it’s sometimes easier for me to have conversations with people when they have a good prompt and I just happen to thinking about and can rattle off (seemingly to me) relevant info, and I get good feedback.
I wonder if it’s an option to learn programming the same way I learned artistic things like video editing and animating etc. For those what I did was just have an idea of where to go, and then pick up sticks and stones as I was proceeding down that road. I can see this being an issue when learning programming because fundemantals and such are very important and honing early/core parts of this skill is a very good investment. Same thing about your wish of me learning the technique of holding a guitar long before trying to learn the fundemantals of playing a riff cuz those early habits are really going to screw you up in the future. But then again I can say that I’ve learned a lot about those core things with editing and such because I was a very methodical applier. I like foldering, organizing and making sure all the fundemantals are set and certain rules are repeated as I start a new project. So maybe just trying to develop a 2D video game without knowing fuckall about programming could work? I don’t know. I would love your insight in this. 1:54 PM
100%
A thinkg I learned with cooking
you can make a certain type of dish and then find that there’ a technique that you can apply to learning other dishes. You have a dish that you want to make and then it opens up a new range of options
“Goals advance technique”
" 64. Often it is the means that justify the ends: Goals advance technique and technique survives even when goal structures crumble. " http://www.cs.yale.edu/homes/perlis-alan/quotes.html
I learned how to make one thing and then I took it apart until I really understood it, and then that gave me an idea for building another thing
it also unlocked the idea that I could build that thing
I think the challenge is picking the battles that you can fight and win so that you’re not faced with a bunch of failures (https://supermemo.guru/wiki/Decoding_failure_penalty)
I don’t know if there’s a perfect way to do it, but “the good plan that works is better than the perfect plan that doesn’t”
Another cool thing is that some of the simplest programs I’ve made are 1. the most useful (my fleetNotes program is stupid simple, and I know I’ve used it over 2500 times) and 2. they still fucking work
needing things to be complicated is a distraction
I made my voicememo2Anki thing and showed a friend (with the program running on a fucked up laptop of mine) and he said, “that thing has a GPU?” me: “no lol” him: “how’d you get it to run without it?” me: “…it just worked”
I didn’t need to know everything, it’s a fucked up program that’s less than 40 lines of code in total, but it still fucking works lol
(I just vomited a bunch of thoughts, does this answer your question?)
Oh, my absolute favorite quote right now is “Always have working code” by Paul Graham (I think I sent this before). I have no clue how to do anything with LLM agents until this program I’m writing right now called for it. And I was able to figure it out because by taking very very small steps and always making sure that my code worked
You sent me some videos a while ago about a front end dude who wanted to recreate certain effects and would break each of them down into simple simple little parts and then figure out how to do those, this is the exact same principle 2:06 PM
It wouldn’t surprise me if you’ve got thoughts of what to make already, what’s the simplest possible version you can think of that would still achieve your vision? If you feel you have to write a huuuuuge fucking program to make your thing work, think about whether or not there’s another existing program that already works that may be able to take place. My fleetNotes thing has all my notes go into Anki mainly because I didn’t want to fuck around with any database shit (and I had also made the voicememo2Anki thing that showed me it was possible). After thinking that, I was able to create that program in like 30 minutes (and it’s never failed)
Paul Graham talks about the dual meaning of the word “hack”, where it could be “that’s fucking awful, a hackjob” or it could be interacting with a system in a way that no one thought previously possible. But the interesting thing is that hackers don’t know which one they’re making when they start, it could go either way
I think you should make some fucked up janky pieces of shit, that work
time to fuck some shit up 2:16 PM