A browse tree is navigation, not an academic map. The nodes that earn a place are the ones where courses actually cluster — everything else is a dead end.
A failing check gets fixed the same day. A check that passes while doing nothing can bill you every push for months, because a green tick looks exactly like success.
Keep one crisp SVG as the source of truth and render a raster only when something needs one. You get sharp images everywhere, tiny files, and a re-skin that costs one edit - without the social-preview blind spot.
Run two automated agents against the same working directory and one will quietly commit the other's half-finished files. The fix is one working tree per line of work, not more discipline.
A query that's instant on a new blog can crawl on a large one - not because the server got weaker, but because the database is scanning millions of rows it could have jumped straight to.
You're mapping a messy source vocabulary onto a clean target taxonomy. The target has about 5,600 slots, so the job must be roughly 5,600 decisions — right? The backlog is already past 8,900, and still climbing.
You shipped a fix to how one field is computed, re-ran the pipeline across everything, and it reported a clean run. Then the consistency check still flags those records as wrong. The fix is in the code, the rebuild "ran" — and the records never changed. Your own optimisation skipped them.
The report said 80 duplicate groups still to resolve. You'd been resolving them for days. You run it again after a big batch and it *still* says 80-ish. The backlog isn't stubborn — the scan is counting work you already finished.
You hardened a skill in the repo, shipped the PR, and the agent kept loading the old one. Or you fixed the live skill by hand, and the versioned copy quietly stayed behind. Same root cause: the copy you run and the copy you version are two different files.