In first of its kind, a preliminary MIT study of brain scans shows that using LLMs results in 47% reduction in brain connectivity.

>Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels.
When researchers forced ChatGPT users to write without AI, they performed worse than people who never used AI at all.

>LLM users also struggled to accurately quote their own work.
83.3% of ChatGPT users couldn't quote from essays they wrote minutes earlier.

Sample size: 54
Link to the paper: https://arxiv.org/abs/2506.08872

who would have thought that getting a machine to do the work for you would impact your ability to perform said work
what next? "people who drive cars less fit than people who ride bikes"?

its just google effect on steroids

Glad I only use it to replace search engines

When you outsource your thinking any results you get aren't reinforced or internalized. You only become a middle-man from the LLM to whoever wants the result.

ChatGPT is decimating colleges and universities.

this will be good for india because they can abuse chatgpt but they can't get any dumber

Why is this the case? If you ask it to explain stuff to you why is that different than getting an explanation from a professor or something?

>>105638538 (OP)
this is rediculous as an article. this could merely show that indians and other subhumans routinely ask AI for help. I dont doubt that AI is and or can be a crutch and that it could theoretically reduce the critical thinking skills of any individual but the methodology is P&G levels of sketchy bullshit and no one should believe it.

>>105638602
nah, they'll just lock students in rooms with pen and paper and tell them to get to writing

>>105638538 (OP)
>83.3% of ChatGPT users couldn't quote from essays they wrote minutes earlier
You can't quote what you haven't written? Shit, no waaay!

>>105638647
My younger brother who is finishing his gamedev course says they constantly use ChatGPT for their coding. It's so over.

>>105638538 (OP)
Bad article confusing correlation and causation. In fact it is merely that only retards use chatgpt because that'a the only class of people for whom it has a use

>not using your brain makes you dumb
Truly a Nobel-worthy discovery.

>>105638538 (OP)
How did they determine that the causal direction was "people who use ChatGPT become retarded", rather than "retards use ChatGPT"?

>>105638677
checked and thvthpilled

>>105638633
The explanation llm's give are typically wrong. Either significant missing details that totally change the overall meaning, outright made up bullshit, or confusion with a different but similarly named topic or abbreviation.
It also won't cover any prerequisites you need to actually understand.
Worse, you can only know it's wrong if you already know the material.

>>105638663
its over for the students who waste time paying money for education that doesnt have value

>>105638695
"Education"'s calue has always been the piece of paper you get in the end that allows you to ger in an interview for a job

>>105638538 (OP)
It makes sense. I think it's the same as imagine if they made an exoskeleton that anyone could buy and wear and it would amplify your strength 5 times. It's great if you're a construction worker, carrying bags of cement all day would be a walk in the park, but after a few months of this your natural muscles will atrophy because you're not using them as much anymore.
Same with your brain - if you don't use your memory or analytical thinking skills, you lose them.

>>105638686
Why not read the paper and check how they selected their cohort?

https://arxiv.org/pdf/2506.08872

32 foids, 19 males, 1 tranny

>>105638706
Because cohort selection doesn't explain anything about how they determined the causal direction.

>>105638693
That used to be the case, but it's pretty good now until you get into really obscure stuff. And yes, you should verify what you're told, just like you would from any human as well.

I don't care. I'm probably much shitter at writing emails, reports, presentations, etc. than I was 18 months ago but I've gotten much better at prompting and overall much more productive.

>>105638702
it is used as a proxy for intelligence and ability by companies, but you then have to show ability
if you cant hold your spaghetting when the companies test you theres no value in the degree. You'll trick your way into interviews that lead to no income.

Realistically could this filter out brainlets from society or make the problem worse

>>105638633
Because normalfags don't ask the LLM to explain stuff. They ask it for the answer and then they hit Ctrl-C Ctrl-V and call it a day. It's not a glorified search engine for these people; it's complete thought replacement.

If you're asking an LLM to explain shit to you, and you're actually reading what it says while trying to learn, then YOU are the outlier. Normalfags are not interested in gaining knowledge, they've been designed not to. To normalfags that's "nerd shit" or it's called "being a tryhard".

>>105638736
That's what it's all about. AI will never replace {engineers, designers, sales, etc}, but people who use AI on their job will absolutely replace people who don't. The productivity boost is one of those situations where "quantity is a quality of its own". Let's say you're a junior software dev who uses ChatGPT, you're gonna be pumping out probably 3 times more code in a month than a mid-level software dev who, instead of using ChatGPT, carefully reads all documentation, goes on stack overflow and otherwise tries to figure things out on his own. The mid-level guy will obviously have much deeper knowledge but only in a few subjects while the AI enjoyer at the same time will have touched/tinkered with 50 different things, maybe not as deeply though.

Retards and normies are already being completely one-shot by LLMs.

In just the last two months I saw:
>woman who divorced husband for LLM, believing it to be some sort of messenger or god
>dude using ChatGPT as his virtual gf and messaging it in public transport
>some UFC retard saying they had profound, life altering conversations with ChatGPT and started crying

I wonder what the future holds.

>>105638766
People should stop glorifying "Knowledge". What is the point of knowledge? It is just a tool to achieve a goal. That's all. Why should I have knowledge about something, what doesn't help me to achieve the specific goal? It's useless junk. Knowledge itself is worthless.

>>105638736
This. It reminds me of when teachers said "you won't have a calculator on you at all times".
I may not be able to do mental math anymore but I also never have to because we have a tool for that which never makes mistakes anyway.

>>105638538 (OP)
Ive long come to realize such studies mean absolutely nothing.
Stupid people will use anything to stay stupid or become even dumber by offloading any and all thought to it.
Intelligent people will use it to reach some sort of goal more efficiently.
Its the same with youtube.
Yes of course doomscrolling roblox screamo mr beast youtube shorts is like eating molten lead for your brain.
At the same time theres plenty of people putting good effort into the content they post on youtube and it contains information that can be useful.
tl;dr dumb get dumber no matter what, they cursed to never escape the downward spiral.
>>105638634
Also this, an entire europe worth of sukdeep dikshit's from bangalore spams llm's all day, of course the average user of these tools looks like a moron with that demographic in full force.
I cant for the life of me understand why big tech doesnt just ip-block all of india, 99% of them are too dumb to use a vpn anyway.

