kirurik

You can run some models pretty decently using CPU inference only, things like Gemma 3 that are built for exactly that use case or some tiny speech to text models via llama.cpp that I have tested out (not so good). Although not the best for "heavy" tasks, if you just need a decent text generator that can produce more or less sensible, generic output you are good to go.

Thank you for replying to me and recommending a feature - that I didnt know about!

Saving this

I agree, I think there is such a thing as AI overuse, but I would rather someone uses AI to form their points more succinctly than for them to write something that I can't understand.

To be fair, you are assuming that the input wasn't garbage to begin with. Maybe you only notice it because it is obvious. Just like someone would only notice machine translation if it is obvious.

True, it is meaningless how good one is at vimgolf if it doesn't flow naturally

I did the neovim tutorial which I accessed via :Tutor, I see now that vimtutor has some divergences, I will do that tutorial also, to improve my muscle memory and knowledge.

Thanks for your other tips as well. I'll try and approach it that way, just doing it little by little, I don't always need VSCode and for simple edits its already way better than nano, although tbf I didn't take the time to learn all of nanos functionality.

It probably is best to take it slow and not cut corners. Play some vimgolf and take it slow, try to get the base stuff like regex search and replace down. Learning regex well can be applied to other things as well, so it is worth it to practice it in the context of vim.

I'm currently learning vim and usign neovim, coming from VScode, wondering how y'all learned it if you made a similar switch, how long it took to learn and your opinions on purely text based editors vs smth like VScode that has a ton of features and plugins.

It seems obvious, but I hadn't thought about it like that yet, I just assumed that the LLM was finetuned to be overly optimistic about any user input. Very elucidating.

True, and many not so intelligent people may overestimate their intelligence and do it. So maybe it does somehow end up being representative.

The self-selection bias is definitely something to consider. I’d guess people who feel relatively confident in their intelligence are more likely to take the test.

There definitely is an element of shooting oneself in ones own foot, but sometimes it seems unavoidable to me, or the effort just isn't worth it e.g. if I am using sklearn or numpy and the return types are ambiguous, then I'd have to overload each function at the head of the file or wrap it although it is clear what it does. What do you think? I think that if it's only my own code, then yes this is certainly avoidable with good composing functions.

I am not an experienced programmer, but I liked python because of the dynamic typing, but tbh no type hints are a nightmare (as I used to use python). These days I gravitate towards using type hints unless I am using an ipynb because it looks clean, but it can be a little much, it can look quite ugly. Not every usecase needs type hints is what I've learned.