> I’ve consistently failed to make writing elisp net positive for me for basically anything.
That's sad to hear, because I somehow have the exact opposite experience. First, it was about fixing things in my config. Then I often would catch myself in a "why can't I do this better" state. And it could be anything. Like one day I was copying the current URL in the browser, and then switched again to copy the description. Next day I caught myself doing that again. And I realized - "shit, I never thought before how often I have to do this crap". Then I wrote a helper.
These days, I don't even blink, if I need something, I'll just start whipping up some shitty Elisp in my scratch buffer. And with help of LLM packages like gptel - it's a breeze. What's awesome is that I can play with that code even without having to save that shit anywhere. Sometimes, I realize the solution is not as easy to make as I initially thought, so I'd just leave it there in my scratch buffer. It sits there, sometimes marinating for weeks - I have a persistent scratch buffer. I even get excited when I stumble on a problem I want to optimize.
Few weeks ago I was in a pair-programming Zoom, and my colleague, let's call them Matthew, was screen sharing. I couldn't help distracting Matt asking questions like, what is that? Can you share this link? Wait a minute, don't scroll away, I need to write that down. etc.
That's sad to hear, because I somehow have the exact opposite experience. First, it was about fixing things in my config. Then I often would catch myself in a "why can't I do this better" state. And it could be anything. Like one day I was copying the current URL in the browser, and then switched again to copy the description. Next day I caught myself doing that again. And I realized - "shit, I never thought before how often I have to do this crap". Then I wrote a helper.
These days, I don't even blink, if I need something, I'll just start whipping up some shitty Elisp in my scratch buffer. And with help of LLM packages like gptel - it's a breeze. What's awesome is that I can play with that code even without having to save that shit anywhere. Sometimes, I realize the solution is not as easy to make as I initially thought, so I'd just leave it there in my scratch buffer. It sits there, sometimes marinating for weeks - I have a persistent scratch buffer. I even get excited when I stumble on a problem I want to optimize.
Few weeks ago I was in a pair-programming Zoom, and my colleague, let's call them Matthew, was screen sharing. I couldn't help distracting Matt asking questions like, what is that? Can you share this link? Wait a minute, don't scroll away, I need to write that down. etc.
After that I thought of solving that dilemma. I sat down and wrote a helper that calls tesseract to OCR text from an image in clipboard. Took me no more than 20 minutes. https://github.com/agzam/.doom.d/blob/main/modules/custom/wr...