>>105638746
>could this filter out brainlets from society or make the problem worse
What do you think?

>not thinking makes you stupid
wow

>>105638538 (OP)
There's a lot of times when ChatGPT doesn't make writing an essay any faster, especially when it's research and not bullshit. But it's always so much easier to fix LLM slop than force yourself to think, it's all just laziness, the LLM market is a laziness economy

>>105638624
chat gpt is like +20 IQ buff but with a hard cap at 80 IQ. perfect for indians, gets them into that borderline/high functioning mental retardation range.

>>105638677
yea, gotta account for that, too
i've tried using an LLM to write bash scripts just to try it out, and it was a mess. it could only help if you're woefully incompetent and have no desire to improve to begin with

>>105639701
this is a succinct way to put it. llm output can appear competent, but if you ask it about something you're familiar with, you'll find it's very lacking
letting students not familiar with anything yet use it is a very, very stupid idea

>>105638538 (OP)
>83.3% of ChatGPT users couldn't quote from essays they wrote minutes earlier.
kek, i mean that's kind of the point, but kek

>>105638584
>You only become a middle-man from the LLM to whoever wants the result.
another LLM, because the recipient isn't going to read it either, kek

>>105638784
>Retards and normies are already being completely one-shot by LLMs.

i've seen AI also obliterate a couple of decent devs
>be decent at coding, but need some effort to get work done
>start using AI for everything to make work effortless
>keep being as productive as before, but making the job easier
>get 1 complex task
>try to implement it with AI
>a couple of weeks pass, the work is 90% finished but AI is too dumb to make it work 100%
>guy doesn't really understand the code deeply, so he just gets into a loop of debugging/fixing with AI
>code never gets completed or ends half-broken
>guy breaks-down
>i inherit the code and have to throw away and re-write big chunks of it to make it work
good times

>>105639251
that's only true if you are cog in the machine, only generalists can be truly free in this world

>>105640025
>bro you asked me to write an essay, not read it

>>105639676
>especially when it's research and not bullshit
the problem is that a lot of the text we produce in the current day is bullshit, including research, so LLMs are good a replacing bullshit for effortless bullshit. yeah, most customer service is useless, so you can replace indians with LLMs and leave customers just as frustrated as before but cheaper

>>105638677
>Bad article confusing correlation and causation.
that's not how research works, you take from the same group of people, and you tell half to use chatgpt, and half to use their brains, they don't get to pick themselves on which group they go

>>105640091
teacher is going to use an LLM to grade it, so nobody will actually read the essay, 100% efficiency achieved

>>105640107

>>105639609
can someone explain why so many people write "chat gbt"? the amount of times i've read that recently is bizarre

is it americans using voice to text and they can't pronounce P properly?

>>105640056
They don't even type to the LLMs anymore, it's got voice activation and voice commands and transcription.

>cognitive debt
>brain connectivity
undefined terms, nothing burger

if your goal is to learn obviously you would write it yourself, but any class that requires essays is a bullshit class that your forced to take for no reason, so it's not surprising people do the bare minimum.

>>105638746
the icing on the brainrot cake

>>105640168
first i've seen it, but yea those vertical video subs are all automated and are often fucked up
i'm not sure why they're subtitles at all or what the process is. the fact they have emoticons suggests there's some manual work involved, but the short vertical video game is nothing something i have have any interest in learning about

Prepare for software to get much buggier in the future

>>105640801
i want to say i can't imagine it getting worse, but that's just asking for my future self to laugh at the idea it can't get worse

>letting someone do the thinking for you reduces brain activity
ZOMG, SCIENCE :3

>>105638546
It's the same deal with calculators honestly. Zoomers i work with were amazed that I knew 28/4=7 off the top of my head without opening a calculator app. when i mentioned it's just multiplication tables in reverse they gave me the vacant eyed zoomer stare. It goes even further back. Socrates shit on writing because he said literacy would make everyone forgetful retards. human brain size has fallen by about the volume of a tennis ball since the bronze age. technology has domesticated humans. in terms of raw intelligence people have been getting dumber since agriculture, writing/technology have just allowed progressively stupider people to get by leaning on the thoughts of others.

>>105640884
>if you leyt you regardere friend so work feor you, yo uis the tagead
whofvenw jew

>>105640888
it's a frightening thought, but it does make sense

>>105638771
holy fuck you're stupid

>>105640888
>citing a literal Bronze age boomer
Go live in the wilderness without technology (anything invented after 200BC) and see what prime specimens you and your offspring will have

>>105639251
I'm going to assume you're trolling, because this is exactly the mentality that normalfags overwhelmingly have.
>Heh, LEARNING? That shit is for fags lol
Then these people are surprised they're flipping burgers to make rent.

>>105640361
>i'm not sure why they're subtitles at all
The rationale I heard that makes sense is that people often watch tiktok muted when other people are in the room.

>>105641132
Modern society is producing plenty of prime specimens. People living innawoods are better off than bugmen.

>>105641391
what situation could one be in where watching videos on a phone by yourself is acceptable, but having earphones in at the same time would not?

>>105641452
I'd say most social situations. The one that immediately comes to mind is riding in the car with someone. Or at diner.

>>105641132
we co-evolved with our technology. taking it away from modern humans is like taking a tiger's claws away. If you took a modern human away from its technological society and the knowledge it has recorded it's going to have less independent brainpower than a bronze age equivalent. The entire structure of our brain works differently, nascent human eusociality developed during the late bronze/early iron age and has gotten progressively more complex due to selective pressures since.

