that's purely based on the amount of cognitive effort we output when achieving a task, isn't that the same kind of worry people had when the internet became a thing?
Using an LLM to handle a task for you seems a lot like letting a car move you. Cars will make you “fat and lazy” if you never move your body otherwise, but it’s fairly clear to see that this is avoidable.
The research seems to always get (intentionally?) misconstrued at headlines that LLM is “bad for you” as opposed to more mundanely stealing opportunities for exercise and practice of mental activities if you let it.
I often find AI makes me angry and stressed out, especially when it suggests dumb solutions to problems. Honestly makes me wonder if I'm more likely to die early from chronic AI-induced stress rather than dementia.
I remember when my school introduced calculators and my parents got upset about it: "They won't learn to do sums in their heads!" Yet it opened us up to working on more interesting, larger problems, at a faster pace. LLMs could atrophy skills if used solely out of laziness (like the cover letters in the post), but they could also help you punch higher, and learn more, and faster, if you're motivated and mentally integrate them properly.
Beware of the false equivalence. AI is not just a calculator that is automating a tedious and repetitive task. It is also automating away things like creativity, research and critical thinking. In addition, we are seeing more and more people use ChatGPT as a source of truth instead of academic literature.
No matter what way you run a math problem, the answer will be the same whether you use a calculator or work it out by hand. This is not the case with AI and the tasks it replaces.
Feels like one of these things that's been known for decades in the general form: tools that take cognitive load off your working memory (a calculator, writing) free your brain up for higher level thinking make you "smarter", whereas tools that take the higher level tasks off you and load up your working memory (hypertext, AI) make you "stupider".
My biggest worry isn't that it will make me dumb (it won't), or that it will make me lazy (it will), but that people raised with it wont learn things in the first place. I'm split on if this is a real issue or an old man rants about slide rules and the decline of mental math kinda situation.
I recommend people look at the actual study and think about how representative are the subjects, the tasks involved (SAT essay writing), and the way LLMs are being used.
To be concrete, this is taking a task in isolation that LLMs can do much better than humans (writing garbage essays) and using LLMs to do that task. In the real world, tasks have parts and they exist in a larger context. When we use LLMs for one part of a task, there are other things we're doing that the LLM is not helping with. If you compared people doing arithmetic by hand and with a calculator, you would also see very big differences in how active their brains are. But it's not anyone's job to add up numbers. Adding up numbers is a subtask of a subtask in someone's job.
What makes me stupid is hearing about "AI" day after day like it is the best thing since sliced bread, and yet 99.9% of useful things that ive seen come from LLMs is just low level programming tasks or fluffed up nonsense that any manager could spew. I can't even trust what LLMs tell me unless the answer is so simple a 2015 google search's top result would be just as adequate. Except now the top 20 google results are all AI answers from the same source material, packed full of fluff but stripped entirely of nuance or useful adjacent knowledge. Just changing the question slightly can give contradictory answers with both given with full confidence.
I would believe you if this were 2024, but Opus 4.7/GPT 5.4 are genuinely impressive in many domains and do perform work one would expect of a college-educated, experienced dev. People don't understand that "AI" is not a static thing, failure modes that make LLM's completely useless one year could be generally solved by the next year (remember the whole thing about image models not being able to do hands? Or LLM's not being able to do math?)
For you, no. For the services you depend on and will continue receiving your data, and may jack up the prices/add limitations knowing that your dependency won't be easily broken, yes.
22 comments
[ 2.6 ms ] story [ 43.1 ms ] threadThe research seems to always get (intentionally?) misconstrued at headlines that LLM is “bad for you” as opposed to more mundanely stealing opportunities for exercise and practice of mental activities if you let it.
"I have one job on this lousy ship, it's stupid, but I'm gonna do it! Okay?" -- Sigourney Weaver
[1] - https://www.youtube.com/watch?v=W4CgQMJCpZI
"Someone said something about AI"
There it is.
No matter what way you run a math problem, the answer will be the same whether you use a calculator or work it out by hand. This is not the case with AI and the tasks it replaces.
https://arxiv.org/abs/2506.08872
To be concrete, this is taking a task in isolation that LLMs can do much better than humans (writing garbage essays) and using LLMs to do that task. In the real world, tasks have parts and they exist in a larger context. When we use LLMs for one part of a task, there are other things we're doing that the LLM is not helping with. If you compared people doing arithmetic by hand and with a calculator, you would also see very big differences in how active their brains are. But it's not anyone's job to add up numbers. Adding up numbers is a subtask of a subtask in someone's job.