Should I learn C or C++ to understand low-level programming?

I’d start with C. It has fewer abstractions and tends to make pointers, memory layout, compilation, and the relationship between source code and assembly more visible. C++ can do everything from low-level embedded work to large applications, but its features and complexity can hide what is happening underneath. Once you understand C, moving into modern C++ will be easier. Just remember that C++ should generally be written using its own idioms rather than treated as C with extra features. Neither language alone will teach you computer architecture, though, so learn to read assembly and compare your C code with the generated disassembly.

C is a useful starting point because it is widely used in operating systems, embedded software, and libraries that other languages depend on. It also provides useful historical context: many systems languages were designed partly to address limitations or safety problems in C. After learning the fundamentals, try modern C++ and compare its resource-management tools, abstractions, and object-oriented features with the manual approach in C. You don’t need to wait two or three years before starting the second language—working on focused projects will teach you more efficiently.

If your main goal is learning how computers work, you can begin with a structured hardware course rather than waiting until after learning C or C++. Build simple logic circuits, study CPU instructions and memory, and then write an emulator or small kernel. C will be useful for the kernel, but the architecture and operating-systems material is what provides the deeper understanding. Also consider Rust later, especially for systems programming, while recognizing that C and C++ remain important in existing codebases.

Don’t treat the language choice as a substitute for studying computer science. C and C++ abstract away plenty of hardware details, so use computer architecture, operating systems, digital logic, and processor documentation to learn how computers actually work. Projects such as building a simple CPU, emulator, or tiny operating system can connect those subjects together. Assembly is also essential if you want to understand a specific processor architecture.

Your list covers several different areas, and they won’t all use the same language. Operating-system and embedded courses often use C, C++, or Rust, while search engines and many graphics or game projects may benefit more from Java, C#, or another higher-level language. Data structures and algorithms are a more important foundation than spending years studying one language in isolation. Learn enough of a language to build real projects, then let each project introduce the next concept.