After working with dummy data, I decided to create a small admin panel to create lists and do post-editing before publishing them.
The panel has a basic authentication view that helps me connect with a Google AI library to draft the lists and generate preview images. It usually struggles with connecting links, for example: IMDB links for movies, so I still do some post-editing, but it saves tremendous time.

Now whenever I feel inspired with a new marathon or have a new idea, I can just login and draft new lists and publish them. The friction is minimal, which means I'm more likely to actually add content when inspiration strikes.
I also added analytics and search console connection, which will give me signals in a few months if it's taking off organically. This is a very low effort project for us, and we have many small bets like this.
These kinds of projects also allow us to have different types of easy boilerplates to start new interests along the way with each iteration. They're not meant to be huge commitments, but rather experiments that can grow naturally if there's genuine interest.
The admin panel makes the whole process sustainable: low maintenance, high potential for organic discovery.