I was also slightly offset by this, particularly the vague phrasing "ultimately the same", which by reflex I would've asked to taboo. However, by charitable interpretation, I think the intended meaning is that everyone is running on the same source code. Even if the source code contains modules that take set values according to runtime events and then become irreversible (or extremely difficult to alter), which leads to the same "program" doing vastly different things and having different capabilities.

An example intuition pump here might be to imagine a standard PC running a custom OS that enables or disables a bunch of its key features and messes a bunch of its parameters or will use different optimization subroutines and garbage collection procedures during it startup routine all according to some hidden, unknown algorithm that takes pictures of the user during said startup as input.

Obviously the sourcecode and hardware are the same, but the behavior and capabilities will be radically different depending on the user. You might even be able to hack parts of the OS during runtime to enable certain disabled features or tweak some parameters, but how much can be hacked and how to do it is unknown at first.