When should I stop struggling and look up a programming solution?

Use a progression instead of jumping straight to the complete solution. First identify whether you're dealing with an error, a missing concept, or a logic problem. Read the exact error, check the official documentation, isolate a minimal example, and use a debugger or print statements to inspect what the program is actually doing. If the logic is unclear, break the task into smaller pieces and write down what each piece should accomplish.

Try explaining the problem line by line to an imaginary beginner or writing a detailed question without necessarily posting it. This is basically rubber-duck debugging, and the act of explaining your assumptions often reveals the mistake. If you still need help, ask for a hint or for the underlying concept rather than requesting the finished answer. That preserves more of the learning process.

Looking something up isn't automatically harmful; looking up the exact answer too early is usually the problem. Search for narrow concepts, documentation, or error messages instead of searching for the solution to the entire exercise. If you use an AI assistant, ask it to question your reasoning or provide one hint at a time, and verify anything it says against reliable documentation.

There isn't a universal time limit. Before searching, try changing your approach: take a break, write the problem down, solve parts of it on paper, and describe the algorithm in plain language or pseudocode. A fresh perspective often helps more than staring at the same code. Experienced developers can spend days or weeks on genuinely difficult problems, so don't assume every solution should appear within thirty minutes.