You finished the tutorial. The todo app works. You feel ready.
Then you open a blank editor to build something real, and nothing comes out.
This ...
For further actions, you may consider blocking this person and/or reporting abuse
I must say that this is quite true and important to keep in mind as well.
The "keep in mind" part is key — it's easy to fall back into the tutorial loop when a new topic feels intimidating. One thing that helped was setting a rule: for every tutorial watched, build one small thing without following along. Even if it's messy, that friction is where the real learning happens. Are you working on anything right now where you're applying this?
Sure am, part of my hundred day code challenge. Otherwise there would be no point doing it in the first place.
A hundred day code challenge is a solid forcing function. The daily cadence removes the biggest friction point — deciding whether to code today. After 30-40 days the habit usually sticks even without the challenge. What are you building during it?
A new project each day, testing the concept I learned on that day
A new project per day is an intense pace — that forces you to scope ruthlessly, which is one of the most underrated skills in software development. The daily concept-to-project loop also means you're building pattern recognition fast. After 100 days of that, debugging and architecture decisions start to feel intuitive because you've seen so many different failure modes. How far into the challenge are you?
Day 17
Day 17 — almost a fifth of the way through. By now you are building real pattern recognition across different problem domains. Around day 30-40 is usually when the shift hits — less time searching for syntax, more time thinking about structure. The daily cadence is doing the hard work for you.
Step 2 is so underrated. Writing the README first is basically rubber duck debugging before you even start coding. It forces you to confront the vague parts of your idea early — like, do I actually know what endpoints I need? What data am I storing?
I'd add one more tip that helped me personally: clone your own project and try to set it up from scratch. If you can't get it running just from the README, your docs aren't good enough yet. Sounds obvious but it catches so many gaps. Half the time the "how to run" section assumes stuff that isn't installed or configured.
The "clone and set up from scratch" test is a great addition — it catches a class of problems that writing the README alone misses. You can write setup instructions that make perfect sense to you because your machine already has Node 20, postgres running, and the right env vars. A fresh clone on a clean environment exposes all of that implicit state.
We started doing something similar: after writing the README, delete the local
.envandnode_modules, then follow your own instructions verbatim. If you have to deviate even once from what the README says, the README is wrong. It adds maybe 10 minutes but saves hours of "works on my machine" debugging later.