, i dont think its super important writing wise for the audience to learn about it either)

that SAID i do ascribe to the idea that ALL of the cast members worked for c&a before the ended up in the circus, no exceptions. i dont really know how say gangle or zooble wouldve ended up there, but i feel like they all ended up there one way or another. and i think most of the cast doesnt talk about their life in the real world bc its gone and theres no reason to dredge it up (or theyve forgotten it or most aspects of it), but it does make it fascinating that not even pomni, who has the least reason to avoid these topics at least early on, broached the subject of what in the WORLD is c&a UP to. it literally is never brought up ever as of ep 4. which makes me think the company was probably always super shady in some way to the pt that its employees were kinda desensitized to the idea that its c&as fault that horrible things happen

[sysig]

*chanting* Second pet, second pet, second pet! (Patreon)

[umm-whats-going-on]

Hey all regarding my prior post about AI effect on environment the user @reachartwork immediately blocked me for sharing some information on the environmental impact of AI.

To be fair I was criticizing, but to not be fair it's a product and we should always be critical of everything.

Here are the links I had in the prior post (hopefully in the same order)

1. 2. 3. 4. 5.

[moderatetoaboveaverage]

"Boil water to turn fan" as if multistage steam turbine generators are not one of the sexiest kinds of machines every made

nuclear power is impressive until you get up to why. "we use the most precisely engineered machinery ever created to split atoms to release energy" oh yeah how come? "boil water to turn a fan" get the fuck out

[donotdestroy]

Here are some examples of activities or uses that typically require that amount of water:

Using 500 milliliters (ml) of water is relatively common in everyday life. Here are some examples of activities or uses that typically require that amount of water:

Drinking Water: 500 ml is equivalent to about two standard glasses of water or a typical small bottle of water that people might drink.

Cooking: Many recipes involve boiling water, such as making pasta or rice, where 500 ml might be a starting quantity.

Handwashing: A brief hand wash, depending on the faucet flow rate, can use around 500 ml of water.

Brushing Teeth: Leaving the tap running while brushing teeth for two minutes could use about 500 ml or more of water, though turning it off saves water.

Showering: A very short shower using a water-saving showerhead might consume around 500 ml in a few seconds, as shower flow rates often range from 6 to 12 liters per minute.

Watering a Small Plant: Many houseplants require about 500 ml of water per watering, depending on their size and type.

Coffee or Tea Preparation: Brewing a couple of cups of coffee or tea uses around 500 ml of water.

These examples help illustrate how everyday activities can equate to this seemingly small quantity of water.

By ChatGPT

[jcmarchi]

Can AI automate computational reproducibility?

New Post has been published on https://thedigitalinsider.com/can-ai-automate-computational-reproducibility/

Can AI automate computational reproducibility?

Last month, Sakana AI released an “AI scientist”, which the company called “the first comprehensive system for fully automatic scientific discovery”. It was touted as being able to accelerate science without suffering from human limitations. 

Unfortunately, the “AI Scientist” has many shortcomings. It has no checks for novelty, so generated papers could rehash earlier work. And Sakana did not perform any human review (let alone expert “peer��� review) of the generated papers—so it is unclear if the papers are any good (apparently they are not). While these flaws are particularly flagrant in Sakana’s case, the lack of good evaluation affects most AI agents, making it hard to measure their real-world impact.

Today, we introduce a new benchmark for measuring how well AI can reproduce existing computational research. We also share how this project has changed our thinking about “general intelligence” and the potential economic impact of AI. Read the paper.

Visions of AI automating science are enticing, but aren’t within reach, and lead to flawed science. In contrast, using AI for well-scoped tasks such as verifying computational reproducibility can save a lot of time and redirect effort towards more productive scientific activity. AI could also help find relevant literature, write code to rapidly test ideas, and perform other computational tasks.

In a new paper, we introduce CORE-Bench (Computational Reproducibility Agent Benchmark), a benchmark for measuring how well AI can automate computational reproducibility, that is, reproducing a paper’s findings when the code and data are available. The authors are Zachary S. Siegel, Sayash Kapoor, Nitya Nadgir, Benedikt Stroebl, and Arvind Narayanan. CORE-Bench is a first step in a larger project to rigorously evaluate progress in automating research tasks of increasing difficulty.

