That's my hypothesis: _that we have yet to create the "best" programing languages_ we can and I'd expect AI to be able to do so faster than we do.