How do I stop tutorial hell and start building my own projects?

Tutorial hell isn't a knowledge problem, it's a comfort problem — tutorials feel productive because they're smooth, and building feels like failure because it's full of being stuck. The way out is deliberately choosing the uncomfortable one, at a size small enough that you actually finish. The escape plan that works: 1. Build something absurdly small this week. Not a clone of a big app — a tip calculator, a page that shows a random quote from a list, a countdown timer to a date that matters to you. Small enough that finishing is realistic, because finishing is the habit you're missing, not complexity. 2. Follow one tutorial, then immediately rebuild it from scratch without the video. Same project, blank editor, look things up as needed. This single exercise is the highest-leverage move available to you — it converts passive recognition into active recall, which is the actual gap. 3. Then modify it. Add a feature the tutorial didn't have. Change the design. Add local storage so it remembers. Every modification forces genuine problem-solving because no video covers it. 4. Only after that, build something original at the same size. What to do when you freeze at the blank editor, because you will: - Write the steps in plain English first, as comments. 'Get the input value. Check it's a number. Calculate the tip. Show the result.' Then fill in one comment at a time. The freeze is usually your brain trying to solve logic and syntax simultaneously; separating them removes it. - Build the ugliest possible working version first. No styling, no edge cases, hardcode things. Get one path working end to end, then improve. Perfectionism at step one is a major cause of the freeze. - Give yourself a rule: 20 minutes of genuine attempt before looking anything up, and when you do look, search for the concept ('how to read an input value in JavaScript'), never for the whole solution. The reframe that helps most: being stuck is not evidence you didn't learn enough. It's what the job is. Professional developers are stuck constantly — they just have a process for it and don't interpret it as failure. Every hour you spend stuck and eventually unstuck builds more capability than five hours of following along, precisely because it's harder. And set a hard ratio if you need structure: for every hour of tutorial, two hours building without one. Not forever — just until the blank editor stops being frightening, which usually takes three or four finished small projects.