The Skill I Wish I Had Committed To Two Years Earlier

If I could go back, the skill I would commit to two years earlier is not AI, not frontend, not any stack. It is writing. Specifically, writing technical content in a way that converts trust into work. Every other skill compounded slower than I expected. Writing compounded faster.

Why Writing Beat Everything Else

It is the only skill where every hour produces a permanent asset
It is the only asset that works while you sleep, for years
It is the only marketing that also makes you a better thinker
It is one of the few skills not being automated away

Every other skill has a shelf life. My 2023 Django knowledge is half obsolete. My 2023 articles still get traffic.

The Compounding Timeline

Month 1: nobody reads it
Month 3: colleagues notice, referrals trickle in
Month 6: search traffic begins
Month 12: inbound leads reference specific articles
Month 18: articles start ranking for competitive queries
Month 24: the library itself becomes a sales asset

The curve is flat, flat, flat, up. Most people quit in the flat. The quitters miss the whole return.

Why I Almost Did Not Start

Three reasons I hear from other consultants who almost did not start:

Everything is already written somewhere
I am not a good enough writer
Nobody will read it

All three are true. All three are irrelevant. The value is in the writing, not in the novelty of the ideas. Your articles become sales assets even if they rephrase ideas others had first.

The Habit That Made It Work

Fifteen minutes, every morning, before anything else. Not one hour. Not when I have time. The bar has to be low enough to clear on a bad day. A bad day with fifteen minutes still produces an article. A bad day with one hour produces nothing.

The Stack I Use Now

Apple Notes or iPad for first drafts
A markdown file per article in a local folder
Artifact scripts to publish in batches
The blog platform I built myself (see Phase 0 of the peaklight.ai project)

Deliberately boring. The stack was never the bottleneck.

What I Would Tell Year-One Me

Start tomorrow. Do not plan the blog. Do not pick the tech. Do not design the taxonomy. Open a file called 2024-04-19.md and write four hundred words. Do it again on the 20th. The strategy emerges from the corpus. The corpus does not emerge from the strategy.

Paul Graham's old post on writing and not writing captures the stakes better than I can.