How can I turn basic programming knowledge into real projects?

Don’t worry too much about the processor architecture you’re using. ARM is a capable and widely used architecture, and the M2’s limitations are mostly about macOS tools and compatibility rather than ARM itself. You can use virtual machines, cross-compilers, containers, or another computer when you need an x86 environment. For OS development, begin with a very small kernel that boots in an emulator, prints text, handles basic input, and eventually manages memory or simple processes. There are plenty of beginner OS-development tutorials and emulator-based workflows for this. FreeBSD can be useful for learning about a mature Unix-like system, while Plan 9 is historically interesting, but neither is required for learning the fundamentals. Linux and small educational kernels may provide more examples and tooling. More importantly, knowing several languages does not automatically teach CPU architecture, operating-system design, networking, data structures, or debugging. Treat programming languages as tools and deliberately study those underlying concepts while applying them to projects.

Try not to choose a project just because it sounds impressive, such as writing an entire operating system or jumping straight into security research. Pick a narrow milestone: write a command-line utility, implement a small data structure, build a toy shell, create a simple network client, or boot a minimal kernel in a virtual machine. Once one project works, improve it with tests, documentation, error handling, and one new feature at a time. That process will reveal what you need to learn next much more effectively than studying languages in isolation.

The best way to improve is to choose a small problem you actually care about and build a solution, even if it seems simple. For example, you could make a small program that monitors a Bluetooth device and triggers a light or another output when you arrive home. Projects like that force you to learn the pieces you’re missing— APIs, hardware interaction, debugging, and system design—instead of just memorizing language syntax. Start small, finish something, then gradually add features.