Computationally reproducing a study is a far more limited task than replication, which requires re-running experiments that might involve human subjects. Even the limited reproducibility task is hard: In the 2022 Machine Learning Reproducibility Challenge, over a third of the papers could not be reproduced even when experts reproducing the papers had the code and data. 

If AI could automate this mundane yet important task, researchers could automate the implementation of baselines, reviewers could more easily assess if a paper has flaws, and journals and conferences could more easily verify if submitted and published papers are reproducible.

We created CORE-Bench using scientific papers and their accompanying code and data repositories. We used Code Ocean to source papers that were likely to be reproducible. We manually reproduced 90 papers from computer science, medicine, and social science, and curated a set of questions for each paper to be able to verify the answers. 

We release CORE-Bench with three difficulty levels. Tasks in all three levels require the use of both language and vision capabilities. The hardest version closely resembles real-world reproduction attempts, and we expect that improvements on the benchmark will translate to agents that are actually useful to scientists.

To implement baselines, we tested the generalist AutoGPT agent and also implemented a task-specific modification to AutoGPT, which we call CORE-Agent. While the task-specific version improved accuracy significantly, there is still massive room for improvement: the best agent (CORE-Agent with GPT-4o) has an accuracy of 22% on CORE-Bench-Hard.

Computational reproducibility requires setting up the code environment correctly, running the code, and seeing if it produces the same results as reported in the paper. Using the shell and other tools correctly is still tricky for LLMs. When we evaluated generalist agents like AutoGPT, we weren’t surprised by their poor accuracy (less than 10% on CORE-Bench-Hard). 

Yet, with a few person-days of effort, we were able to build CORE-Agent by modifying AutoGPT, which more than doubled accuracy on the hardest level. We also built a task-specific agent from scratch, but modifying AutoGPT was far less time consuming while also resulting in a stronger agent. We are cautiously optimistic that this approach can be pushed to yield agents that perform well enough to be useful in practice. 

Simple task-specific modifications allow CORE-Agent to outperform AutoGPT. 

If this pattern of being able to easily adapt a generalist agent to produce a task-specific agent holds in other areas, it should make us rethink generality. Generality roughly translates to being able to use the same model or agent without modification to perform a variety of tasks. This notion of generality underpins how Artificial General Intelligence (or AGI) is usually understood and the hopes and fears that accompany it. 

But at least from the point of view of economic impacts, generality might be a red herring. For a task such as computational reproducibility on which expert humans collectively spend millions of hours every year, being able to automate it would be hugely impactful — regardless of whether the AI system did so out of the box, or after a few person days (or even a person year) of programmer effort. 

In the AI Snake Oil book, we define generality as the inverse of task-specificity, and analyze how the history of AI (and computing) can be seen as the pursuit of gradually increasing generality. Increasing generality means decreasing the human effort it takes to build an AI system to perform a given task. From this perspective, systems like AutoGPT may be more general than most people (including us) gave them credit for.

Yet, definitions of AGI typically insist that a single system be able to do everything out of the box. There is no systematic effort to track how the human effort needed to build task-specific AI is changing over time. Just as we’ve argued against flawed conceptions of generality that overestimate AI progress, we should avoid flawed conceptions of generality that underestimate it. 

Read the CORE-Bench paper here.

In our recent paper, AI Agents That Matter, we found several shortcomings with AI agent evaluations. While building CORE-Bench, these shortcomings informed the design of our benchmark.

We recently organized an online workshop on useful and reliable AI agents where leading experts shared their views on better agent design and evaluation. The workshop videos are available online.

Ben Bogin et al. released the SUPER benchmark to evaluate if AI agents can set up and execute tasks from repositories accompanying research papers. It is another interesting benchmark for measuring AI agents’ capability to automate research tasks. It differs from CORE-Bench in many ways: 

CORE-Bench consists of tasks across scientific disciplines (computer science, medicine, social science) whereas SUPER consists of tasks from AI.

CORE-Bench requires the use of both vision-language and language models, and consists of multiple languages (Python and R) as opposed to SUPER (language models, Python).

