A functional prototype is worth a thousand GDD words.

This means that it might be better to invest your time and work on creating a prototype that implements your game idea (or feature) instead of putting a lot of work upfront by writing down a GDD.

You can validate your game idea/feature when you have something that people can play, test and give feedback.

It is much more difficult to validate an interactive experience by just reading a document, since people can interpret it in different ways.

At the end of the day, game development is all about people [besides making money]:
You spend most of your days working with a team, which becomes sort of a family. There are great times and times when you all struggle to work out the differences and ship the game.
You connect with other developers on social networks. Go to events and jams. Build a community. Market your game.
Your game talks to players and they talk back. It makes them happy, angry, sad. And players talk about your game to other people.

For their final [gamedev course] project, I ask my students to think about the following:

-What are my goals with the project?
-Which area do I want to pursue?
-What can I do?
-[Maybe more important,] what can't I do? (my constraints/must learn stuff)
-How many hours/week can I allocate to the project?
-How can I communicate with the team?
-How will I publish/make money with it?
-Did I check the project triangle? (fast, good or cheap, pick two)

I believe these questions help when starting a project.

It's dangerous to go alone!
Even if you are creating a game all by yourself, don't go alone on this journey. Find a fellow game developer to join you. You two don't need to work on the same project; the idea is to support each other: talk to your friend everyday (or at least a few days a week), explain what you're doing in your project, complain about that feature/bug that got you stuck for days, share that cool video/article you found the other day. The path will be easier and more enjoyable. Try it!