I've spent most of this week learning about the architecture of a very complex system. As someone who has always been fascinated by puzzles, I think this part of software development feels like candy for my brain.
I’ve been calling it the venture into the why. Why did they solve this issue in that particular way? You eventually find out, because the moment you let your guard down—when you think you’ve found a shortcut—you get slapped by the backhand of reality. Still, backhands and all, I enjoy the process of learning, of challenging myself.
If I had my way, if I had the right skills, this week would have ended with me being able to do a lot more to help a friend. But I’m not feeling too defeated either. On one hand, I know far more than I did a week ago. More importantly, I think I helped my friend at least get going in the right direction. Like when a car person tells a buddy exactly what the mechanic needs to do, so they avoid the typical repairman grift.
Along these lines, I was talking with my brother about it. He helped guide me through dissecting the piece of software, so it was only natural we discussed where things are headed.
It’s obvious to me that the people who will remain relevant in the near future of software development will be those who can design solid architectures. Not so much the grunts who just carry out the work. If I compare it to construction, to building a home, the architect and the engineer might keep their jobs—assuming they’re good enough. But it sure seems like AI will take over all the sweat and tears.
In this future—which I think is very near—the people who will hold the keys are the ones with better ideas. It’s fascinating to imagine, exciting even. But I can admit this while also knowing AI will be extremely disruptive economically. Whole hierarchies of tech workers will be destroyed—or at best, rearranged.
Today was productive for me. I finished a tool for Snapie that makes moderation much easier. I’m already thinking of more features to add to it. As I dream about what it could become, I keep circling back to the idea that soon, very soon, what now takes me a day or two could be reduced to a single prompt.
I don’t know if that’s something to celebrate or to fear. Somehow both feelings coexist in my head. Maybe there’s a part of me—a naive part—that thinks I can survive this switch. But the voice that reminds me I’m getting older isn’t getting any quieter with time.
The thoughts of a Sunday night, it seems.
-MenO
I can relate, to the brain candy, the getting slapped by reality on supposed shortcuts, and the current air of the developer landscape.
I can't afford to play with replit and others like it... but from what I hear, its already "there". Full apps being built with a few prompts.
I can't afford any of those either. I have a local llm and I pay 10$ a month just to be able to work with "how do i do x" and go from there.
same boat as me! hahahaha
Really interesting thoughts! I like how you connected software architecture with the changing role of AI. It’s true, understanding the reasons behind design choices makes a huge difference. The future of this field looks exciting, even if a bit uncertain.