I started using python type hints when you needed to apply patches. This was years ago. At work we had a database that changed it's schema randomly every night.

I started using python type hints when you needed to apply patches. This was years ago. At work we had a database that changed it's schema randomly every night. With type hinting in our build process we were able to determine whether we could handle the new schema. I have also been developing python for about eight years now. I can't really stand the language anymore. It feels to me that unless you have a large python code base, or a large python employee base. It's not just worth the hassle. It was game changing when it first came out. But now,outside of data science, machine learning, and devops I'm not seeing it. With projects like mypy/type hinting, and pypi. You're one step away from using a static language, that is either native or running on a VM (CLR, JVM, BEAM, etc.) Additionally almost all teams I've worked on, have encountered. Either maintainability issues. Where a static type checker would aid. I.E. changing a class or function definition, and not seeing it updated across the board. But more common is performance. Specifically the more asynchronous/parallel work stream. We have asyncio, and others. But it's more complex than some of the static languages. Since type hinting first came on the scene it has been a requisite in any project I've worked on. I've lead a number of teams, who want to move from python. To the tune of we complete each sprint developing the project in python and $LanguageOfChoice. Begging the business to move on, but no dice. Yes I speak in business language why we should move on. But the business won't let us.