Tutorial hell is not a discipline problem

The story is always the same. Four tutorials finished. Three of them excellent. You followed every step, you understood every line as it went past, the app worked at the end. Then you opened an empty file to build something of your own and sat there for twenty minutes with nothing.

The conclusion almost everybody draws from this is that they are lazy, or not cut out for it, or need to “just build something” and lack the willpower to.

That conclusion is wrong, and it is wrong in a way that keeps people stuck for years.

Following is not a weak form of building

It is a different activity. Not an easier version of the same one — a different one, exercising different machinery.

When you build something, roughly this happens: you decide what to do next, you choose an approach out of several, you write it, it fails in some way you did not anticipate, you form a hypothesis about why, and you test it. Repeat a few hundred times.

When you follow a tutorial, the first two steps are done for you and the fifth is trivial. You never decide what comes next — the video does. You never choose between approaches, because only one is shown. And when something breaks, the diagnosis is nearly always “find where I diverged from the instructions,” which is a search problem with a known answer, not a diagnosis with an unknown one.

So a tutorial exercises transcription and comprehension. Building requires decision-making under uncertainty and diagnosis from nothing. Doing more of the first improves the first.

The feeling of understanding is not evidence

This is the cruel part, and it is well documented in the learning research: how easy something feels while you are learning it is a poor predictor of whether you will be able to do it later, and quite often an inverse one.

Watching someone competent do a thing fluently produces a strong, genuine sensation of understanding. It is fluency — the material is being processed smoothly — and your brain reads fluency as mastery because in most of life it is. Here it isn’t. The tutorial’s clarity, its good pacing, the fact that nothing is confusing, are all properties of the tutorial.

Which means the better the tutorial, the more confidently wrong your self assessment. People do not get stuck after bad tutorials. They get stuck after good ones.

Why “just build something” fails as advice

It is correct and nearly useless, because it describes the destination rather than the step.

The distance between “follow along with an app being built” and “build an app from an empty file” is not one step. It is a cliff. Told to jump it, most people reasonably fail, conclude they were not ready, and go back to a tutorial — which is the only artefact anyone has offered them. That loop is what tutorial hell actually is. Not a motivational failure. A missing rung.

The rungs exist. Nobody sells them, because they are different for everybody.

What actually converts

Everything below has the same shape: put a decision back where the tutorial removed one.

• Predict before you reveal. Before the instructor writes the next line, say what it will be and why. Being wrong here is the single most valuable event available to you, and following passively means it never happens.
• Break it deliberately. With the finished project working, go and delete something. Change a parameter. Then, before running it, predict the failure precisely — what error, from which line. Diagnosing a break you caused is the training-wheels version of diagnosing one you didn’t.
• Rebuild from memory, a day later. Not from the video. From nothing, with the docs. You will get about forty per cent of the way and stall, and the exact point you stall is the most useful piece of information you have received all week. It is a map of what you actually know.
• Change one requirement. Same tutorial project, but the data comes from a file instead of an API, or it has to handle two users. Small enough to be possible, different enough that no video covers it.

Notice what these have in common: each one is a rung, not a cliff. Each is a problem just past your current edge.

The thing that is actually hard to get

Which raises the question the whole genre dances around: how does anyone know where your edge is?

A tutorial cannot know. It is written once, for whoever arrives, and it must assume something. It nearly always assumes nothing, because assuming nothing is the only choice that never insults a beginner — and that is exactly what makes it useless to you on the fifth one, when the assumption has stopped being true and the material has not moved.

A human tutor solves this instantly. They watch you work for ten minutes, find the edge, and hand you something one step past it. That is most of what makes tutoring effective, and it is the part that has never scaled, because it requires knowing something specific about you before the material is chosen.

If you are stuck, you do not need more instruction. You have plenty. You need somebody to establish what you already know and start one step past it — which is, not coincidentally, the only thing every tutorial you have ever taken declined to do.