Starship is a fuel-hungry beast - it can get to LEO by itself, but it needs a lot of tanker launches to go beyond. And if your goal is a Mars colony, you don't want to be limited to one launch per launch window.
We could just construct 200 Space Shuttles and spend months refurbishing them after every flight, and still send one up every week.
The goal is to drive down launch costs, time is money, and a system that requires time consuming refurbishments is more expensive.
Once the tanker version is needed, a ship ship could go up 5+ times a day. The logistics of backfilling a pad with a new ship is much more involved
As you say, they reïnforce each other by speeding up the learning curve and deployment of learning to the real world, serving as both a bolstering of the product and experimental validation.
Elon always talks about a city on Mars but seeing for the first time the gargantuan size of Starfactory it dawned on me that SpaceX are true believers. It is still a big IF, because the dimension of the mission is absolutely bonkers, but IF the goal is to send every two years hundreds of Starships to Mars (everyone needing around 3-4 tanker missions) you need large scale production of ships.
Anyone who thinks they can’t do stuff is not seeing the whole picture.
It seems like if they can get boosters to rapid reuse (a much easier goal), and churn out ships at sufficient scale, they can afford to take time inspecting/refurbing each ship as part of a pipelined approach.