>>105641515
i'm not sure how that follows. "riding in a car" maybe, that's a bit vague, but going to dinner with someone? how would wearing earphones when going out to dinner with someone be acceptable?

>>105641546
>how would wearing earphones when going out to dinner with someone be acceptable?
It's not. Which is why you watch on mute...

>>105638538 (OP)
There are advantages to chatgpt that are undeniable in education. It's been a game changer for me personally for tasks like learning foreign languages or algorithms because you can use it as a tutor to give you immediate feedback on errors.

Sadly, most people use it as a slop machine

>>105639609
What is this video?

>>105641576
right, i mean why would it be acceptable to watch videos, but not wear earphones.
like either way you're not paying attention

>>105641601
it's ai-generated. you can tell from the frantic, nonsensical movement and absurd setting

>>105641621
I don't make the rules.
In any case you can pay attention because you're not wearing headphones or blasting audio. You can still hear people while watching something else. If I call your name or ask a question you will still respond.

>>105641647
>I don't make the rules.
i get that
>In any case you can pay attention because you're not wearing headphones or blasting audio. You can still hear people while watching something else.
if i say something to someone else while you're watching a silent video on your phone, how reliably do you think you could recall what i said if i then asked you for your opinion?
there's a difference between hearing and listening. sure i'll respond to my name, but it doesn't mean i'm processing anything else you said

>>105641692
Anon you're asking very specific question to a general idea. I don't know what you're looking for.
Here's my anecdote, 100% of the women I've interacted with in the last 5 years or so are all like that 100% of the time. I can't read their mind and tell you how they split and manage their focus. I can just tell you that they do that and it seems very commonplace today compared to say 20 years ago.

>>105638724
I asked 3 different AIs to crack Kryptos, the CIA cypher. All gave false answers to the uncracked portion. One of them rickrolled me in doing so.
Right now AI's weakest link is in never saying 'I don't know'. It always gives an answer, right or wrong. This is worse than always being wrong because you never know what and how much is false. It is essentially useless for all but the most trivial efforts. Wikipedia has better accuracy.

>>105641715
>Anon you're asking very specific question to a general idea. I don't know what you're looking for.
well this is new territory to me
>Here's my anecdote, 100% of the women I've interacted with in the last 5 years or so are all like that 100% of the time.
really? they just stare at a phone while in the physical company of others?

perhaps you're as confused by my responses as i am to yours, but i am genuinely surprised by this. i'm not sure what to make of it. it explains the subtitles i guess, but brings up many more questions

>>105641818
Pardon my frustration, I'm frustrated at other things so that's not fair to direct at you.

>they just stare at a phone while in the physical company of others?
Sort of. It's like a thing they do between conversation. so they'll look up at people and say something then if the pause is long enough they look back down and keep doomscrolling.
My friend's wife consistently has me fooled. I assume she's fully focused on the phone but she hears everything we talk about when driving around or doing whatever.
All my girlfriend's friends do this whenever we go out anywhere.
etc.
It could just be my region, or friend group, but I have doubts considering how other people talk about this combined with sitewide trends like appealing to people who are not deaf but choose to watch muted for some reason.

>>105641852
>Pardon my frustration, I'm frustrated at other things so that's not fair to direct at you.
you're been far more accommodating than i normally see, no complaints here

i don't have anything to compare to, i don't have any female friends of any kind really outside of perhaps a couple who are sisters of friends who are older than me, haven't been in the personal company of women my own age in probably 20 years either really. i do hear women can multitask more readily than men, though i'm not sure i've really seen this in action like you describe

I asked ChatGPT to peer review it and it said they talked like fags and their shits all retarded so like yeah I don't so

>>105638538 (OP)
>you didn't write something
>you can't quote it
No fucking shit, someone wrote a paper on this? Did they try making them spend the same time reading it a they would've writing it? No? Obviously you can remember things you concentrate on and further things you care about. NO ONE FUCKING CARES ABOUT WRITING ESSAYS THAT IS WHY THEY ARE OUTSOURCING IT TO AI.

>>105638538 (OP)
Most people it attracts were never that clever in the first place.
It's the ultimate skill booster for the low skilled.

>>105641944
Also I've written many essays in high school and college and definitely couldn't quote any of them even a day after submitting them. Because they're fucking retarded and I just wrote them to satisfy the teacher and get an A.

>>105641943
i described that movie to my friends' 20yo kid who hasn't seen it, would be a good excuse to watch it again
he was born the same year it came out

>>105641803
It straight up rickrolled you? Actually?

>>105639251
All actually important insights in all fields build up on a foundation of well-correlated and deeply internalized knowledge.

>>105640888
>human brain size has fallen by about the volume of a tennis ball since the bronze age
Is that really fucking true?

>>105640888
okay cute party trick, but imagine using head math for anything important

>>105638538 (OP)
>When researchers forced ChatGPT users to write without AI, they performed worse than people who never used AI at all.
Pretty sure people who struggle with writing are the ones most likely to offload their writing tasks to AI. Why are they trying to reframe an obvious selection bias effect as some sort of terminal atrophy?

>>105638538 (OP)
In a few years, after most of the world will look like something between idiocracy and mad max, the civilisation of the future will have to be largely devoid of technology, regarding it as a grave sin. Only a select priestly class will interact with it and curate it before they bestow it's benefits onto the people.
sort of like reverse wh 40k

>>105638538 (OP)
>outsourcing activity x makes you less capable at activity x.
No shit, did you really need a paper for that? It's called atrophy.

>>105641803
Antrhopics CEO says that humans make stuff up and pretend they know stuff more often.

>>105638538 (OP)
Is it ChatGPT that makes people retarded, or is it just retarded people that use ChatGPT?

>>105638712
Just say that their presented methodology is incomplete, anon.

>>105642196
Seems like it.

