[fulafel]

Using a lot of execution resources is not a direct benefit from the user POV.

Making programs execute faster could be. And there using many cores is one implementation strategy among many complementary ones. But the current landscape of programming languages shows that languages focusing on execution speed and parallelism have smaller user bases.

Also, re this data interpretation of Python's killer app, this usage of Python is something fairly recent and happened after Python had already broken into mainstream usage in the industry. A lot of small and big applications have been built using it before the data and ML things got in vogue (like Youtube, Dropbox, Spotify etc at the bigger end).