Mental Model
Rote memorization doesn't work well for my brain. As a kid, I couldn't memorize poems or formulas without understanding how they worked. Later on, I cared much less about memorizing textbook definitions than understanding actual system behavior.
I think in data and state flow: Did the packet arrive? Did the socket read it? Did the application process it or silently drop it? Where did the assumption fail? Instead of memorizing interface steps, I check what the machine is expected to do, read the docs and --help flags, and trace the execution path.
I treat AI as extended memory and leverage. It handles routine syntax lookups and boilerplate while I focus on architecture, data flow, and root causes. Problems that typically stall on guesswork get diagnosed much faster.