We need to talk about data centres.
For the 2nd or 3rd time this week I've seen someone comment on a new data centre build with a stat about how 80% of data is never accessed. Then they talk about the energy and cooling used in modern DCs.
The reality is that data storage is actually incredibly efficient, and uses fuck all power. A hard disk is less than 10w and stores multiple users data.
Storing data, our photos, our memories, our history. Is not the problem.
What is? 1/n
The thing driving the need for the bigger more power and water hungry data centres is AI. Sparkling autocarrot. Where as a machine in a rack full of hard disks might consume a couple of hundred watts. A machine loaded up with a typical load of 8 "AI accelerators" can be pulling in the region of 5kw. Over an order of magnitude more power than the energy needed to store the lifes photos of hundreds of people.
And why ? To what end?
2/n
I've worked in this industry for over a quarter of a century.
At no point have I found myself thinking "I wish I could just ask the computer to write this email for me" "I wish the computer could write my code for me". MS is adding co pilot function to lots of products. Not as opt in. But opt out. And it's a right hassle to turn it off? Why? So someone can ask it to right a longer email from a prompt that the recipient can then ask the AI to summarise for them ?
3/n
There are certainly some areas where machine learning (note, I'm not calling it AI) has it's uses. Medical research springs to mind. But a ubiquitous AI assistance rolled into all our products? Why? It's just using too much power, too many resources, and for what ? Sparkling autocarrot.
Encoding the worst of our society in a bit stream. Exacerbating inequality, prejudice, and hate.
In comp sci there's a term. GIGO. Garbage in. Garbage out.
4/n
These large language models are being fed on the combined mass of the world's online content. Your tweets, Facebook posts, forum posts, that blog you forgot you started. All of it is being fed into the black box of the LLM. The internet provides us unprecedented access to the world's information. But it is also an unprecedented collection of hate. We've seen this time and time again. From chat bots that start shouting nazi propaganda, to CV vetting systems that won't hire women.
5/n
Garbage in. Garbage out.
And what makes this even more terrifying is that when you look at a webpage, it's often hard to tell if it's been generated with sparkling autocarrot, or written by a human. If we can't tell, then what hope does the LLM? And so we're gonna end up with the next generation models being fed on the output of the previous. This is going to create feedback loops. Reinforcing the worst the model has to offer. Strengthening the hate. The prejudice.
6/n
And because we don't know what has been created how. There's no way to control what feeds the models. It's just gonna enshitify. And fast.
AI has all the hallmarks of a bubble. Like crypto before it, and half a dozen other bubbles before that, that all share their heritage right back to the south sea bubble (no, not tulip mania, but that's something for a different thread).
Except this bubble has gone more mainstream. It's consuming way more resources. Than any before it.
7/n
Water is going to become the next big inequality front. As the climate changes. Clean fresh water is going to become harder to come by. More expensive, and more unequal. That same water is being poured over panels I'm data centres to cool the servers. To cool the AI accelerators, generating content noone asked for. Enshitifying the knowledge base of humanity. Just so a few people can make some money.
8/n
Storing our data, our memories, our photos, on servers in data centres that are built in sensible places isn't inherently a bad thing. And we shouldn't allow ourselves to fall for the trope of 80% of it is never accessed. But building datacentres that use ten times the energy, and need even more water, in deserts, and water stressed areas, to drive sparkling autocarrot that noone asked for. That we should be more vocal about.
9/9.
@quixoticgeek Training GPT-3 took as much water as 460 hamburgers.
You raise a very good point about location of data centers. There's plenty of room along the Wisconsin and Michigan coastlines with access to more fresh water than could ever be consumed in a thousand years. If water were priced appropriately in drought-ridden areas, I imagine data centers would be happy to relocate. First step is to fix the broken politics that subsidizes silage/beef
@quixoticgeek as a practical matter, don't think we can guilt trip users into pretending a given convenience a) isn't cool after all or b) should not be used. People get into accidents because they're idiots and use their cell phones while they're driving. Even so, cell phones are part of life now.
