Lets say we have an enemey ai and we want it to be able to act like a real enemey in a stealth game, do we handle all its behaviour in a single script or should we seperate the scripts out into each behaviour such as following player script and patrolling script.
Making such decisions varies from person to person and condition to condition. Writing whole behavior at one place saves a lot of messaging(communication) effort but you can’t reuse partial behavior anywhere else then. I would suggest you to what ever way you prefers to write just keep the Single responsibility principle in mind too.