Reading a tutorial and terms like exponential backoff, jitter, and idempotency keys makes sense. Then you build the thing, the API returns a 500, and you watch your system crash.
The happy path is the part everyone writes about. The request succeeds, the response looks like the example, and the system behaves. That is maybe 20% of the actual code you end up shipping.
The other 80% is what you only discover by running the thing under real conditions. A queue that is supposed to be FIFO but is not under load, a dependency that is “always available” until it is not, or a database connection pool that gets exhausted when 3 slow queries stack up during a deploy.
Hence, reading 10 blog posts about a system is no substitute for spending a weekend building it. The bugs you hit while building it (or, better yet, while running it in production) are the actual curriculum. The blog post is just the syllabus.
Reading teaches the happy path. Building teaches the failure modes. So keep coding and keep shipping.