Popular posts

That depends on various things:

• What language are you learning? You might want to do something parallel when learning Erlang, something data science for R (or Python) and something low-level for C.
• Why are you learning it? That's probably the most important question. If you want to learn to make a web framework, don't build a 3D graphics engine for practise.
• What do you like to make? Especially if you're doing it in your spare time, you'll want to stay motivated, so do something you like.
• What is different from what you know? Say you're learning Kotlin. If you're coming from Rust you'll want to focus on inheritance and polymorphism, while if you're coming from Python you'll want to focus on static typing.

What's most important, I think, is that whatever you work on, you stop and think about what's going on. You should certainly not stop as soon as it works, you should understand why it works, and preferably ask yourself "but what if I ...".

In general I find that trying to abstract whatever you make into a reusable framework tends to push you towards harder things. Maybe consider that for the second project. Maybe this is the language that's finally going to solve ORMs?

And finally, if you really want to understand the differences between languages, maybe make your own, better language. It'll never go mainstream and it'll keep getting harder the more you think about it, but you'll learn a lot.