What's the difference between flash and SRAM on an MCU?

Flash is non-volatile memory that retains its contents without power; it holds the program code and read-only/const data, and the CPU typically executes directly from it. It's larger but slower (often needing wait-states at high clock speeds), and writing to it is slow, requires erase-before-write, and wears out after a limited number of cycles. SRAM is volatile, it loses its contents at power-off, and holds your runtime data: variables, the stack, and the heap. It's fast (often single-cycle access) and freely writable, but smaller and consumes standby power to retain data. So at runtime your code runs from flash while your mutable data lives in SRAM, and at boot the startup code copies initialized variables from flash into SRAM and zeroes the rest (the .data/.bss init, covered in the boot lesson). Knowing the split explains why globals persist, why string literals are read-only, and why you can't just "write a variable to flash" casually.