One answer I am trying for what to build today is a platform and an API as an intermediate layer, something solid to build against while I keep exploring. I will probably need data for most projects in the spaces I work in.
Downside: I might distract myself from a better opportunity (which I do not see yet). Maybe the right path will not require gathering data at all.
For now, I assume we take this direction.
Given the tools we have as software engineers, keeping everything in a single repo starts to make sense, at least for internal tools, static sites, and experiments. I want momentum and a setup that wastes less time testing ideas, does not block me with hypothetical costs, and still lets me learn as I go.
The technical walkthrough lives in engineering: Self-Hosted Platform. This writing page is the why and the shape of the bet; that series is the how.
I also want to keep a journal about this platform. It helps me document and share context, and it surfaces blind spots, mistakes, and gaps in what I notice.
I would rather pick the best next step each time than plan everything up front. Locking into fixed assumptions makes it harder to adapt as context, requirements, and information change, and it can bake the wrong biases into the next move.
There are other reasons to care about a data-oriented platform:
- Many products need data behind their apps.
- Mine do too.
- It can be a faster way to run experiments.
Next I want to study how companies that sell data package and price it, and to read their terms and conditions carefully. Business models and legal constraints matter as much as the tech stack.