Tasks in SUPER require access to a Jupyter notebook. In contrast, tasks in CORE-Bench require shell access and allow the agent to modify the sandbox arbitrarily.

[nasa]

Seeing the Invisible Universe

This computer-simulated image shows a supermassive black hole at the core of a galaxy. The black region in the center represents the black hole’s event horizon, beyond which no light can escape the massive object’s gravitational grip. The black hole’s powerful gravity distorts space around it like a funhouse mirror. Light from background stars is stretched and smeared as it skims by the black hole. You might wonder — if this Tumblr post is about invisible things, what’s with all the pictures? Even though we can’t see these things with our eyes or even our telescopes, we can still learn about them by studying how they affect their surroundings. Then, we can use what we know to make visualizations that represent our understanding.

When you think of the invisible, you might first picture something fantastical like a magic Ring or Wonder Woman’s airplane, but invisible things surround us every day. Read on to learn about seven of our favorite invisible things in the universe!

1. Black Holes

This animation illustrates what happens when an unlucky star strays too close to a monster black hole. Gravitational forces create intense tides that break the star apart into a stream of gas. The trailing part of the stream escapes the system, while the leading part swings back around, surrounding the black hole with a disk of debris. A powerful jet can also form. This cataclysmic phenomenon is called a tidal disruption event.

You know ‘em, and we love ‘em. Black holes are balls of matter packed so tight that their gravity allows nothing — not even light — to escape. Most black holes form when heavy stars collapse under their own weight, crushing their mass to a theoretical singular point of infinite density.

Although they don’t reflect or emit light, we know black holes exist because they influence the environment around them ��� like tugging on star orbits. Black holes distort space-time, warping the path light travels through, so scientists can also identify black holes by noticing tiny changes in star brightness or position.

2. Dark Matter

A simulation of dark matter forming large-scale structure due to gravity.

What do you call something that doesn’t interact with light, has a gravitational pull, and outnumbers all the visible stuff in the universe by five times? Scientists went with “dark matter,” and they think it's the backbone of our universe’s large-scale structure. We don’t know what dark matter is — we just know it's nothing we already understand.

We know about dark matter because of its gravitational effects on galaxies and galaxy clusters — observations of how they move tell us there must be something there that we can’t see. Like black holes, we can also see light bend as dark matter’s mass warps space-time.

3. Dark Energy

Animation showing a graph of the universe’s expansion over time. While cosmic expansion slowed following the end of inflation, it began picking up the pace around 5 billion years ago. Scientists still aren’t sure why.

No one knows what dark energy is either — just that it’s pushing our universe to expand faster and faster. Some potential theories include an ever-present energy, a defect in the universe’s fabric, or a flaw in our understanding of gravity.

Scientists previously thought that all the universe’s mass would gravitationally attract, slowing its expansion over time. But when they noticed distant galaxies moving away from us faster than expected, researchers knew something was beating gravity on cosmic scales. After further investigation, scientists found traces of dark energy’s influence everywhere — from large-scale structure to the background radiation that permeates the universe.

4. Gravitational Waves

Two black holes orbit each other and generate space-time ripples called gravitational waves in this animation.

Like the ripples in a pond, the most extreme events in the universe — such as black hole mergers — send waves through the fabric of space-time. All moving masses can create gravitational waves, but they are usually so small and weak that we can only detect those caused by massive collisions.  Even then they only cause infinitesimal changes in space-time by the time they reach us. Scientists use lasers, like the ground-based LIGO (Laser Interferometer Gravitational-Wave Observatory) to detect this precise change. They also watch pulsar timing, like cosmic clocks, to catch tiny timing differences caused by gravitational waves.

This animation shows gamma rays (magenta), the most energetic form of light, and elusive particles called neutrinos (gray) formed in the jet of an active galaxy far, far away. The emission traveled for about 4 billion years before reaching Earth. On Sept. 22, 2017, the IceCube Neutrino Observatory at the South Pole detected the arrival of a single high-energy neutrino. NASA’s Fermi Gamma-ray Space Telescope showed that the source was a black-hole-powered galaxy named TXS 0506+056, which at the time of the detection was producing the strongest gamma-ray activity Fermi had seen from it in a decade of observations.

