Mr Data, sure there are ways around the slowest parts of Python; and sure Python is the language for ML at the moment. But that doesn't mean that we can't do better. Python's creators know this :)