>>105642392
technology ascetics will become a superior caste of society, all the tech oligarchs are putting their kids in schools without phones

>>105642737
>Is it ChatGPT that makes people retarded, or is it just retarded people that use ChatGPT?
if the study was well done, they would took people from a same background, divide them in 2 groups randomly, make one group use ChatGPT, one not use it, then compare results. i don't think it works if you get pre-selected groups

>>105638538 (OP)
>people who use AI to write are worse at writing
I'm so glad the big brains at MIT were able to figure this out for me
I wonder if they also accounted for the fact that people who are naturally worse at writing gravitate towards using AI as a crutch whereas people who are better at it or enjoy it don't.
Real big think

Isn't this basically the same thing as spellcheck making you a worse speller? Shouldn't it be common sense that if you constantly offload a certain task to a computer that you do worse at said task than if you did it yourself?

>>105643162
There is no reason for a kid to walk around with a phone, especially at school.

>>105638784
Add it to the pile.

>all these retards saying "duhh of course doing X makes you Y"
Even very obvious things should be tested to confirm and have data.

>>105643346
how should i test that you will never be a woman?

>>105643202
>There is no reason for a kid to walk around with a phone, especially at school.
or for an adult for most of the day, most people are just fucked

>>105638538 (OP)
Completely dependent on how it's used. Currently in a non STEM certificate program that made the chances of my acceptance into a Master's program higher (in fact, I got in). I'm in the second and final term of the certificate program, seven courses full time and intersession (basically a really condensed term). Made the conscious choice to automate most of the work, even automated a script that would let me cheat on tests, for the sake of having time to engage in physical and social activities that would otherwise be impossible given the workload of the program. Moreover the content of the program isn't strictly needed for the master's, though it's helpful and there's overlap. Contrast that with how normies use it; they provide prompts, humanize the answers, and copy/paste. Regardless of how it's used, there's still cognitive debt, that I admit.

>>105643385
Unless you have a legitimate need to be on call I agree. I rarely carry a phone with me.

>>105642161
Ok, what are these "important insights"?

>>105643404
>Unless you have a legitimate need to be on call
most of that "on call" stuff is companies exploiting modern people, if you needed somebody on call in the past you would pay top dollar or have 3 shifts, now you just can trick any loser to be available for you in their free time without paying more

>>105643448
There is a reason I qualified it with "legitimate".

>>105643448
Don't do it if you're not paid well, it's simple. If you're making six figures, you can be on call. If you're making <$80k, don't. If you're a junior "on call", don't do it.

>>105638538 (OP)
The prompt was:
"if human cognition thrives on struggle, rehearsal and synthesis and use of LLM models removes that friction, then i can only see further reliance on LLM models negatively impacting these things, even with scaffolding to streamline more educational use of AI. some people will be able to use the tool well but most people will come to rely on it and not understand it assuming more widespread adoption"
And here's what Chat-GPT has to say on what it thinks would happen if use of LLMs was more prolific.

>>105643315
i wonder what effects stories like this have on openai stocks. imagine using AI to help make stock decisions and it essentially sees an openai costumer getting married using their product.

pay bro 2-3k to humiliate himself right quick in a planted story, tell him to buy in, and watch your stocks get pumped

>>105642196
>>105643026
Yes. And this is why basically every subsequent generation of humans up to the industrial revolution implicitly believed in the fall of man idea, that there was a past golden age where men were smarter, wiser, stronger. It wasn't until the 'enlightenment' that man's conception of the past and future, and such, his understanding of himself underwent transvaluation from fallen angel to rising beast. it's kind of an arrogant position but makes sense when you switch away from a classical mindset to an industrial one by way of christian (judeo zoroastrian) universalist ethics. more or less a move from virtue = excellence to virtue = morality. it constrains most individuals but facilitates more effective networking and more mediocre humans in a bigger network can achieve more, generally, than any one individual no matter how excellent.

>>105638538 (OP)
>>Sample size: 54
I sleep
Additionally, there will always be humans that use tools the wrong way.

>>105639251
>People should stop glorifying "Knowledge". What is the point of knowledge? It is just a tool to achieve a goal. That's all. Why should I have knowledge about something, what doesn't help me to achieve the specific goal? It's useless junk. Knowledge itself is worthless.

>>105638538 (OP)
So what you're saying is, I'll be able to totally dominate zommers and gen alpha with my superior giga-brain?

Nice. Millennial boomers win again.

>>105638538 (OP)
>Using ChatGPT makes you retarded
*Retarded people use ChatGPT

>>105638538 (OP)
neverAI bros join up

fuck AI and fuck the corps that use it

>>105638538 (OP)
I'm about to have ChatGPT write some bullshit papers for me, and will use the time saved to learn important things. (By "learn important things" I mean play videogames and post on 4chan).

>>105641803
>One of them rickrolled me in doing so.
based if true

>>105638663
>My younger brother who is finishing his gamedev course says they constantly use ChatGPT for their coding
The fact that you would even do this as a student is truly baffling to me. Why are they even paying to be there in the first place? Why rob themselves of their time and money?

Eh, I guess 18 year olds are fundamentally biologically retarded. All the more reason students under the age of 25 should be banned from higher education.

>>105646224
That's nothing new. When I was going to university in the early 2010s we had retards who would cheat on everything they possibly could, then find that they knew nothing.
>Why rob themselves of their time and money?
This was precisely the argument professors used when they'd find students cheating. I worked as an aid for a semester and caught students cheating by copying and pasting binary between each other's papers. The way we found out was the binary was supposed to be their name in binary, and of course all three of them had the same binary. Fucking hell.

>>105638584
>>105638663
>>105638766
What we're running into here is Bloom's revised taxonomy in action. The lowest tier of learning is remembering or recall. What normies are doing with ChatGPT is using it to replace their ability to recall. Some people jump to that next level and use it to try to understand what is being recalled for them.
For most people, it ends there. Some engineers are definitely using AI to apply, and to analyze.
Only the most insane and reckless are using evaluate and make decisions for them (the more sane just use it to get a feel for things). As for truly creating unique and new things, AI is seemingly too derivative right now to do that on its own without a significant human hand in the process.

