Back in college, I had a Systems Architecture and Design professor. The very first day of his class he stood up in front of us and said:
"One day, you're going to be brought in to help optimize a company. You'll see that they're doing things by using index cards in shoeboxes. Then, after you understand their processes, points of interaction, and audits, you'll find out that the most optimal way for them to work is by using index cards in shoe boxes. Walk away."
That was one of the best bits of advice I ever got in my career. Why? Because it reinforced that in order to solve a problem, you have to understand the problem. You have to understand how the problem fits into it's larger ecosystem. You have to understand what the solution breaks and what problems it will cause. Nothing lives in isolation. Everything is interconnected.
The last thing you do as a systems architect is write code.
I wish like hell more folk understood that, particularly the vibe-coding, rapid iteration, "throw code at it" types.