[amrx101]

As someone who has wrote Python professionally here is my quirk about it. As soon as performance requirements start cropping up in any Python project, all "easy-ness" of Python starts to fade. You have to start looking up for optimisations left, right and centre. Re-writing your code with C-calls to improve performance.

All these has taught me a thing or two. Now if I know that we will be using the product wherein performance or scalability is required, I straight away opt one of Rust, Golang or JAVA. I dont want recurring efforts on tuning the code-base to get better performance. I know I might get some opinions here, but Python is a scripting language for rapid prototyping and I plan to use it as such.

[sumanthvepa]

Yup. Use the right tool for the job. Python's performance is definitely something that the implementers of Python should worry about, but as a user, you have choices and you should exercise them. If your problem requires relatively high performance: Java, Go and Rust are all good alternatives. If absolute pedal-to-the-metal, no compromises, performance is needed: C, C++ are your choices. Heck, you can even inline assembly in some cases.

[Areading314]

This is especially true when writing code that needs to run on many cores -- Python starts getting really cumbersome in this case, which can be avoided using the relatively elegant threading models in alternative languages.

[jashmatthews]

Task based parallelism for a web server or background job server is really easy and that's 99% of the cases covered. Problems which _need_ full multi-threading with shared mutable state are actually pretty rare.

[Areading314]

Agreed -- but they do happen!