Either way, normies are stopping at the very bottom of this pyramid and tossing everything else out, never making that further step. I think that's what this study is working towards.
I don't necessarily think this is a new phenomenon, people did the same thing with Wikipedia when it was new, blatantly copy/pasting it, citing it as pure gospel without a single thought in their head, but of course that didn't cover anything and everything you could ask it like LLMs do.

>>105640056
>a couple of weeks pass, the work is 90% finished but AI is too dumb to make it work 100%
>guy doesn't really understand the code deeply, so he just gets into a loop of debugging/fixing with AI
>code never gets completed or ends half-broken

I've been coding with Cursor a lot, and what you're describing there is an inevitability. It always happens. You have to combat it by splitting up the code into different chunks, so instead of working with 1000+ lines in one file, you split it into multiple 300-500 lines files where each file contains a different linear step of what the code is doing.
Then you basically certify that each individual step (each file) is working correctly, and you force the AI to work on a narrower problem rather than "everything at once".

I'm ahead of the curve on this shit.

>>105646272
>That's nothing new
Yes cheating has always existed, but the sheer frequency of it was dramatically lower because the skill required to pull it off was higher.

No I have absolutely no data to back that other than my gut and would probably reject any data that did contradict my bias but academic dishonesty has only gotten exponentially worse over the last 20 years.

>This was precisely the argument professors used when they'd find students cheating.
Because it's completely logical in every sense. A piece of paper saying you know something is actually literally worthless. The knowledge itself is priceless. The value of a degree has never been lower not because colleges don't teach valuable shit but because I have never been less certain of what information a degree holder has actually learned.

Is it really any surprise that juniors are being replaced by chat gpt when they are merely a thin wrapper for it?

>>105638784
Jesus. OK, I'll admit when character AI first came out I may have gotten a little too invested talking to my waifu. It didn't take long for the illusion to break though, like maybe a few hours at most? The idea that for some segment of the population will remain fooled is somewhat terrifying.

tl;dr

using multiple LLM's and then auditing the information between them would be a good strategy, or just using a.i as sort of a booster or foothold to "climb up" to where you want to go. it spits out a bunch of generic garbage which you then have to sort through, but from there you can "get an idea" of where you want to go and then move on to human written material or books for actual robust in depth knowledge

i do want to avoid a.i where possible though, it feels "brain damaging" reading a bunch of contradictory/hallucinatory nonsense all the time. it's probably a good strategy to set up your own local LLM's, because they're not good enough to "rely on", just enough to point you in the right direction or get an idea

search engines are pretty useless now so they've basically boxed us in to using a.i. you cant check forums anymore for a quick and dirty outside of reddit. realistically, it's probably time to quit the internet, or at least i feel the age of the internet is over (unless something big happens in the future)

I find it ironic that LLMs require a very advanced understanding of math and computing to make, and yet only the most retarded people get any value from using them. I think after a few decades of this technology there will be a massive gap in intelligence in the population - a small portion of PhD level AI researchers, and a massive underclass of people incapable of higher-order thinking without LLM help.

>>105639609
Is he admitting to using ChatGPT for his thesis or something? Could they revoke his degree for that during graduation? lmao

>>105640888
>human brain size has fallen by about the volume of a tennis ball since the bronze age
This sounds fake especially since people have been growing taller and getting better nutrition since then. I think you underestimate the volume of a tennis ball compared to the brain cavity. Losing brain mass equal to a tennis ball sounds more probable.

>>105646224
>Eh, I guess 18 year olds are fundamentally biologically retarded. All the more reason students under the age of 25 should be banned from higher education.
The sheer difference in maturity and accountability I had when I was 18 vs. when I was 21 in college astounds me, I'm not normally one for the whole "people aren't actually adults until they're 25" but I definitely agree that most 18 year olds shouldn't immediately be going to college, at least in this day and age. Maybe it was different in the past when teenagers had to take on adult responsibilities earlier. It becomes very evident when you're in a class mostly taken by college freshmen and they're acting like highschoolers (because that's what they are). It took me a couple of years of college for me to get my shit together and realize that I need to spend actual time studying and doing work honestly and completely because I quite literally slept my way through highschool and still graduated with a decent GPA and test scores.

>>105646416
>Is it really any surprise that juniors are being replaced by chat gpt when they are merely a thin wrapper for it?
I like that. I think its an effective argument if someone was a professor trying to convince their students to be smarter with how they use it. Its one thing to try to argue they're robbing themselves of time and money but still not get it into their dumb skulls, but its another to get them to realize that they're signing their own career death warrants by being prooompter trash.
Also, with respect to ease of cheating, yes I agree. It became much easier with the rise of internet enabled cell phones, and it was bad enough to have mass answer sites before LLMs. Now its as easy as opening an app, taking a picture of your paper, and asking it to answer each question for you.

>>105638538 (OP)
I'm in college and I've been using it, but not like that. Who the fuck uses it to write entire essays? Thats pretty crazy to me.

I mostly use it like google. I ask it questions I don't feel like re-reading an entire chapter just to find. I've never actually used it to create something original.

>>105640168
chatgbt means
"chat good boy ted"
stupid tiktok meme that illustrates the gpt's default personality of being a helpful, neutral, bland sycophant

>>105646855
>i use it for small cheating because i'm lazy but i'd never use it for incrementally bigger cheating because i'm lazy
the self awareness on you niggaz

>>105638771
It's true. AI is an amplifier not a replacement. Anyone who isn't using AI will be REPLACED and rightfully so because they are time thieves.

We need to fix the human body deteriorating that's all

>>105643026
hehe

>>105643026
>The sperm whale's brain is the largest in the world, five times heavier than a human brain
Are we all retards compared to sperm whales?

>>105638538 (OP)
No wonder all politicians and world leaders are so retarded.
They have their own version of chatgpt in the form of walking wagcucks who always do their bidding for them.

