Depends on how experimental the idea is. Usually you want to start with something you know will work, and then try out one new idea at a time, each time persisting until you understand the results you're seeing fully.

→ View original post on X — @gdb