[jeffreyrogers]

I would bet that how your brain stores information that you read from long-form text is very different from how it stores information you acquire from chatting with an LLM. When I read something challenging or new to me I spend a lot of time thinking about how what I'm reading matches my own experiences or knowledge. Although I'm a fairly fast reader, it often takes me a long time to get through difficult pages since I have to stop and think about what I'm reading. I seem to be doing a lot of integrating and reorganizing my thoughts. When interacting with LLMs it feels a lot more like I'm just receiving knowledge passively and I don't think it gets integrated as well. Not sure why this is and its somewhat counterintuitive since I don't think I'd have the same experience with a human tutor.

[theshackleford]

> When interacting with LLMs it feels a lot more like I'm just receiving knowledge passively and I don't think it gets integrated as well.

This sounds like it's more down to how the individual uses the tool. I am not someone who has even been particularly good at reading > learning the thing. I've only ever been capable of learning by doing and with capability to interrogate on the points I don't get. LLMs allow that in a way that is just not feasible with any human being whose tolerance of me would diminish rapidly.

In most cases, I will be writing down my understanding as I would in isolation from a primary source, building flash cards and actively practicing what I have learned, only with more capability to interrogate on the points I have difficulty understanding or need clarity on. Effectively, I am doing the following in a capacity I personally never had via any other means:

> I seem to be doing a lot of integrating and reorganizing my thoughts.

If you are using it as a slot machine of knowledge, and going from receiving > doing with no intermediate step I see how outcomes could differ.

[tclancy]

Seconded. At a minimum, you can literally tell the agent to guide you but not point you and only help when you are stuck if you're actively trying to learn something. It's been a help for me to bridge a couple of gaps I'd struggled to cross before (mainly hardware things).

[Forgeties79]

> This sounds like it's more down to how the individual uses the tool.

I genuinely agree with a lot of your comment but lines like this drive me crazy. Anytime somebody has a critique of AI, so many people hand wave away with “well you’re doing it wrong.” If we can just chalk up every negative thing with AI to “incorrect tool usage” then we’ll never honestly assess it.

[theshackleford]

I didn’t say they were using AI “wrong.” That’s a stronger claim than the one I made.

Their assessment was:

> When interacting with LLMs it feels a lot more like I'm just receiving knowledge passively and I don't think it gets integrated as well.

My point is that the passivity is not necessarily an inherent property of the tool. It depends heavily on an individual chooses to engage with the output.

Saying LLMs make learning passive is a bit like someone else saying books make learning passive because you can skim them without thinking. When we all know instead of reading a book passively you could in addition annotate it, question its claims, summarise it, test yourself, and ultimately apply what you learned. The medium permits both. The same is true of the output of an LLM. You'll get out what you put in and nothing more.

You can use it like a slot machine, repeatedly pulling the lever for another answer without reflecting on any of them. But the fact that a tool permits shallow engagement does not mean shallow engagement is the only option available to you.

That also doesn’t mean every criticism of AI should be dismissed as “incorrect tool usage.” AI has real limitations and deserves serious criticism, and I assure you, I am not popular with the AI hype bros for this reason. However we still need to distinguish between limitations of the tool and consequences of how someone uses it. Not every potentially negative experience is rooted entirely in the technology itself. Sometimes user behaviour is a significant part of the cause.

[m463]

I listen to audiobooks while I walk and hike, and sometimes when I recall some particular thing I've learned, I can also recall where I was hiking when I heard it.

I think there's a lot to learn on how we really process information.

[justinnk]

I once heard a lecture about „learning how to learn“. There are already a lot of well-researched things one can exploit about the way we process information. What you describe is known as retrieval cue. Our minds seem to be associative^1. It also seems a lot of the learning/processing happens after the actual learning process, which is why resting is important. Spaced repitition is a further method well-proven in practice.

^1 In the early days of AI, the concept of Hebbian learning was popular („neurons that fire together, wire together“). Its implementation is even simpler than gradient descent used today, but never caught on https://en.wikipedia.org/wiki/Hebbian_theory

[jsty]

Perhaps not the same one, but there's a fantastic course by Barbara Oakley by this name. I viewed it quite a while back, but the modern incarnation seems to be [0]. In a similar vein, the book "How we learn" by Stanislas Dehane is also very good!

[0] https://www.coursera.org/learn/learning-how-to-learn

[tony_cannistra]

Same here. I always marvel at this. These recollections can even be years later, and the memory of the place in which I heard the remembered thing is vivid.

[annzabelle]

There's a podcast I listened to as it came out when I was a teenager. I'm now relistening to it over a decade later, and I distinctly remember walking home from the bus stop with it coming through wired earbuds from my ipod.

[ahartman00]

"when events are represented in memory, contextual information is stored along with memory targets; the context can therefore cue memories containing that contextual information" https://en.wikipedia.org/wiki/Context-dependent_memory

[lesostep]

That's spatial memory! I have long walk-and-talks with friends, and sometimes I forget minor details of previous discussions, but not where we were when it was mentioned.

A lot of my friends can't do same trick. But the fascinating part is that it still works for them! Usually they can recall whatever I forgotten after I described the place we were in.

[internet_points]

i read a school-assigned book during the loading portions of a certain computer game. the two stories are completely interlinked in my head.

[nimonian]

What you're describing is very close to what education theory calls "assimilation" and "accommodation". When we assimilate knowledge, it just fits into our schema of understanding. When we accommodate knowledge, we need to change our schema, and that is where the feeling of being challenged (and often the feeling of profundity) comes from.

For example, a child learns that "foreigner" means someone from outside their country. Then, when they're 11, they go on their first holiday abroad and realise "Wait! _I_ am a foreigner here!"

So, maybe one way to frame what you're saying, is that LLM output tends towards being easily assimilable.

[raincole]

> When interacting with LLMs it feels a lot more like I'm just receiving knowledge passively

It's... almost an oxymoron, and very different from my experience.

[jeffreyrogers]

In college I used mathematica a lot while taking linear algebra. I ended up having to relearn a lot of that math later since I never really understood it at a deep level, although I was able to use it and apply it under class conditions. I feel like a lot of my LLM derived knowledge is similarly superficial.

[vouaobrasil]

That's absolutely right. Math is difficult to understand without writing it down on paper. As one of my professors said, "something magical happens when you write down a problem from scratch and work it out". LLMs make people believe they are improving their productivity and that may be so, but it also removes that magic and removes that deep learning.

[altmanaltman]

Any such bet would be grossly misguided given how different people acquire information differently. Just because you don't like using LLMs doesn't automatically extrapolate to a general realisation about the human brain or how one approaches knowledge.

Your subjective experience matters in this case. Almost completely.

[jeffreyrogers]

I like using LLMs and one of the main things I'm working on would be impossible for me to tackle without them, but I do think they cause some skills to atrophy despite being a big productivity booster overall.