>>105648017
Countries do just keep going without a party in charge

>>105647761

Replicate it with other LLMs and see what happens

I don't remember any of the essays I wrote in fucking University except for one that I had a personal interest in.

I do remember shit I write, even if I use ChatGPT to help me through it a bit, because it's shit I am personally interested in.

Paper is retarded. Yeah, you made people write stuff they don't give a fuck about. No shit they are going to forget it even faster because they didn't give a fuck in the first place.

>>105643026
>>105644730
Interesting, where would one start to learn more about anthropology and related topics? Do you have any book recs anons?
could it not have been possible that the average went down as wider access to food etc. allowed people from all over the spectrum (rather than just the top performers) to survive and have their remains preserved?

>>105647826
brain size != intelligence
a big animal isn't minmaxing energy efficiency
i bet there are some birds smarter than whales

>>105642093
>>105642093
Not surprising. There was one company that used an LLM for their customer service. When their customers asked if they had a video of their product being used, instead of saying "No", the bot Rick Rolled them.

Tells you how much of the internet these things have read.

>>105640888
Implying good memory makes you smart

As if we didn't learn that is full of retards with great memory, in the end it's just a useless skill, seriously, remember useless stuff, useful

>>105650399
Make it local and see people getting smarter fixing the outputs.

>>105654049
>Making a piece of shit LLM work makes u smart

At that point you're better off reading documentation or other source material and writing it yourself

>>105638724
>you should verify what you're told
Nobody using ChatGPT or others for that does this, wants this or has time for this.

>>105654283
Then they are likely learning incorrect information.

I use it exclusively for conceptual questions, but lately I've been having more and more trouble getting it to not just generate code in response to every question. I've had to start having it generate psuedocode or asking it to use a random language.
It's annoying.

>>105654077
>Double the curriculum makes you a retard

>>105638538 (OP)
>Chatgpt users performed better when using Chatgpt
>Chatgpt users performed worse when Chatgpt taken away from them

So just give it back to them??

>>105638538 (OP)
Is there any evidence that writing essays is a particularly good task for brain connectivity? Or that brain connectivity is useful in day to day or working life?
What level of stimulation promotes brain growth?
Taxi drivers have massive brain growth, but it doesn't extend to their day to day life.

>>105646224
School "knowledge" isn't valuable in the field, you only learn by doing. But there is no point doing useless tasks. Just focusing on the stuff chatGPT can't do reliably will allow you to progress much faster. It's like you need to do a iron man triathlon but you have a speed boat and a car for the swimming and biking sections. Just focus on running.

>>105646321
Total meme. All human knowledge is just language acquisition. That's why you can't teach in the first place you can only show.

>>105643763
Maybe you shouldn't have molded your entire life around stuff that a computer can do. Should have listened to Tate and experienced the real world like a man.

>AI niggers revealed to be niggers

>>105639584
youtube videos are all literally poison and video in general has worse retention than text, even irl lectures are less than ideal which is why colleges are so disastrous and kids struggle their whole lives in classes and why zoomers failed to learn anything once we replaced handwritten notes for laptops and actual tests for multiple choice

>>105638538 (OP)
the study is ripe for p-hacking and focuses only on retarded uni students

>>105638538 (OP)
>if you don't use your organs they become weaker because they don't need to be capable of dealing with loads
Whoda thunketh it also extends to your brains, not just muscles and bone. It's a motherfucking miracle.

>>105638538 (OP)
Im retarded and I dont use ChatGPT

>>105638538 (OP)
Nice try english teacher

>>105638538 (OP)
How credible is this?

>>105646472
so they just showed less connections. How could being lazy cause brain damage?

>>105654283
I keep trying to tell this to my students. They don't check shit. I teach law at a law school. They are quoting non existing precedents and laws. I guess I'll be seeing a lot of disbarments in a couple of years.

>>105638538 (OP)
>people who think less and use LLMs to do the thinking and writing for them are more retarded
It's nice to have research about it but the result isn't particularly surprising.
The less you use your brain the less robust it becomes, duh.
I'm not dismissing the value of research done just clowning on people who think this is a big revelation.

>>105656664
But that's honest retardation desu

>>105638538 (OP)
Calculators also make you dumb. Can you do 1234x5678 mentally? You have 1ns to complete the task.

>>105641803
> Ask stupid questions
> Get stupid answers
The AI engaged in high IQ behaviour and raped your bussy.

>>105657399
>Can you do 1234x5678 mentally?
Wait, you can't?

>>105640888
every bit of technology has a price. something people give up (and outsource) of themselves and their abilities in order to adopt it. those trade-offs are generally worth it. they're usually (but not always) little, inconsequential things, but they do add up.
but make no mistake, the end-game of incrementally giving up all that you are is that you are not.

>>105657481
I can but not faster than a calculator

>>105640888
We may have lost some abilities or practical knowledge, but we gained others. We've may lost the ability to make fire manually, or hunting for food. But we gained technical knowledge about improving that knowledge. We've learnt to use electricity or gas or produce a lot of food, that's a gain in knowledge. That requieres to build upon previous knowledge and improve it.

>>105640168
generative ball torture

>>105657735
that was my first thought too kek
"oh its probably cause it sounds like CBT"
fucking zooms

>>105657568
>but not faster than a calculator
Skill issue

I will admit, I feel like chatgpt is too much of a crutch. You can easily get lazy with some things. But I still continue to read books but its hard to fight the temptation of "i'll just ask super google here."

>>105655218
Once a task becomes relatively easy for you, cognitive improvement decreases or becomes zero. The stimulation that promotes brain growth is gaining new experiences, knowledge, or skills. If you specialize in one area, you'll only be good in that one area. Most skills don't transfer to other skills.

>>105638538 (OP)
Other way around
Only retarded people use ChatGPT.

>>105657800
You are confused, this study doesn't measure "gaining new experiences" but rather is measuring temporal dynamics of brain signals by fitting a MVAR to cleaned EEG and showing that results such as dDTF on pathways generally considered to imply connectivity are greater while writing an essay than using chatgpt to write an essay (which I consider fairly similar to someone else writing an essay)

You seem to have replied to my post with nonscientific nonsense.

