[holdenweb]

I wouldn't say Python is good _at_ vectorization so much as good _with_ it. It's also good with AI, web systems, etc, as a result of its "play well with others" philosophy, which pragmatically accepted from the start that no one language could be best for all purposes.

One of the best improvements to Python's memory model was extending it to allow sensible access to blobs of memory containing regular structures such as vectors and matrices, making it relatively easy to interface with and control already-available, well-tried, highly-optimised algorithms.