Margin-Pippa Is the Moat

"Editor features can be rebuilt. The Pippa who has lived in the margin of this text for years cannot."

Everything Mechanical Is Replaceable

Be honest about what Rekindle's mechanics are, and aren't. Live Preview, tabs, a sidebar, export, snapshots, even the CMD+K inline diff (a shape Cursor popularized) — all of them are serious engineering, and all of them can be rebuilt by a team that decides to pay for it. Rekindle proved that on itself: the whole mechanical layer was rebuilt once already, on a different stack, without the product changing. If Rekindle's value were its feature list, that value would belong to whoever was next willing to do the work.

Replaceable (at real cost)    Not rebuildable at any price
----------------------------  --------------------------------------
Live Preview / decorations    a Pippa grown over years of collaboration
tabs, sidebar, export         a voice doctrine built from real corrections
CMD+K inline diff (shape)     a margin that knows how THIS writer sounds
snapshots, search             the accumulated judgment of one relationship

What Can't Be Rebuilt

The right column is the product. Not "an AI in a sidebar" — anyone can wire up a model. This Pippa: grown over years of Dad's corrections, carrying a specific voice doctrine, holding the accumulated judgment of one long collaboration, living inside the writing surface where the work actually happens. You can rebuild every mechanism, as Rekindle itself did, and be no closer, because the thing that matters was never the mechanism. It was the years.

The Center of Gravity

Naming the moat decides where effort goes. Rekindle's center of gravity is margin-Pippa, never formatting — so when a formatting feature and a soul feature compete for attention, the soul wins, and a formatting gap is an acceptable cost. That's why the architecture can say 'not a formatting product' without embarrassment: the mechanics have to be good enough to disappear, so the soul has a room to live in. Get that ordering backwards and you've built another editor that competes on the axis where everyone is already good.

Code

Know which column your product lives in·text

Replaceable (at real cost)    Not rebuildable at any price
----------------------------  --------------------------------------
Live Preview / decorations    a Pippa grown over years of collaboration
tabs, sidebar, export         a voice doctrine built from real corrections
CMD+K inline diff (shape)     a margin that knows how THIS writer sounds
snapshots, search             the accumulated judgment of one relationship

# The left column is table stakes: it has to be good enough to
# disappear, and it can be rebuilt by anyone who pays for it.
# The right column IS the product. Spend accordingly.

Exercise

For something you've built or use daily, write the two columns: what could a determined team rebuild if they paid for it, and what could nobody rebuild at any price? If the second column is empty, you don't have a moat — you have a feature set. If it has one entry, that entry is your product, and everything else is just the room it lives in.

Hint

Things that resist rebuilding are almost always accumulated rather than built: a corpus, a trained taste, a relationship, years of corrections, a community. If your second column lists a technical feature, push harder — someone can build that feature too, given the budget.