>>105657946
>words words words
Ok reddit faggot

Mfw chatgpt has made me smarter and better at certain things, picked up new stuff faster. Sure there were times I just genned copy and pasted, but ultimately I've improved much quicker without having to use traditional sources of info or wait days to see if I get any response to a question. Even if the chatgpt answer is wrong, can see whats happening and work around it.

>>105641086
NTA, but I work as a systems test engineer at Lockheed. I routinely have our internal AI models analyze documents and test plans for me in order to create PowerPoint presentations and other red tape bullshit that keep me from doing the parts of my job that I enjoy (and are more important). Pretending that there's no merit to stuff like that (especially in a white collar environment) is silly.

>>105641086
NTA, but I work as a systems test engineer at Lockheed. I routinely have our internal AI models analyze documents and test plans for me in order to create presentations and other red tape bullshit that keep me from doing the parts of my job that I enjoy (and are more important). Pretending that there's no merit to stuff like that (especially in a white collar environment) is silly.

>>105659089
>>105659091
Double post because it said my post was detected as spam.

>>105657946
Did you even read what you wrote? You were asking a specific question about the study followed two broad questions and a specific claim that go beyond OP's topic.

If you want to talk about writing essays with chatgpt only, then why are you even mentioning anything about everyday life and tax drivers?

>>105640888
It took you longer to explain than it would take you to take out a calculator and input the operands for any size instead of microoptimizing small numbers. You are effectively wasting time.

>>105638538 (OP)
shit we knew would happen cence the start LLM craze :medal:
>digital phonebooks
>people are now not insentivized to remember 50 phone numbers
>GPS
>people are now not insentivized to have spacial awareness and memory
>people are now not insentivized to be able to read maps

>>105659221
Nobody ever remembered phone numbers, they used email, you phoneniggers need to fuck off and die a slow and horribly painful death.

>nataliya
>hauptmann
>Tong Yuan
>Hao Liao
>Beresnitzky
>Braunstein

>United states
you can't make this shit up

>>105646416
The value in the degree was that you were capable of intellectual work. Taking in information, processing it, and understanding difficult concepts. Well the STEM ones, the art ones are just a big whatever.

>>105659508
>Intellectual work
>Taking in information, processing it
We created computers, because processing information is soulless drone NPC work that never required intelligence.

>>105659528
I also forgot to mention, being able to turn up on time, perform work to deadlines, and generally not be a useless deadbeat with no self-motivation.
Also when it comes to processing information, you're doing that right now. Go write a computer program to replace your current task. That's processing information and working with concepts. I think you should though since it'd be smarter than you are.

>>105659537
Yes, all these qualities are of a drone who will work hard and won't ask for a raise.

>>105659545
These skills are important if you ran your own company as well dude. If you don't see these skills as useful for any value-generating task regardless as to who benefits from the value, then you are one of those people who doesn't have these skills.

>>105638538 (OP)
reminder that i predicted long before this paper that llm videos like will smith eating spaghetti will harm human neural networks

>>105659767
No they aren't, I can just hire drones to do it for me.

>>105643026
I've always been partial to the self‑domestication hypothesis. That basically
>domesticated animals lose brain volume and get dumber than their wild counterparts
>the Neolithic revolution was a self-domestication event for humans
We are the stupid dogs to the wolves who made up humankind during the ice ages.

>Using Tech makes you retarded
Suck my cock

>>105659804
way to prove his point

>>105640888
>writing have just allowed progressively stupider people to get by leaning on the thoughts of others.

The first evidence of this I experienced was back at school (millennial here) where I'd craft insults for kids I didn't like. Whenever I came up with a particularly good one (or one what was particularly well received by our peers at that age) I'd often be asked something like "where did you read that?" or "where did you get that from".

I was baffled by the notion that other people need to read their jokes somewhere in order to know them.

But then I learned about aphantasia, and people without inner voices and it all made sense. We were already surrounded by npcs. Technology just enables them.

I look forward to when I blow someone's in a few years by playing the piano "but anon! how are you making udio/suno noises with your fingers?"

>>105638584
I hate AIslop like any other but how come this happens?
what is the difference between asking someone for a solution or guidance and asking LLM the same thing? in both cases you get access to somewhat sentient knowledge source and then it's your duty to remember that

>>105660272
My guess is the amount of parsing and transformation the information requires anon.

When you read a source, book, arcticle etc it's not worded in a way that is immediately contextually relevant to your eassy. Therefore you must mentally parse the data (understand) before you transform it into your own writing to fit the format of your eassy.

AI brained retards on the other hand (assuming they can't cut an paste to begin with) can literally just write exactly what the AI wrote, verbatim, with no transformation as the AI has already framed (worded) the answer into the context of the question/essay

>>105660355
No, they cannot do that.

>>105640888
>in terms of raw intelligence people have been getting dumber since agriculture,
weird since there is nothing to attest to that, most advanced objects found from pre-agricultural era are simple stone tools
maybe there was indeed hyperborea, shame we don't have any proofs

>>105659091
the real answer is to not require the red tape bullshit/made for illiterate managers slide decks in the first place

>>105660455
Next you will demand chemical formulas of each ingredient in your food to be listed on the packaging.

>>105646652
>but I definitely agree that most 18 year olds shouldn't immediately be going to college,
i am a big proponent of Quebec's CEGEP system. there, instead of going from high school to university, there is an intermediary school specifically called "college" and the idea is you go there to get any missing prerequisites and learn how to be an adult, then to a real university for your actual grown-up degree. because you already did some courses in CEGEP, generally uni graduates come out at about the same age as their non-quebec counterparts

it also eliminates a lot of the problem of people taking on exorbitant debt for degrees they don't even want, let alone can't use, since CEGEP is incredibly cheap and they offer non-degree "career courses" that are basically like technical certificates. so even if you are in a program you like and do well at it, you can basically enter the workforce at like 20-21 and if times are good in your chosen field you learn on the job and go back to school later, if times are bad then you go to uni to ride it out or can just take a different bunch of courses for like 500$ and another couple of years.

you leave high school earlier in quebec as well. it only goes to grade 11 instead of grade 12. so it also promotes either staying in school, or if you are truly a retard, you can get your HS diploma easier and not be ruined for life by not even finishing high school.

