Six months of learning in public — what actually helped

Six months into building and writing about it in public, the honest scorecard is mixed in a useful way — some of what felt worth sharing at the time wasn't, and some of what felt too small to mention turned out to be the most useful thing on the page.

What didn't help: polished retrospectives written well after the fact, once the messy middle had already been smoothed over into a clean story. They read fine. They're not that useful to future-me, because the actual decision points — the moment something almost went a different way — get lost in the smoothing.

What did help: writing down the specific, boring detail in the moment it happened. Not ‘I learned to verify assumptions’ as an abstract lesson, but the actual number that proved an assumption wrong, the actual command that failed, the actual line where a bug was hiding. The abstract lesson is easy to reconstruct later from the specific detail. The specific detail is gone forever if it isn't captured when it's fresh.

The other real shift: writing for future-me first, and treating ‘other people might read this too’ as a nice side effect rather than the goal. The posts that tried to be broadly useful to a general audience took longer to write and were less useful to the one reader who mattered most for a project log — the version of me picking this back up in three weeks with no memory of the details.

Not every lesson is worth sharing, and that's fine — the ones that are tend to be the small, specific, slightly embarrassing ones, not the sweeping conclusions. Those are the ones this whole workshop-notes section is actually built around now.