It is hard for me to accept that historical interpretation as I started using Python back in the 1990s, when it was already making in-roads in steering high-performance computing codes. NumPy's roots date from that era.
In 2002 it seemed that half the attendees of the Python conference were there because of Zope.
I was programming full-time by 1999.
So for me Python was well-established in several areas far before its wide-spread use in computer programming education or its use in data science.
I never had any exposure to HPC, scientific or numerical computing, even through university and in my career - that world is still comparatively silo'd off from the wider dev ecosystem IME; I know you are correct in what you say, but I imagine Millennials like myself (who were still in middle-school when you were using Python professionally) only ever saw Python in less serious applications.
In 2002 it seemed that half the attendees of the Python conference were there because of Zope.
I was programming full-time by 1999.
So for me Python was well-established in several areas far before its wide-spread use in computer programming education or its use in data science.