it’s getting to the point where I notice people say it a lot, especially IRL now for whatever reason recently.
And for clarity I’m not in research or anything, so these people just mean ‘LLM/image gen’, not utilities like OCR or (usually not) transcription.
Some have argued it’s just more efficient (which I can kind of get), while others think you’re actively hindering your intelligence somehow.
On the first point:
I’ve tried it occasionally to see how it compares to my own skill, and while it produces a functional result, it’s always very derivative work to the point where you can find things with the exact same names of other ‘public’ (but not libre) works, and often isn’t the ideal solution to what it targets. So I can see how you can get things out of it, but it never felt really that profound to me.
But for the second… isn’t this supposed to be the tool for people to do things they aren’t experienced in? If anything, you probably need to be able to understand how to write pertaining to the task so the token probabilities are biased toward writing from that area.
And even then, if all you end up doing is prompting AI, then wouldn’t you ultimately serve no purpose outside of being glorified QA?
I guess I’m trying to figure out what exactly non-users would be ‘falling behind’ in that affects them more than those who use AI?
Agreed. Most people are not going to personally benefit from LLMs, in particular those in non-STEM fields. But…
The first sentence is unnecessary for the other two. It is insufficient for the second.
For example, I use a LLM as a LaTeX assistant. I still end up heavily modifying whatever the machine gives me by hand. And I always do my own reading… because the whole point of me reading papers and books is to put them in my brain. And I still volunteer to do certain exercises on paper or whiteboard simply to keep my mind in good shape.
What LLMs allow me to do is focus on the mathematics and the theory of whatever I’m writing, and offload some of the unenlightening cognitive effort of writing and debugging LaTeX code to the machine.
Similar remarks can be made for non-typographical programming in certain situations.
The solution to not becoming a slophead is, more precisely, to occasionally (say X% of the time) put your foot down and actually do the thing you wanted to give to the LLM. The most extreme versions of this solution are to set X := 0 (never use your brain) or X := 100 (always use your brain, your solution). And while for many people setting X := 100 might be the best solution, for some of us, particularly in information-dense technical fields, setting 0 < X < 100 (both inequalities strict!) might be able to make our lives easier without slipping into slophead behavior.
Firstly thank you for such a thoughtful and long reply. You are right about LLMs reducing cognitive overload and helping you concentrate on the core problem. I avoid them for environmental reasons too.
I too am not a fan of boilerplate and tedious languages like latex.
My approach for dealing with it is using better abstractions like typst. When I was seriously writing LateX years ago I had emacs macros and shortcuts to power through the tedious parts.
I also choose modern programming languages like gleam instead of java or js. They are easier to reason about and write fast manually. IME you can overtake the productivity gains that LLMs offer after 3-6 of working on a project.