Mastodawn

Lauren Weinstein

One of the Big Lies you'll hear proponents of AI data centers spout is that if you stand in their way you will block progress toward new cancer cures and other medical progress. This is total bull. The vast amount of this AI data center capacity is aimed not at that kind of AI work, but at LLM generative AI SLOP: Creating sexualized and other deep fakes, automated phishing scams, misinformation-filled search query answers that starve sites of views, agentic agents that destroy your data and wallet while the AI firms take no responsibility, AI chatbots that urge users toward murder and suicide, cheating machines for kids in school, vibe coding systems that generate garbage code and create an excuse for mass layoffs -- and so on. That's the reality of Big Tech LLM AI, and that is what these horrific data centers are mostly all about.

Spot on. Analytic AI combined with machine learning has vast potential. It has already yielded breakthroughs in many areas, and will continue to do so.

Generative AI... produces slop and, more dangerously, misinformation. At great expense. And environmental destruction. The demand for data centres is almost entirely driven by generative AI.

Oppose those data centres

Things AI will be good at (even if it isn't yet):

Interpreting medical images - mammograms, X-rays, scans

Interpreting other images - eg, maintenance of motorway bridges used to require a scaffold for a person to take a look. Drone imaging plus AI will be more efficient.

Generating hypotheses - seeing patterns in data, for humans to question and interpret.

AI will learn to do these things better than humans.

Humans are still needed, to check and be accountable.

:EA DATA. SF:3d ago

@regordane @lauren But not generative AI which is what these data centers are for.

@regordane @lauren

This is why I got crook at Firefox when they wanted to make blocking "AI" in Firefox also block machine translation.

@regordane @lauren Exactly. I try to keep repeating this everywhere, but it doesn't feel like the difference sinks in.

And it's super-important. Because the successes coming from (let's call it?) "expert" AI rubs off on stupid "autocomplete" AI where it's simply an outright lie.

@lauren So just like humans then (only half joking here!)

@lauren
Oh, don't forget hoarding all the content of our computers, phones, records of every site we've ever visited, everywhere we went in our vehicles, everything you bought, etc.

gotofritz 3d ago

Isn't that what people used to say about the web in the early days?

Lauren Weinstein 2d ago

@gotofritz Eh? What are you referring to?

gotofritz 2d ago

Your rant: "LLM generative AI SLOP: Creating sexualized and other deep fakes, automated phishing scams, misinformation...."

Isn't that what people used to say 30 years ago about the web?

Lauren Weinstein 2d ago

@gotofritz Nope.

gotofritz 2d ago

Au contraire....

[it was a rhetorical question]

@lauren
Part of why I don't like the term AI anymore. Confuses LLM with other machine learning activities. Many things previously called machine learning appear to be moving to being called AI to get on the band wagon.

Cassandrich 3d ago

@sab38 @lauren Their loss trying to get a quick buck rebranding to ride the scam. "AI" has always meant deception.

Mousa Al Bateh | Cloud Security & Architect 3d ago

@lauren My main concern is that many data centers use water for cooling.

tribactam 2d ago

they always wheel this one out. Researchers have been using machine learning for decades before the rise of AI slop (chemo- and bio-informatics).

It can't be said often enough, these data centres are about control, surveillance and ensuring stupidity.

Oh, and in any case, don't expect to get treated with the medicines discovered using specialised ML systems - you won't be able to afford them.

@lauren oh, man, you are not kidding. Almost every ML-based life science invention that I work on as a patent attorney runs on a laptop.