Try the idea that could flop. A flop you learned from beats a hit you copied.

A copied success teaches almost nothing, because the important decisions were already made by somebody else and you only executed them. A flop that came from your own reasoning shows you precisely where the reasoning went wrong, which is information you can use next time. One gives you a result. The other gives you a skill.

When you have to choose between the safe version and the one you are curious about, look at what each teaches if it fails. Then take the risk on the smallest scale that still tests the idea honestly. Failing cheap and often is how the useful knowledge gets built.