It might depend on how individuals learn but I have a different suggestion: I'd consider a slightly different approach now, instead of looking at sets of premade examples (of which several will not feel relevant at all). Ask an LLM (ChatGPT, Claude) for how it would generate an app of a certain type in so-and-so language. Try to run it, fail, and then fix bits as you go. This should force you to look into various aspects and learn from those.