How many studies do researchers need to do before the threat of LLMs is taken seriously? This technology *might* have some useful niche applications, but widespread deployment will be a disaster for humanity.
This shit is an existential hazard, and not in the way the AI companies love to talk about. It's not going to take over the world like Skynet, it's a cognitohazard that turns anyone that interacts with it into an idiot.
I should stop talking about AI. Like so many other social issues, I'm only preaching to the converted. All it does is remind people of Yet Another Bad Thing You Have No Power To Change.
Or to put it a different way, whatever levers you have available to influence the situation are probably already being pushed. Me reminding you doesn't do anything other than make you feel worse.
@malcircuit These conversations are useful, if only to help us articulate our shared concerns to others. While we may not individually control the situation, we are able to collectively improve our odds.
For example, I've been reading more about cognitive load and debt. I mentioned it a work conversation on Monday, and my boss used that language with her leadership yesterday, then we had a followup conversation today about how difficult it is to train humans to catch unintuitive mistakes.
These posts, including the amulet and atrophy metaphors, will directly help me to continue those conversations and start new ones. Articles and posts like these help us to build definitions, cite examples, and reframe the narrative, so thank you for forwarding this and for adding your perspective and imagery.
@malcircuit People don't care about studies. They care about money and right now LLMs promise them more money.
They're wrong. They'll be proven wrong. But it'll take time and some MASSIVE failures, first.
Fortunately, if the news is to be believed, Zuckerberg is doing his hardest to destroy Meta with AI so it may come sooner than later.
@faithisleaping @malcircuit I wish I could believe that it will eventually be proven to be bad. I'm afraid it's going to become a subliminal part of society and stick with us forever. Technologies have a tendency to do that no matter how harmful they are.
Broadcasting brought us manufactured consent.
Cable brought us cognitive overwhelm.
Social media brought psyops to the individual.
I'm afraid AI be around forever and destroy conceptual diversity.
@sabrina @malcircuit It will likely be around forever in some form. We're not going to just shut it all down. Too many people have spent too much money.
But it will likely become part of the background noise that we learn to navigate.
But also, you say social media brought psyops to the individual but here we are talking on social media. It also brought human connection across continents.
Which isn't me saying that we should look for the silver lining to AI. It's pretty shit at 95% of what they're trying to use it for. I'm just pushing back a bit on the "technology is destroying everything" narrative. It's changing everything. It's destroying some things. Other things get created in its wake. Humanity will probably survive this one, too.
@faithisleaping @malcircuit Mostly I'm frustrated that we continue to dive into technology after technology without learning to moderate the harms. With AI we know a lot of the harms and we're still failing to moderate them. (I don't know how well we knew the harms of past technologies early in their use as well as we know the harms of AI this time.)
Intuitively I don't agree that technology is destroying everything. I don't want to make a "technology is destroying everything" argument. I also don't want to moderate my argument just because I don't want to be making that argument. I'd have to give more thought to what argument I want to make and how.
I definitely agree the technologies I chose have done more good than harm. I chose media technologies by accident, but since they go together...they have progressively increased the spread of culture and human connection and that's amazing.
And now I notice that my fear of the harm of AI is that it will undo that...flatten culture and erase human connection.
@sabrina Yeah, and it absolutely is doing harm and it's absolutely a play by the rich and the powerful to control the thoughts of the masses.
There's a great quote from the classic Dr. Who episode The Green Death, where the doctor says something along the lines of,
Humans! Whenever they discover something new, the first thing they do is try to figure out how to kill each other with it, then how to make money off it, and only later do they study it to know whether or not it's safe.
I think about that quote a lot.
@faithisleaping @malcircuit I see the implicit argument now. Sorry, bad wording from a bad mood. I didn't mean to imply the world would be better off without those technologies.
I'm not so sure about AI. Hypothetically if we started from scratch and built it in a good ethical framework it would be fine. I don't think it can be redeemed as-is.
@sabrina I think the fundamental tech has some real and valuable applications. It's a huge jump in parallel compute power.
But LLMs just aren't it. They're cool and flashy but at the end of the day it's just a better markov chain and there's no real intelligence there. And thinking there is is just delusional. And, yeah, they can make pretty pictures, too, but pretty pictures aren't art and we need to stop thinking they are.
But if we used those data centers to study protein folding, maybe we could figure out better vaccines or something.
@malcircuit Sure, it’s the ladder they’re *standing on,* but anyone who’s watched Road Runner cartoons knows that you keep standing in the air after you take away the platform.
This connects directly to the current administration being filled with people who love gangster movies because they never bothered to watch them all the way to the end. (And every other story, or history, that they try to cite in their favor.)
It's #Aislop, it's not a serious study.
It doesn't even measure what it claims to measure.
Do you folks even read these "studies"
Or do you just mash "boost" on every "AI bad" trash?
@[email protected] @[email protected] The Kosmyna "Your Brain on ChatGPT" preprint at least hooked up EEGs and measured actual brains doing actual tasks. Flawed, preliminary, n=54 — but real science trying. Gerlich? STOP. CITING. GERLICH. (It's Business school FFS) The viral "AI makes you dumb" paper doesn't test if AI makes you dumb. It asks people to self-rate their thinking on a Likert scale. That's the whole study. A vibes check dressed up in ANOVA. It claims to use the Halpern Critical Thinking Assessment. It doesn't. Zero HCTA items. That's not a simplification, that's fraud-adjacent. Let me restate IT DOES NOT MEASURE CRITICAL THINKING!!! Forbes ran it. Big Think ran it. PsyPost ran it. Psychology Today ran it. None of them read it. A paper about the death of critical thinking, amplified by the total absence of critical thinking. This is what we deserve. Nebu takes a hatchet to it an leaves a twitching corpse. https://nebu.substack.com/p/highly-cited-ai-erodes-critical-thinking #AI #AISlop #CriticalThinking #PeerReview #psychology #education
So adults get lazy & rusty, while kids never learn…
Mallory's Musings & Mischief
Jake
Huntn00
Aaron Soto
Amin Girasol
Adam Millerchip
Gregly
James Smith 💾
Faith, Purple haired feminist 🧋 
Phie Lux 
St. Chris
Anthony
Jason Bowen 🇺🇦
Epic Null
Cyberlyra
Mother Anarchy
Wulfy—Speaker to the machines
Patrick
F4GRX Sébastien
Rita ⁂
tired
Looking for explanations…
gudenau