How can I make programming fun again without rejecting AI?

A structured course or book can provide a good balance. Learn a language that seems interesting, work through the exercises, and then build a small project without immediately outsourcing the hard parts. When you need help, consult documentation, search for explanations, or ask AI about a narrow issue. That preserves the useful struggle without turning every problem into an all-night debugging session.

Struggling on purpose is not necessarily pointless if you are doing it to understand the material. If AI gives you code, compare it line by line, ask why each choice was made, consider alternatives and tradeoffs, and then rewrite parts of it yourself. That turns an answer into a learning exercise instead of a shortcut.

Try choosing a niche project with unusual requirements or poor documentation. AI and search tools are less likely to provide a complete solution, so you will naturally have to investigate, test, and debug things yourself. You can still use modern tools when they save time, but the core discovery process remains yours.

Tune out programming-content social media and spend more time actually building things. Conference talks, technical books, and long-form courses tend to be much more inspiring than algorithmically optimized videos. There are still plenty of enthusiastic programmers, but the most interesting ones often are not making flashy content. Use AI or search when you are genuinely stuck, but do more coding than watching other people talk about coding.

You do not have to recreate the exact feeling you had years ago. Choose a project or subject that interests you now and see what programming feels like at this stage of your life. AI can be useful as an assistant without becoming the author: ask about a specific algorithm, request examples, or have it review your work, but write the implementation yourself when learning is the goal.