Seems like it would really slow you down though if you adopt it too early. Sometimes you don't even know if the thing you want to do is possible. My developmen

Seems like it would really slow you down though if you adopt it too early. Sometimes you don't even know if the thing you want to do is possible. My development style is: - prototype ideas in a quick and dirty dynamic typed language just to gain confidence it's going to work (a lot of ideas die here) - rewrite in a more performant and safe language with a "type checker and a smattering of unit tests" (usually here I'm "done" and have moved onto the next idea/task. If there's an issue I fix it and add another test case) I'm trying to imagine where formal verification comes in. I'm imagining something like: - prototype ideas in a quick and dirty dynamic typed language just to gain confidence it's going to work (a lot of ideas die here) - Formally model the requirements - rewrite in a language that can be formally verified and which is hopefully performant and lets me do things like simd and/or cuda if needed - Never have to fix a bug unless there was a bug in the requirements (?) To me, it just seems like it would take an order of magnitude longer to develop things this way for not much benefit (I've traded development time and potentially runtime performance in exchange for correctness)