5. Neutrinos

This animation shows gamma rays (magenta), the most energetic form of light, and elusive particles called neutrinos (gray) formed in the jet of an active galaxy far, far away. The emission traveled for about 4 billion years before reaching Earth. On Sept. 22, 2017, the IceCube Neutrino Observatory at the South Pole detected the arrival of a single high-energy neutrino. NASA’s Fermi Gamma-ray Space Telescope showed that the source was a black-hole-powered galaxy named TXS 0506+056, which at the time of the detection was producing the strongest gamma-ray activity Fermi had seen from it in a decade of observations.

Because only gravity and the weak force affect neutrinos, they don’t easily interact with other matter — hundreds of trillions of these tiny, uncharged particles pass through you every second! Neutrinos come from unstable atom decay all around us, from nuclear reactions in the Sun to exploding stars, black holes, and even bananas.

Scientists theoretically predicted neutrinos, but we know they actually exist because, like black holes, they sometimes influence their surroundings. The National Science Foundation’s IceCube Neutrino Observatory detects when neutrinos interact with other subatomic particles in ice via the weak force.

6. Cosmic Rays

This animation illustrates cosmic ray particles striking Earth's atmosphere and creating showers of particles.

Every day, trillions of cosmic rays pelt Earth’s atmosphere, careening in at nearly light-speed — mostly from outside our solar system. Magnetic fields knock these tiny charged particles around space until we can hardly tell where they came from, but we think high energy events like supernovae can accelerate them. Earth’s atmosphere and magnetic field protect us from cosmic rays, meaning few actually make it to the ground.

Though we don’t see the cosmic rays that make it to the ground, they tamper with equipment, showing up as radiation or as “bright” dots that come and go between pictures on some digital cameras. Cosmic rays can harm astronauts in space, so there are plenty of precautions to protect and monitor them.

7. (Most) Electromagnetic Radiation

The electromagnetic spectrum is the name we use when we talk about different types of light as a group. The parts of the electromagnetic spectrum, arranged from highest to lowest energy are: gamma rays, X-rays, ultraviolet light, visible light, infrared light, microwaves, and radio waves. All the parts of the electromagnetic spectrum are the same thing — radiation. Radiation is made up of a stream of photons — particles without mass that move in a wave pattern all at the same speed, the speed of light. Each photon contains a certain amount of energy.

The light that we see is a small slice of the electromagnetic spectrum, which spans many wavelengths. We frequently use different wavelengths of light — from radios to airport security scanners and telescopes.

Visible light makes it possible for many of us to perceive the universe every day, but this range of light is just 0.0035 percent of the entire spectrum. With this in mind, it seems that we live in a universe that’s more invisible than not! NASA missions like NASA's Fermi, James Webb, and Nancy Grace Roman  space telescopes will continue to uncloak the cosmos and answer some of science’s most mysterious questions.

Make sure to follow us on Tumblr for your regular dose of space!

[easternblot]

Generative AI is the type artists and writers hate: the kind that makes up nonsense "facts", creates hands with six fingers, that sort of thing. It's different from the kind of useful AI that, for example, helps doctors find tumours in medical scans or helps scientists spot patterns in data. Generative AI generates new things that didn't exist before, and if that was not enough reason to dislike it, it's also TERRIBLE for the environment!

[jackels-in-space]

I’m sorry but please stop using AI.

Not only does it use other people’s work as base training (let’s not forget about the fact that C.AI creators have been taking fics to build up chat bots) it also like…destroys the environment.

They need computers capable of giving you quick responses and that generates heat and what cools things down? AC and water.

Genuinely go search it up it’s insane.

Stop using AI.

If you feel like you want to post and you’re like “oh but I don’t think I’m good enough” I sympathise I really do every writer friend I know has had that exact same experience but you’re never gonna get better if you don’t try.

You’re never gonna learn anything if you use AI so stop assisting big corporations in destroying the world and actually try.

If you feel scared about your actual writing if it’s any good or not I am so happy to beta it for you but please just try that’s all I’m asking.

I’ve supplied a couple of links so you know I’m talking bullshit but please do your own investigation <3

[athenaeum-of-the-herald]

"Hermes is the god of thieves so he's fine with art theft!" No.

