There's definitely some truth in the Feynman algorithm. If I had to explain how I did some of my research, I'd probably say it similar to that. While sometimes

There's definitely some truth in the Feynman algorithm. If I had to explain how I did some of my research, I'd probably say it similar to that. While sometimes the manual work to actually show that it works is quite small, other times it may take a month or even longer just to write down the derivation properly. Ideas that come up after step 2, i.e. "think really hard", are ultimately just that: ideas. You need to ensure that you're not fooling yourself or that you have a case of confirmation bias (skipping steps in a proof because you think it's obvious while it actually isn't). Thus, I'd say step 3, "write down the solution", is at least as important as step 2.