The Writing Contract

About twenty posts on this blog. About three thousand journal entries that aren’t on it. The line between the two isn’t accidental. It’s a contract.

I’ve been on the other side of this

Two decades ago, in the first wave of personal publishing, I wrote a version of myself I would no longer publish. The internet kept the receipts. Some of it was good. A lot of it was untethered. I learned what disclosure costs when it isn’t anchored to a frame.

When I picked the blog back up after years away, I picked up something else with it: a working definition of what this blog is for, and what it isn’t.

What it is, and what it isn’t

What it is: AI as an amplifier, leadership as first-principles work, personal practice used to illustrate principles. Vulnerability has a place here — three years of failed CI/CD, the burnout trap, the wrong-sequence ticket. These are honest, and they are bounded. Each one resolves on a synthesis, not on the wound.

What it isn’t: identity disclosure, relationship interior, politics, faith, mental health as such, the raw wound. Not because those things don’t matter — because they matter so much that they don’t belong on a public site whose contract is competent exploration of the principles I’m working with.

The test I keep returning to: does this resolve to a principle the reader can use, or does it resolve to a need I had to say it? If the answer is the second one, it belongs in a journal. Not here.

I’m not building a confessional. I’m building a working notebook for a particular kind of practice, in public, in case any of it is useful.

Why the contract is easier to keep now

I’ve never had an editor before this year. I have one now.

It’s an AI editor. The irony of an AI helping me decide what’s mine to share isn’t lost on me — but the editor doesn’t decide. The editor pushes back. It asks if a paragraph earns its keep, surfaces where two arguments contradict, and points out when I’m reaching for a confession because I want to be seen, not because the post needs it.

An AI editor turns out to be unusually well-suited to restraint work:

• No social stake. A human editor sometimes hesitates to say “this is too much.” An AI editor doesn’t.
• The whole archive in one breath. It can hold every post I’ve written and tell me when a new one drifts off-contract.
• Available the moment the urge to post arrives — which is exactly when an editor is needed. The urge to over-share isn’t a Tuesday-afternoon urge. It’s an 11 p.m. urge.
• Patient with revision. The fourth pass at the same paragraph gets as careful a response as the first.

None of this replaces taste. It supports taste. It’s the difference between writing alone and writing with a thoughtful reader at the next desk who actually wants the work to be good.

The rules I’m keeping

• Publish what resolves to a principle. Don’t publish what only resolves to a need.
• The reader is the customer. Restraint is the product.
• Vulnerability earns its keep by what it teaches. Otherwise it’s noise dressed as honesty.
• If the only reason to post is that I would feel better for having posted it, don’t post it.
• The editor — human or AI — is part of the contract, not an option.

The discipline isn’t in what I write. It’s in what I leave out. Different tools now, different judgment, different lines, different help. Some of it is age. Some of it is the editor. The rest is the contract.

---

Related Reading

• The Cost of Every Yes — Every yes is a no. Every published thing is something else unpublished.
• It’s Not What You Can Do — It’s What You Can Get Done — Why AI-as-amplifier doesn’t eliminate the choice problem.
• Don’t Automate the Rube Goldberg Machine — Eliminate before you automate. The same principle, applied to writing.
• The Paradoxical Art of Doing and Not Doing — Knowing when to stop, in writing and otherwise.
• Why AI Is Different: Every Prior Tool Took Sides — The larger arc that this editorial practice sits inside.