"Dionysos wouldn't mind it he's chaotic!" No.

"Hephaistos is a machinery god and AI is machinery so it's okay!" No.

"The gods wouldn't even care they just want offerings!" No.

There is not, and never has been, a reason to offer the gods AI art. It requires absolutely no effort or thought on your part as a worshipper, which defeats the purpose of an offering.

Do not support the use of AI art at ALL, much less as something given to Them.

"But it's all I can give!" Water. The Theoi fully accept a glass of water.

Sing.

Do a craft.

Write a letter.

Draw.

Dedicate a task.

Please do not offer those you worship something so lackluster and thoughtless as a computer generated image that destroys our environment and piggybacks off the works of others.

[yuumei-art]

First, I want to address the rumors that I used AI to paint A Thousand Skies. Here is a condensed version of the 10 part video recording of how I painted everything. I normally reserve the full hours long video recordings for my patrons but here they are in case there is still any doubt https://drive.google.com/drive/folders/1d-3hxjZWROiZPaI8XeU93cKBBKB1Gz9Z?usp=drive_link

All of my past paintings have full video recordings for my Patrons, going back to 2018 when I first got a Patreon.

Long time followers will also know that I care deeply about the environment, not only in the themes of my art but also in my actions to the best of my ability and knowledge. So much so that every year since 2016, I’ve donated 100% of my Black Friday sales to environmental charities, and you can see all the receipts of the organizations I donated to here https://www.yuumeiart.com/blog 

I realized a comment I’ve made at 2am and while sleep deprived was confusing. I want to clear up the misunderstandings from yesterday's post where I said AI isn’t that bad for the environment. I thought I was specifically answering the question of how much energy it takes to generate 1 AI image on my own computer without any servers, but I realize now that people were talking about how much energy it takes to train AIs by big companies. This was not what I was trying to refer to in my comment, but I recognize it is one of many legitimate concerns with the way current generative AI is developed.

I want to reiterate that I did use AI for a gag scene in my comic—the character, Vance, was objectifying women by seeing them as anime cat girls and pasting them over AI backgrounds. I thought that such a character having AI goggles was making a point. I recognize that regardless of the context, the use of AI carries other ramifications. I will be taking all of this into consideration with my future artwork.

I know AI art is a very heated issue with very real concerns attached to its use. I don’t expect people to change their minds. I will continue to make art as I always have for the past 2 decades before AI existed, and continue to make available to my patrons full video recordings of the paintings I’ve made and will make. Thank you.

[powdermelonkeg]

Cheat Code #3 for accommodating disabled characters in sci-fi/fantasy:

If you want your setting to be accommodating, change the environment more than the person.

i.e.: On a worldbuilding level, if you want to portray a society that keeps disabled people in mind, then that needs to be reflected more broadly, even without your disabled character on screen. Because this means that your society was considering disabled people as part of itself when it was figuring out what's necessary.

If your computer takes voice commands, it should also have an optional keyboard in case someone can't speak.

If your magic school has multiple floors, it should have a teleporting rune circle for those that can't take the ever-changing stairs.

Whenever you have a feature you're adding, ask yourself—"If my character couldn't use this, what would they do instead?" And if the answer is "they'd have to wait until they could" or "they need someone else to use it for them," then your setting isn't accommodating. An accommodating setting always has an actionable answer to that question.

And as a bonus, if you follow through with it, oftentimes you'll end up with a more interesting world and story overall. Spells most people can speak can be written in ancient elven instead? That means you can have a character sneak a spell into a magic-banned city by writing it on their hair ribbon, and that it's possible that a book might be a self-generating spell on its own. Your spaceship has textured lines on the walls to let blind people navigate without guidance? Not only can you make it look artistic (different colored paints, glowing patterns), but now your engineer can make it to the warp core when the power's out and oxygen's finite.

Don't limit yourself just to what's needed in the moment. Figure out interesting alternatives to your setting's features, and your world will automatically feel more alive.

