From the reflections section:

> x86 vs. ARM. Working with ao486 deepened my respect for x86’s complexity. John Crawford’s 1990 paper “The i486 CPU: Executing Instructions in One Clock Cycle” is a great read; it argues convincingly against scrapping x86 for a new RISC ISA given the software base (10K+ apps then). Compatibility was the right bet, but the baggage is real. By contrast, last year’s ARM7‑based GBATang felt refreshingly simple: fixed‑length 32‑bit instructions, saner addressing, and competitive performance. You can’t have your cake and eat it.