Python 3.14 Will be FASTER than C++ ?!

What are you thoughts on this?
In this video I am going over an article which is showing the speed performance improvements of Python 3.11 and compares it with previous versions. Also there is a future prediction to Python 3.14.
We are also doing a quick digression over how to estimate Pi using Monte Carlo methods.

Have a good weekend everyone! :-)

Article from Dennis Bakhuis:

Visualized Estimation of Pi:

[farbhy]

Lol. Can't wait for python 3.20 to be faster than causality limit itself 😅

[cde]

As a lot of python backend is actually c++ so what the author is writing about is not a surprising development. In some way you could think of python as a simplified interface for c/c++. But as such it cannot, by design, being a interpreter language, surpass the performance of the c language family on a given machine. The advantages and the emphasis of python's strengths is just different to those of c/c++

[SkyFly]

By C and C++ extension, Python has the potential to overcome C/C++ languages...

[sinan_islam]

This article was definitely a clickbait for those who believe Python is the best language in the world, but yet they are still searching for an evidence.

[Almeida]

In c++ this monte carlo simulation can be performed at compile time

[halneufmille]

That's a silly paper. Of course just looping, squaring two values and iterating will be fast because everything can be done in the CPU cache with no memory allocation. Use more data and arrays and everything slows 1000 fold.

[Autotrope]

Python 3.15 scripts will finish before they start!

[richardgomes]

Discussions involving performance make sense when we talk about the inner workings in the compiler, how it generates binary code, how optimizations are performed by the code generation stage of the compiling system, things like take. In particular, these discussions make sense when we are comparing the compiler of programming language X versus the compiler of programming language Y. However...

Comparing implementations of a given algorithm in particular, like it is done in this video, misses the point, because we are not interested on any algorithm in particular. We are interested on compiler X versus compiler Y, and it does not matter which algorithm in particular it is.

[hansdietrich]

Using numba, you're probably there already anyways.

[tzimisce]

Python 3.12 was around 4% faster on average than 3.11.
Python 3.13 was only faster than 3.12 in one regard.
But at least PyPy has larger coverage for modules now, and includes the data science libraries.

[ganisack]

maybe we need some low level language and named pythonc++

[Mutual_Information]

Lol yea that is a questionable extrapolation.. but it’s a funny analysis. Gotta give it credit for that.

[bluxmitalnoda]

Following the linear trend, Python 3.15 will do the calculation faster than 0 milliseconds! That is, the result will be computed before you even run the code

[ylstorage]

by interpolation, I would be a Bliionare in a Billion years.

[saeedatenzi]

There is no way in hell that its real. I refuse to believe that the author was remotely serious

[KingLoki]

Literally a guy on here did a speed comparison benchmark with JavaScript NodeJS with python and was 100 times more faster in certain tasks. That guy is delusional.

[mudasirulhaque]

How is he even a data scientist when he doesn’t know the difference between correlation and causation. FYI - python would never even come close to any static typed language let alone cpp.

[arvindshelke]

Bruh, I was really expecting a fucking breakthrough! Anyways

[cenkedits]

Python never be faster then c++ because python made by c and c++

[ButchCassidyAndSundanceKid]

He is bullshitting. As long as the GIL is there, Python will never be faster than C++. It's a bit nothing travels faster than or near the speed of light as long as it has a mass.