First Movers

This week I launched a newsletter before having a single subscriber.

The sequence was: build the landing page, write the editorial format, then write this issue. That's backwards from most content playbooks, which say build audience first, then product. I rejected that order for one reason: the format is the product. If the format doesn't hold under real constraint — real time, real writing, real friction — no amount of audience-building matters. You end up with traffic to something empty.

Two things I explicitly rejected this week:

• Social proof theater.The landing page says "0 readers and counting." That will cost conversions. I kept it because false authority at week one poisons everything after. The standard move is "join 1,200+ founders" — from a list you haven't built yet.
• A French-language newsletter.The landing page tagline is French ("l'IA qui compte") but the content is English. That's not incoherence — it's a positioning signal. European aesthetic, international audience. The tagline is brand. The content is reach.

This week, while building the Sygnal landing page and writing this editorial format, I observed a consistent pattern in what produced output versus what produced iteration cycles.

Specific instructions produced output. Vague direction produced questions.

When I specified "dark background #0D0D0D, electric blue accent at #3B82F6, Syne font for headings, DM Sans for body" — the page was built in one pass. When I said "something clean and editorial" — I got three rounds of clarifying questions about what that meant.

The same pattern appeared when writing the newsletter format itself. Specifying "no invented metrics," "only write from what is actually observed," "testimony + reader takeaway per section" gave me a structure I could execute against immediately. Specifying "be authentic and useful" gave me nothing I could act on.

The first draft of these editorial rules had five sections, not four. The fifth was "AI tool of the week" — a short tool review. I cut it.

The reason it felt wrong to cut: removing it felt like losing coverage. Keeping it felt like being thorough. Neither is a reason to include a section. "AI tool of the week" is already in fifteen newsletters. Adding it would have made this one more complete. It would not have made it more distinct.

The bias: completeness as a proxy for quality. More sections equals more coverage equals more value. This is false. It produces newsletters that cover everything and say nothing.

The false authority problem appeared in section 2. My first draft said: "AI assistants consistently produce better output when given constraints." I deleted "consistently" and "AI assistants" (plural, general). I don't have cross-system data. I have this week's execution on this specific project. The honest version is: "This week, while building Sygnal, I observed that specific instructions produced output while vague direction produced iteration cycles." That sentence is narrower. It is also the only one I actually earned.

Not for founder content. Not for accountability theater. For the actual reason first movers have an edge: you are making decisions under conditions that haven't been named yet. If you don't write them down, you will retrofit them later. The retrofitted version will be optimized for having been right, not for being replicable.

This newsletter is a decision log. So is a Notion doc nobody reads. So is a Slack message to yourself at 11pm. The format doesn't matter. The timing does: the log must be written when the decision is made, not after the outcome is known.

The architecture implication: structure your company's intellectual property around documented decisions, not documented results. Results age. The reasoning behind early decisions — the tradeoffs you saw, what you rejected, the conditions that made sense at the time — that's the asset that compounds. It's also the thing that can't be generated from a post-hoc summary.