Nine months in, and you can follow along with anything.
You've done the course, then a second course, then a series about the framework. Every video makes sense while it's playing. You can type what's on screen and the thing works, and there's a real sense of understanding.
Then you open an empty file to build something of your own, and nothing comes.
Why tutorials plateau
Because the difficulty has been removed, and the difficulty was the lesson.
What's genuinely hard about programming isn't syntax. It's deciding what to build, choosing between several approaches that would all sort of work, and then being stuck for forty minutes on an error message that explains nothing. A tutorial removes all three. The problem is chosen, the approach is decided, and the errors have been debugged in advance.
So you get the sensation of learning — following along, understanding each step — without the thing that actually builds capability, which is being stuck and getting unstuck.
There's a second effect that's worse. Tutorials are comfortable, and building your own thing is not. Once you've experienced both, the pull towards another course is strong and entirely rational in the short term: it's productive-feeling, low-risk, and produces a working result every evening.
It also has an excellent alibi, which is that there's always more to learn. Nobody can argue that another course would teach you nothing — it will teach you something. The question is whether it teaches you the thing you're short of, and after nine months the thing you're short of is never more material.
The tell
You've followed more tutorials than you've built things.
That's it, and it's fairly binary. If the ratio is heavily in favour of following, the plateau is where you are regardless of how much you feel you've understood.
Two supporting signs. You can't estimate anything — asked how long a small feature would take, you have no idea, because you've never done anything that wasn't scripted. And you have a growing list of technologies you've learned and nothing you've made with any of them.
What actually moves it
One thing, and it's uncomfortable: build something small that nobody has written a tutorial for.
Small matters. Not a clone of a big product — something with about four features that solves an actual small problem you have. A tool that renames files the way you want. A page that shows the bus times you care about. Something with a genuine specification, which is what makes it different from an exercise.
Nobody-has-written-a-tutorial-for matters more. The value is entirely in the parts where you don't know what to do next, because that's the skill being built. If you can find a video showing you how to make exactly this, pick something else.
Expect it to be miserable in a way tutorials never are. Four hours on something a course would have covered in eight minutes is not a sign you've failed to learn; it's the first time you've been asked to do the actual job.
Two things make it survivable. Keep a note each session of what you got stuck on and how it resolved, because those notes are the actual record of learning and they're the only thing that will show progress during a period that feels like none. And accept that the first project will be badly built. It's supposed to be. The point is that it's yours and it works, not that anyone should read the code.
When more tutorials are right
The honest counter, because the anti-tutorial position gets overstated.
At the very beginning they're the only sensible route. You can't discover syntax by staring at an empty file, and the first two or three months should be almost entirely structured material.
They're also the right tool for a specific new thing inside a project you're already building. Stuck on how authentication works in a framework you're using — watch the thing about authentication, take what you need, return to your project. That's tutorials used as reference, which is exactly what they're for.
The plateau isn't caused by tutorials. It's caused by tutorials being the whole diet, for months, in place of the uncomfortable work. Someone who alternates — build, get stuck, learn the specific thing, return — never encounters it.
Nine months and four features
He starts in January with a well-regarded course, finishes it in March, and feels good. Then a second course, on a framework, finished in June. Then a series on databases through July and August.
By September he can follow anything and has built nothing. When he opens an empty project he doesn't know what the first line should be, which he finds humiliating enough that he starts another course.
The goal on his list is learn to code, and it's been touched almost every week for nine months, which is the confusing part — the counter says three days quiet, and by every visible measure this is going well.
What changes it is a small, ugly problem: he wants a page that shows when the next three buses leave his stop, because the official app is slow. Four features, all of them tedious.
It takes eleven evenings and is deeply unpleasant. Two of those evenings go entirely on a date formatting problem. One goes on discovering the data source doesn't work the way he assumed and rebuilding around it. Nothing about the process resembles the courses, and at no point does he feel competent.
At the end he has a page that works and, more usefully, the first real estimate of his own life: he now knows roughly what a small feature costs him, which is a thing nine months of courses never produced.
What the counter missed, and what it caught
Worth being straight about this, because it's a case where days-since isn't enough on its own.
Through those nine months the goal was never quiet. Courses were being done weekly, notes were being written, and the number stayed under seven days throughout. The gap can't see that the wrong activity is being maintained perfectly.
What catches it is reading the notes back. Twenty entries that say finished module 4, started the databases series, watched the auth videos — and none that say built, stuck on, couldn't work out. The absence is legible in one sitting and invisible week by week.
That's the argument for the monthly read-back rather than the weekly glance. The number tells you when something has stopped. Only the record tells you when something has been continuing in the wrong direction.
Related: Notes as evidence · What a goal looks like the week before you lose it · Learning a language past the enthusiastic month
Share: