I'm not quite sure if even that is correct. The AGC, as far as I understand it, did do quite a bit of calculation on the fly and was essentially the first digital fly by wire system. It did rely on input from the crew and ground control for eg correcting its state vector etc etc, but it even has dedicated vector instructions if I recall correctly. Can't really precompute all that much when you can't be sure things will go to plan and you're dealing with huge distances. It did have eg separate programs for different phases of the flight but they weren't really precalculated as such, more like different modes that eg read input from different sensors etc etc.

The US space program was pretty big on having a human in the loop though, much more so than the Soviet one which relied more on automation and the pilot was more of a passenger in a sense, sort of a failsafe for the automatic systems.

The book Digital Apollo goes into all this this in more detail, I can highly recommend it if you're a ginormous nerd like I am and think that computers we've shot into space are endlessly fascinating