What we can do is tax/regulate the hell out of large commercial entities that, in making cool new tech convenient, move us 1 cm closer to climate disaster. You don't get to make money for doing that.
But that's a handwave. "We" won't tax the hell out of them, any more than we tax the hell out of billionaires.
This "incentivize good choices" model is what's gotten us an environment in which everybody is pissing microplastics and the amount of CO2 in the atmosphere went up by 1% in a year in 2023.
Regulation works if people decide to do it. This argument is what's been convincing otherwise smart people not to try to do it.
Answer: vote and organise. For Democrats, if you happen to live in the USA.
And yes, this is why Europe is farther along on this. Americans can't seem to grasp the absurdly high stakes. Every syllable of what you just wrote is why I'm trying to make sure we don't get distracted/manipulated into electing a climate denier slate nationwide. Govt must take this seriously.
Guilt tripping users isn't the answer regardless. That's another distraction in my humble opinion.
And by the way I would start with Exxon when it comes to taxing and regulating the hell out of people
I agree. But I don't think OP was guilt tripping anyone. OP was reporting what's going on. That's important.
This felt like guilt tripping:
"But building datacentres that use ten times the energy, and need even more water, in deserts, and water stressed areas, to drive sparkling autocarrot that noone asked for. That we should be more vocal about."
I prefer we talk about legislation that keeps companies responsible. I disagree that I don't want AI. I *do* want AI. Inaccurate to say no one is asking for the next wave of development. I am. Even if I don't know what it is.
Okay, but the ask was literally "that we should be more vocal about" [it].
I think LLMs are interesting, and GAI would also be interesting. But we actually already have GAI. It's called corporations and societies. These are artificial cognitive structures, built by humans, running on a substrate that is known to be intelligent.
So if you're interested in how to improve the state of AI, you don't have to wait!
Just give the people who are building responsible data centers a huge subsidy and slap huge taxes and fines all the rest. I don't try to dictate what the marketplace is going to create, or what is going to be cool, a year from now. It's pointless.
Maybe a little polderpolitiek wouldn't hurt, though. I agree that we shouldn't be legislating outcomes, but right now we're deliberately pricing externalities out, which is a subsidy that always and only benefits approaches that are wasteful. We are not starting with a level playing field, so of course we see abuses.
BTW, Americans do grasp the absurdly high stakes, and the majority vote accordingly.
The difference between America and, say, the Netherlands, is that in the Netherlands when Geert Wilders gets the biggest number of seats, the question is, will Yesilgƶz cooperate with him. In the U.S., it's "well, I guess Trump won even though the majority voted against him."
Our news media is Pravda from the 80's, even the "left media." It's very difficult here to get real news.
@YusufToropov @abhayakara @quixoticgeek
*Your* answer is facile. Yes, organize, and slow the collapse with voting when it'll work, but voting is less and less powerful as time goes on thanks to an unpacked SCOTUS and very well-lubed fascist political regime. National Dems are at best nearly completely useless, either unable or--more likely--unwilling to pursue policies that will alienate bourgeois donors.
Dems aren't the answer, and never can be. Nor are you making sure of anything, expat.
@YusufToropov @abhayakara @quixoticgeek
I doubt you've had your foot on the ground in a while, since you don't seem to understand that climate denial was already the entire sweep in 16. In 2000, when SCOTUS decided an election for a climate-denialist and oil exec.
We get the stakes, we're just fucking exhausted. Work is hell and we can barely afford a roof over our heads, and the population that can't grows all the time.
Standard Oil should've been and can still be nationalized.
@quixoticgeek Stochastic Parrots Considered Harmful
Sparkling Autocarrot Considered Harmful
Take a look some time at Ireland's current and short term projected datacentre electricity use as % of total generation. 20% now, 70% by 2030. That's what a friendly corporate tax haven regime gets you these days.
