The single most expensive habit late-start engineers carry over from school is “I’ll figure it out as I go” — a strategy that worked for a 50-line problem set and fails badly on a 5,000-line system. writing-plans is the discipline of writing what you’re about to do, in prose, before opening the editor. The skill makes that discipline cheap enough to actually use.

What it does

Given a spec or requirements, the skill walks you through articulating: the goal, the constraints, the chosen approach with alternatives considered, the risky steps, the testing strategy, and the rollback plan. The output is a plan document you write to your future self — and to anyone who reviews your work — before code exists.

Who it’s for

PhDs and postdocs entering industry whose academic instincts say “just code it and iterate” — and whose teams expect a plan first
MS holders taking on multi-day tasks for the first time, where pure-instinct execution starts to break down
Career-switchers building the engineering-management instincts they didn’t get in their previous career
Anyone whose past three projects ended up bigger and uglier than expected

What to watch for

Plans rot. Write them, but don’t worship them — when implementation reveals the plan was wrong, update the plan, don’t fight reality
Plans are not specs. A plan is what you are going to do; a spec is what the system will do. Don’t conflate
The skill won’t make you slower. People avoid plan-writing because it “feels” slow. The actual time cost is 30–90 minutes; the cost of a wrong implementation is days

Verdict

The most valuable single discipline a late-start engineer can adopt — and the one most resistant to adoption because it feels like overhead. Write the plan. The plan saves you.