Cheat Code 1: How to avoid eliminating disability in your setting

Cheat Code 2: What kinds of aid to use to accommodate disability

Cheat Code 4: How to personalize your character's disability aid

[bloomzone]

STUDY SYSTEM : DAILY STUDY ROUTINE ( EXAM EDITION)

hii looves so this blog is all abt how to optimizing your day for peak academic performance. This four-part daily system is the exact routine used during exam season especially finals to consistently achieve top grades with minimal stress. While the routine may appear complex at first glance it is built upon simple, intuitive principles rooted in human biology and psychology. It is not only practical but highly effective when followed consistently. You’ll learn to structure your day around ur natural energy cycles, use tools to boost focus, and incorporate essential periods of rest and release to maintain motivation and productivity.

SECTION 1: UNDERSTANDING UR ENERGY RHYTHM

The routine is built around the concept of the circadian rhythm, your body’s natural energy cycle throughout the day. On a standard day for example :

Energy peaks shortly after waking typically around 6:00 AM.

Energy dips mid-afternoon often around 2:30 or 3:00 PM.

A secondary energy peak occurs in the evening approximately around 7:00 PM.

Energy tapers off as bedtime approaches.

This predictable fluctuation is key to optimizing your study schedule. The two energy peaks will be your core study sessions, while the dip will be used as a rest period, and the late-night wind-down becomes your release period or bedtime .

SECTION 2: THE MORNING STUDY SESSION (STUDY SESSION #1)

◜⁠✧ Start Within One Hour of Waking Up

Your goal is to begin studying as soon as possible after waking, ideally within the first hour. Use the high energy of the morning to tackle your most challenging subjects.do ur morning routine quick as possible don't do intense workout or stuff like this cuz u will waste ur energy so always have a specific morning routine for days like these ! And u can workout go to the gym or whatever at the rest period !

Pre-Study Essentials:

1. Set Your Daily Goals (5 minutes)

Before starting, sit with a notebook or your computer and write down what you intend to accomplish. Be specific. Define exact tasks e.g. which past papers you’ll solve, which topics to review so u will be more organized during the day

2. Activate Focus with Three Optional Tools:

Caffeine — coffee, yerba mate, or tea to boost dopamine and adrenaline.

Cold Showers — a physiological wake-up that increases alertness (personally I don't do that lmao but if u can that's good !)

Focus Warm-up (1–2 minutes) — pick a point in your environment and concentrate on it intensely. This warms up your cognitive focus system before you start.

◜⁠✧Deep Work Sprint Format

Commit to a 2–4 hour study block.

Use timed work intervals: e.g., 25–30 minutes of focused work followed by 5-minute breaks.

Use a visual timer to create urgency and focus. This serves as a “deadline generator,” helping you push harder and maintain hope by offering visible progress.

─ ⊹key principle: work Like a warrior

Study in focused sprints. The more intense your focus, the less time you’ll need to study. The idea is depth over duration not 12 hours of mediocre attention, but 2–4 hours of deep concentration.

SECTION 3: MIDDAY REST PERIOD

Timing: After First Study Block Ends (~Early Afternoon)

At this point in the day, your energy naturally dips. It’s essential to give yourself permission to rest. This period is not for distractions like Netflix, YouTube, or social media.

◜⁠✧ Approved Activities:

Exercise or light sports

Socializing with friends/family

Taking a walk, especially outdoors

Napping (ideal: 20 minutes)

The goal here is active recovery choose activities that contrast focused work. Avoid anything with dopamine stimulation that mimics your “vices” or release behaviors.so this break allows your mind to reset, preventing burnout and increasing productivity in the next session.

SECTION 4: EVENING STUDY SESSION (STUDY SESSION #2)

Timing: During the Second Energy Peak (~6:00–8:00 PM)

Return for your second battle. This session is similar in structure to the morning study session, but with a few differences:

◜⁠✧ Change Your Environment:

Consider studying in a different location e.g., library, a new room, or another productive setting. (Personally I move from my desk to the guest room cuz it's far from family chaotic activities ifykyk )

This provides novelty and reduces boredom, which helps counteract distractions that are more likely to arise in the evening.

◜⁠✧ Eliminate Distractions:

If possible, leave your phone behind take that shit in another room

Create a space where your brain associates the environment with productivity.

◜⁠✧ Study Format:

Continue using timed sprints.