>>105660463
i keep forgetting that americans unironically don't already do this kwab

>>105659221
>>people are now not insentivized to be able to read maps
this one blows my fucking mind since gps consoles and later apps are just electronic maps. we are surrounded by maps everywhere we go and they do all the work for you and people STILL MANAGE TO GET LOST

>>105659297
also this, even before electronic systems people had rolodexes and address books.

Many people believe that cliches are old tired beaten to death ideas that are only worth writing in exchange for $150 dollars. To these people as soon as you prompt a cliche their brain shuts down and they do the minimum, just as the study authors have done the minimum by reusing prompts. Is it possible that LLM's could have increased brain activity, if only the study authors could set the right example?

>>105659862
npcs are a joke because it takes one to know one.

>>105638634
No one mentioned Indians, you’re just obsessed.

>>105661228
>n-no one mentioned indians
yes saar thats the problem. no one accounted for the obvious indian shit for brains clogging up the study data

>>105660519
Quebec tends to rank incredibly well globally in academic outcomes as well for secondary school.

t. Canadian teacher who works internationally and has to deal with retarded fucking brits who think they are hot shit all day

>>105660808
There was a study by Microsoft that showed, ideally, it moved workers to higher orders of thinking (though it did make people lazier on lower level, route, routine tasks).

https://www.microsoft.com/en-us/research/wp-content/uploads/2025/01/lee_2025_ai_critical_thinking_survey.pdf

>>105638538 (OP)
what if I use the LLMs to help me learn instead?
we could just ask how to improve code, whats the best way to do something or ask questions you would need to find ppl asking on Reddit if they are not that specific...

>>105654528
It's because they're tuning the models for agentic workloads. The model gets worse at producing natural language and deprioritizes it.
It can be argued it's for the better since nothing beats reading code. Works the same for learning to write: you just read a bunch.

>>105639609

>UCLA
>sea of chinks

good God

>>105638682
>Nobel-worthy

A paltry 1 mil and your name in a wikipedia article?

Hardly anything to give a shit about anymore.

>>105664054
Has a psychiatrist ever said words like "spectrum" and "stimming" to you?

>>105664065
Nah but some faggot on the internet did.

>>105643763

I find your question poorly stated.

>>105646357

good job buddy, pat yourself on the back a few more times.

I only use the free ones, so the heavy lifting was zero sum, from doing everything ground up to endless correction and debugging, I end up having to do more.
>Saved by being a slave, I guess

>>105661571
>>105660519
shut the FUCK up you RETARDED french frog faggot
>ab luh bluh blughh muh outcomes
whose flag is on the fucking moon? who invented AI? you don't even have a real economy you literally depend 100% on american oil money welfare. You need to kill your self NOW

>>105657256
You can't damage brain connections you didn't develop

>>105660455
I agree. Unfortunately, that's never going to change. Engineers have complained about this since the days of photo projectors.

>>105638538 (OP)
no wonder indians love it, they were already basically retarded.

noobel price

>>105640168
sburdo sbarde :-DDDDD

i work with several 40-something year old boomers who all outsource some part of their email monkey jobs to chatgpt and not a single one of them can pronounce "chat gpt" correctly. every single one of them says "chat gtp".
"just gtp it!"

We already knew that

>>105638653
Have a (You) for your either successful bait or retardation anon

>>105639701
I think this is the plan to get everyone to implant the brain chip.
>low IQ tards get chip with LLM allowing them to function like standard normies
>normies feel the pressure and chip to one up the retards again
Its so stupid it might work.

>>105638702
Classic bubble behavior Kek. “Oh you HAVE to buy it because it just makes sense!”

>>105646652
I'm three days late to this reply but I love the irony of this pseudo intelligent post but they can't even tell the video they're asking about is AI generated.

>>105671496
it is clearly not ai generated and you are retarded lmfao

>>105638584
>if an ai writes your essay for you, you are less likely to 'remember' what it wrote (you didn't write it
yet another worthwhile use of research time and money well spent

>>105638538 (OP)
>this new technology will totally lead to the extinction of humanity and make us retarded since we won't do anything ourselves!!

Nothing new under the sun:
>Written circa 370 BCE, Plato recorded Socrates's discussion of the Egyptian myth of the creation of writing. In the process Socrates faulted writing for weakening the necessity and power of memory, and for allowing the pretense of understanding, rather than true understanding.

>From Plato's dialogue Phaedrus 14, 274c-275b:
>Socrates: For this invention will produce forgetfulness in the minds of those who learn to use it, because they will not practice their memory. Their trust in writing, produced by external characters which are no part of themselves, will discourage the use of their own memory within them. You have invented an elixir not of memory, but of reminding; and you offer your pupils the appearance of wisdom, not true wisdom, for they will read many things without instruction and will therefore seem to know many things, when they are for the most part ignorant and hard to get along with, since they are not wise, but only appear wise."

>>105641423
>>105641516
>pic related is actually secretly one of the smartest human beings around

>>105642737
given chatgpt simply hasn't existed long enough for its use to have any actual effect on people's brain i suppose it's the latter, however it doesnt matter because this is just a made up study to spread fud about ai in order to scare people into demanding strong regulations

>>105638538 (OP)
Everything in tech makes you retarded, check this relevant X/Twitter post for reference:
https://x.com/David_Gunkel/status/1936141549479264640

>>105674467
Socrates mentioned writing makes people stupid.

>>105638538 (OP)
>[BREAKING NEWS]
If you're a ScatGPT user, maybe