EST. 2026
The Curator's
Manifesto
We are uncovering better ways of solving problems by building solutions faster than we can deliberate about them. Generation is now abundant. Judgment is now scarce.
// through this work we have come to value
That is, while there is value in the items on the right, we value the items on the left more.
// the values
Problems over backlogs.
A backlog is a queue of stale guesses — solutions decided months ago for problems that have since moved. When building is instant, stockpiled solutions are inventory rot. Curate the problem space instead: what hurts, for whom, why. Generate solutions fresh at the moment of need.
The backlog isn't an asset anymore. It's a museum.
Selection over estimation.
Story points are a futures market on labor — and the market just crashed. When three working alternatives cost less than a planning-poker session, predicting is strictly worse than trying. Set criteria first, build in parallel, choose by comparison.
The working artifact is the estimate.
Verification over generation.
Generation is now abundant. Knowing whether the thing is correct, secure, and honest barely accelerated at all. That's the new bottleneck — so that's where human expertise goes. The senior engineer's job is no longer writing the hard code.
It's being the inspector who can't be fooled.
Coherence over throughput.
Cheap generation creates a failure mode Agile never met: sprawl at machine speed. Ten teams shipping fast produce five auth patterns, three duplicate services, and an architecture nobody chose. Output is cheap; fit is scarce.
When the rate and the whole conflict, the whole wins.
Keeping deliberately over shipping continuously.
Shipping isn't the expensive act anymore. Adopting is — every kept system must be secured, staffed, patched, and maintained for its entire life. Deploy continuously; admit into the portfolio deliberately.
The demo is the cheapest moment in a system's lifetime. The decision to keep it is the costliest.
Judgment over procedure.
The immune system for everything above. Every gate here can calcify into a checklist, a review board, a compliance ritual — exactly how Agile became SAFe. So: when the procedure and the experienced eye disagree, the eye wins and the procedure gets rewritten.
Process is the fossil of past judgment — never a substitute for live judgment.
// the principles
- 01.Our highest priority is understanding the problem. A fast solution to the wrong problem is not speed; it is debt compounding at a faster rate.
- 02.Treat built software as a question, not an answer. Prototypes exist to make a decision better-informed, and their natural fate is deletion.
- 03.Build alternatives in parallel and choose by comparison. Estimation predicted cost when construction was expensive; when construction is cheap, the honest forecast is the working artifact itself.
- 04.The critical gate is no longer "should we build this?" but "should we keep this?" Guard the keeping decision with the rigor that sprint planning once pretended to have.
- 05.Every kept artifact is an operational commitment. Budget for the tail — the security review, the migration, the on-call burden — not the launch. The demo is the cheapest part of its own lifetime.
- 06.Spend human attention where machine leverage is weakest: verification, judgment, and taste. The senior practitioner's job is no longer producing the hard part; it is being the inspector who cannot be fooled.
- 07.Discard without ceremony and without grief. Deleting a prototype is progress. Hoarding working code you don't need is the new technical debt.
- 08.Make learning compound. Every cycle should leave behind context, evaluations, and patterns that make the next cycle cheaper. This residue is what turns a pipeline into a flywheel.
- 09.Coherence is a feature. Five solutions that fit the whole are worth more than ten that don't. Someone must hold the city plan while everyone else pours foundations.
- 10.Synchronize continuously, not on a calendar. The cadence of meetings must never be slower than the cadence of the work, or the meetings become the bottleneck they were meant to remove.
- 11.Reflect on judgment, not velocity. The retrospective question is no longer "how much did we ship?" but "did we choose the right problems and keep the right things?"
- 12.Trust those closest to the problem to generate freely; reserve authority for keeping decisions. Hierarchy belongs at the expensive gate, not the cheap one.
// the curator
The curator is not a gatekeeper. A gatekeeper says no; a curator is someone whose yes means something. Gatekeepers subtract. Curators compose.
[ problem space ]
A living map of what hurts, for whom, and why. This is the permanent collection; the backlog was the gift shop.
[ prototypes ]
Many things are made; few are exhibited. The curator's eye decides which, and that eye cannot be fooled.
[ operational estate ]
What enters the portfolio must fit the portfolio, and everything admitted must be conserved: staffed, secured, maintained.
The curator moves fast precisely because most decisions are "no."
Everyone builds. Curating is the discipline.