@quixoticgeek I'd like to highlight that "AI" isn't actually anything new, its just larger than ever now. Which is why this serverfarm issue isn't new. Before it mostly came in the form of surveillance advertising & personalized recommendations.
*Our need for an income is what drives waste!*
The legit value of serverfarms (though I am sympathetic to arguments that we over-rely on this) is near-entirely in data storage/publishing. And message routing.
The legit value costs near-nothing.
Burn.
The.
Servers.
@quixoticgeek one of the things that bothers me is the Snake Oil that is big data.
Computing 101 : data + context = information.
Google, Facebook and other similar horse traders do not have this context element. But sell their wares as customer information.
I donāt need Microsoft to be my editor. I am purposefully using verbose language, because there are steam powered airships in my writing.
There is no place for AI in my life exept for my x-rays. Else more snake oil.
There are over 550,000 Private pools in AZ and 1,300,000 pools in Ca.
This doesn't include resorts.
Then there are the golf courses..
@quixoticgeek
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
We all need to start signing our work.
-----BEGIN PGP SIGNATURE-----
iGwEAREKACwWIQTiPCHthk/086cRTN/KRx/VAWGKSQUCZdtADg4cZGFsZUByZG1w
Lm9yZwAKCRDKRx/VAWGKSageAKCHXHNhCsgXCuCg853RYmdx+6ssoQCeJDnXNTLk
nNPxcQO+qq1BXHStaXA=
=kZbF
-----END PGP SIGNATURE-----
Stable Diffusion revolutionised image creation from descriptive text. GPT-2, GPT-3(.5) and GPT-4 demonstrated astonishing performance across a variety of language tasks. ChatGPT introduced such language models to the general public. It is now clear that large language models (LLMs) are here to stay, and will bring about drastic change in the whole ecosystem of online text and images. In this paper we consider what the future might hold. What will happen to GPT-{n} once LLMs contribute much of the language found online? We find that use of model-generated content in training causes irreversible defects in the resulting models, where tails of the original content distribution disappear. We refer to this effect as Model Collapse and show that it can occur in Variational Autoencoders, Gaussian Mixture Models and LLMs. We build theoretical intuition behind the phenomenon and portray its ubiquity amongst all learned generative models. We demonstrate that it has to be taken seriously if we are to sustain the benefits of training from large-scale data scraped from the web. Indeed, the value of data collected about genuine human interactions with systems will be increasingly valuable in the presence of content generated by LLMs in data crawled from the Internet.
@quixoticgeek That is the term I've been looking for, to summarize LLMs and generative systems to non-tech people. GIGO machines.
It's not just AI.
AI is ofc exponentially worse than data centers but the latter are a huge problem.
Absolutely!
And yet, DCs are a big problem.
I'm not saying abolish them! But they must be powered by renewables and the waste heat (is this the correct term? Or Denglish? š ) capitalized on.
Quixoticgeek
aekis projects
Jamie McCarthy
Yusuf Toropov
Ted Lemon
SouprMatt
Steve Popovich
Psittacus cafeina
Eric Schultz
Jeremy Kun
ā“ā“ā“ĆšĘ.Ęźā“š„Ńɼš©š»āš»
MacBalance
Cassandrich
Alan Sill
Final Round Player š·šŖšŗšø
(moved)Galbinus Caeli šÆ
BobHy
alcinnz
Gerard Cunningham āļø
Nefarious Celt
Ćtienne Parmentier
Nuncio Bitis š³ļøāš šµšø šŗš¦
Sex Ed for Bi Guys 
Gatekeeper..
Kevin Karhan 
Tom Carlson
Khleedril
David (he/him)
Chris Samuel
Guido Stevens HAS MOVED
š»Frā“ HaĆ«nš»
OldGeek
Matthew Toohey
UpLateGeek
David Reinertson
Lil Meow Meow