Strategy for subsequent tutorials

After the feedback I've received in #2705, I'm wondering how to proceed with the subsequent tutorials. I think we kinda have two quasi-conflicting goals:
- on one hand, we want to keep the tutorials short, focused and independent of one another
- on the other hand, I don't think we want the user to keep starting over from scratch

For example, my earlier PR could effectively be split into the following tutorials:
1. Generating your first recipe (already in).
2. Building your recipe (optional, based on the recipe prepared in 1.).
3. Running `pytest` tests.

The point is, 2. and 3. either need to require the user to go through 1. first, or I'd need to start each of them with a bunch of steps to get the initial recipe.

So I'm thinking the right approach here would be to have a dedicated repository for tutorials, where the entry points are available as branches or tags. For extra points, the repository could have a feedstock-like structure, so it would fit as "after your submission has been merged and feedstock repository created". This would also let us cover steps such as rerendering in tutorials.

WDYT?