What pains me in Python adoption, beyond its use as Perl replacement, is that we have so much better dynamic languages with advanced JIT implementations, but have to reach out writing extensions in native languages instead.
At least Python as DSL for GPU JIT compilers is a thing now.
Yes, I know about PyPy in the corner looking for attention.
> we have so much better dynamic languages with advanced JIT implementations
What are some of these better languages that you're referring to? (The usual dynamic language JITs I hear people praise are LuaJIT and Chez. And V8. And the JVM?)
Smalltalk, Self, whose research lead to Hotspot and V8, Common Lisp.
And the usual, in Python everything is dynamic, well it is even more so in an image based live coding environment, where any break into the debugger, with code changes and resume execution can come back to a complete different world.
Additionally there are features like Smalltalk become: message, where two objects can change places everywhere they are used in the image, and current execution.