The weekly review that's allowed to say no
Most systems only ever add. Once a week I run the opposite — a standing appointment to demote, park and kill, and to check the week's reality against what I claimed I'd do.
A weekly review is the scheduled subtraction pass: it checks proof-of-progress, enforces the limit on active work, promotes what proved durable, and escalates when the warning signs show. The hard part isn't the checklist — it's removing things I'm attached to, on a cadence so it actually happens.
Every system I've described so far shares a bias: it adds. Capture adds notes, the diagnostic lab adds verdicts, the learning system adds episodes. Left alone, a system like that silently fills up — more active projects, more parked ideas, more half-true claims about where things stand. The weekly review is the one routine pointed the other way. Once a week, for fifteen or twenty minutes, the job is to take things away.
A standing appointment, not a mood
It's fixed, because the work it does is the work you least feel like doing: admitting something stalled, parking an idea you were excited about last week, deleting a claim that's no longer true. Leave that to willpower and it never happens on the weeks that matter. So it's scheduled, and a routine walks me through the same questions every time. It's the wide end of the same bracket that opens and closes every working session — see the routines that make the rules stick.
What it checks
A handful of blunt questions, asked against the week's actual evidence rather than my intentions:
- Proof of progress. Did each active stream ship something visible this week? A priority that's shown nothing for two weeks gets questioned or killed — not quietly carried for another month because cancelling it feels like failure.
- The limit holds. There are three active slots, no more. If something new crept in, what did it displace? "Nothing, I'm just doing four things now" is itself the finding.
- What's no longer true. Plans, state files, claims — anything the week falsified gets corrected or deleted at the source, before it rots.
- Escalation signals. Too many open fronts, a parking lot filling faster than it empties, the same decision dodged three times, work being done out of resentment rather than judgement. Each is a named trigger the review goes looking for on purpose.
Promotion is the other direction
Subtraction is the headline, but the same pass also promotes. The durable lessons of the week — a concept that sharpened, a decision worth keeping, a pattern that proved out — get lifted out of the daily noise into a permanent knowledge layer. Capture is wide and cheap; promotion is narrow and deliberate, and conflating the two is exactly how a knowledge base turns into landfill. The review is where that line gets drawn: most of the week is disposable, a few things are permanent.
It keeps the public version honest too
There's a newer item on the list. This site is a hand-curated snapshot of the workspace — capability levels, counts, claims — and snapshots drift. So the review now includes a reconciliation step that checks what's published here against what's actually true in the system, and flags anything stale before it can mislead. The review isn't only private hygiene; it's what stops the outward-facing version of me from quietly going out of date.
Why it's the hard part
Anyone can keep a list of ideas. What's actually rare is looking at your own work the way an outside manager would and removing what isn't earning its place — including the parts you're fond of. Adding is easy and feels like progress; cutting is uncomfortable and feels like loss, which is precisely why it can't be left to a good week. A system that only ever grows isn't a system. It's a hoard with better branding.