Python won because Google picked it over perl in the early 2000s, declaring that it was more uniform in having one preferred way to do things rather than allowing (in contrast to Perl) programmers having a zillion ways to do things.
Larry Wall, creator of Perl, coined the phrase, "make the easy things easy, and the hard things possible". But the things perl made easy were not always things that the market needed. Perl was very distrustful of enterprise disciplines and the perl culture "let a thousand flowers bloom" mentality (which they/we called "There's more than one way to do it"/(TMTOWTDI)) was not actually useful for organizations maintaining larger code bases/systems with large groups of mixed skill level people.
Python threaded that needle of centralizing a right way to do things (not just whitespaces) more effectively.
I also observed Scala suffering from similarly disfunctional "TMTOWTDI" dynamic in organizations I belonged to; different people had wildly different styles of solving even basic problems and it really made code hard to maintain even in smaller organizations where people came and went and new people were onboarded frequently.
Plus data science use cases saved Python's web/scripting lifespan in a way slightly analogous to AI/CUDA extending NVidia's 3D graphics-centric roots. (Although the python data science ecosystem nowdays does have some TMTOWTDI weaknesses that remind me of perl/cpan's choices of web development frameworks of the mid-2000s!)
Larry Wall, creator of Perl, coined the phrase, "make the easy things easy, and the hard things possible". But the things perl made easy were not always things that the market needed. Perl was very distrustful of enterprise disciplines and the perl culture "let a thousand flowers bloom" mentality (which they/we called "There's more than one way to do it"/(TMTOWTDI)) was not actually useful for organizations maintaining larger code bases/systems with large groups of mixed skill level people.
Python threaded that needle of centralizing a right way to do things (not just whitespaces) more effectively.
I also observed Scala suffering from similarly disfunctional "TMTOWTDI" dynamic in organizations I belonged to; different people had wildly different styles of solving even basic problems and it really made code hard to maintain even in smaller organizations where people came and went and new people were onboarded frequently.
Plus data science use cases saved Python's web/scripting lifespan in a way slightly analogous to AI/CUDA extending NVidia's 3D graphics-centric roots. (Although the python data science ecosystem nowdays does have some TMTOWTDI weaknesses that remind me of perl/cpan's choices of web development frameworks of the mid-2000s!)