rediscovered magic in computing
although i don't quite remember what our first family computer was i have vivid memories of spending countless hours on them. i had no idea what i was doing, but that was the fun of it for me. as a kid who couldn't get enough of puzzles, computers were the biggest puzzle of them all, and in every error message was a hint that i followed until it went away and i had learned something new. it was like a treasure hunt of sorts.
when i was in the third grade, my teacher gave me my first gig work teaching her how to use a word processor after school. i worked for my school district in high school, and having enough credits to graduate early, i got to do tech work for the back half of every day for my senior year; ride the bus, take a few classes, then be done with classes after lunch and get paid to help out in the lab or run around and help teacher's with their PCs.
later on in life though, corporate sucked all of the joy out of it entirely: endless amounts of people without hands-on skills who only got their jobs through memorizing certification questions, "scrum masters", managers who cave to every single ridiculous ask from outside teams, people in lead roles that have no business being in them, no time to fix bugs... absolutely maddening.
then the LLMs came....
i was quite underwhelmed at first with the earlier generations because things would completely fall apart after only a few prompts, or the resulting code that came out was complete shit, and early in-IDE was little more than glorified line completion, but man, has that changed.
there are still a lot of detractors saying things like "it's just vibe coding" or "prompt and pray, eh?", etc. maybe. but here's the thing...
you have to have a decent idea of how to actually build robust stuff, understand how something should/could be architected, and be able to know when a particular design choice is going to be unsound down the road.
if you have those basic qualities, you can indeed build useful things with AI, and i challenge you to try. last night i had an idea to write some alerts and flag staging for CTFd-powered CTFs. this morning i sat down, wrote a design doc, set up a test CTF, then prompted through the entire thing in under three hours and it came out waaaaay better than i had originally wanted it to.
through my experiences like that (and many others prior), i'm pretty sure those detractors are dead wrong.
either way, i don't care because computing is magical to me again, and that's what i'm in this for.
-agatha
