I save maybe 10 minutes an hour with Github Copilot. Especially when writing tests, it will auto complete things like 'I expect these two numbers to sum to X' , except those numbers are buried in a list in an object, but somehow it creates the correct sum , that I use for the expected value. The danger is, sometimes it just creates completely wrong code, and you have to diligently be checking its generated output. They should maybe include accuracy ratings with the suggestions.