Session length: 2 to 3 hours, depending on your focus reserves.

The goal is to extract one final productive effort from your remaining focus reserves for the day.

SECTION 5: NIGHTLY RELEASE PERIOD

Timing: 1–2 Hours Before Bed

This period is crucial and often overlooked. It functions as your psychological release valve a scheduled time for indulging in your “vices” or desires.

◜⁠✧ why it matter

Without a controlled release period, distractions tend to creep in throughout the day. When you tell yourself you’ll “resist” TikTok or YouTube for three straight weeks during exams, it almost always backfires. You end up scattering distractions across the day, killing momentum and u will feel like shit

◜⁠✧ so solution:

Contain those activities to this specific window. Give yourself full permission to indulge whether it's gaming, scrolling, or Netflix. The only rule: Only do it at night.

◜⁠✧ psychological benefit:

You’ll find it easier to say “no” to distractions earlier in the day when you know you can give in later. It reduces the mental burden of constant suppression.

Caution:

This is not a prescription to develop new addictions or deepen existing ones. If you don’t feel the need for this release, skip it . But if you’re honest with yourself about your impulses, this structure helps you keep them in check.

◜⁠✧ ADAPTATION AND FLEXIBILITY

✧ Everyone has a unique biology. Some wake up at 5:00 AM, others at noon.

✧ Adjust the energy curve and study blocks to match your personal circadian rhythm.

✧ This is a template, not a strict prescription. Principles stay constant, execution varies.

KEY TAKEAWAYS:

🗝️ Two deep-focus sessions aligned with your body’s energy peaks yield greater results than dragging your mind across a 12-hour marathon.

🗝️ Midday rest and nightly release are components of a sustainable routine.

🗝️ Use tools like caffeine, timers, environment changes, and goal setting to maintain momentum and focus.

🗝️ Structure breeds freedom. When your day is mapped with intention, your brain is free to focus trust me with this one

✧ This daily routine is not about rigid hours or perfection. It’s about aligning your habits with your biology and respecting your mental bandwidth. When implemented consistently, this system transforms exam season from a stressful grind into an enjoyable and productive challenge.Now take what you've learned and design your daily routine with intention ✧

@bloomzone

[sandersstudies]

So the AI ask wasn't spam. I'd highly encourage you to do some research into how AI actually works, because it is neither particularly harmful to the environment, nor is it actually plagiarism.

Ignoring all of that however, my issue is that, fine, if you don't like AI, whatever. But people get so vitriolic about it. Regardless of your opinions on if it's valid art, your blog is usually a very positive place. It was kind of shocking to see you post something saying "fuck you if you disagree with me, your're a disgrace to the community." Just felt uncharacteristicly mean.

Even if you insist AI isn’t actively harmful to the environment or other writers (and the research I have done suggests it is, feel free to send me additional reading) and you simply MUST use prompts to generate personal content, nobody has any business posting it in a creative space for authors, which was the specific complaint addressed in that original post. While I’ll never say “fuck you for who you are as a person” on this blog, I might very well say “fuck you for harmful or rude actions you’ve taken willingly,” which is what that post was about.

Ao3 and similar platforms are designed as an archive for fan content and not a personal storage place for AI prompt results. It is simply not an appropriate place. If you look in the notes of the previous ask you will see other people have brought up additional reasons they have concerns about this practice.

A note on environmental effects for those who might not know: Generative AI requires MASSIVE amounts of data computers operating. As anyone who has held a laptop in their lap or run Civ VII on an aging desktop computer, computer équipement generates a lot of heat. Even some home and small-industrial computers have water-cooling systems. The amount of water demanded by AI computers is massive, even as parts of the world (even in America) experience water shortages. Besides this, it consumes a lot of power. The rising demand for AI and the improvements demanded to keep it viable mean this problem will continue to scale up rather than improve. Of course, those who benefit from the use of AI continue to downplay these concerns, and money is being funneled into convincing the public that these are not real concerns.

I have been openly against the use of generative AI, especially for art and writing, since its popularity rose in the last couple years. I’m sorry I wasn’t clearer about this stance sooner. I have asked my followers to alert me if I proliferate or share AI content, and continue to do so.