Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: How did Python become the lingua franca of ML/AI?

news.ycombinator.com

Post reply on HN
Print view View original

31–40 of 99 posts

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#31

Post by aasasd »

karmasimida wrote: post #18

Python was already the lingua franca before the whole deep learning/AI thing. It has numpy/scipy/pandas/scikit-learn, etc. And when did numpy happen? It was in 1996. Arguably its biggest competitor then was R, but R is not well accepted by programmers. Yet another alternative is Matlab, but OMG, using matlab for anything string related is killing me. While there is some history to it, Python won in the end isn't a su…

Yup, it was already everywhere before ‘data science’ was a thing. Turns out, people in actual fundamental science want to slap calculations together fast instead of fussing with memory and whatnot.

I have a friend who writes C or C++ (can't remember which) for clusters processing data from particle accelerators, but he will still reach for Python when he wants anything simpler than that—and, I guess, less interactive than Matlab.

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#32

Post by mynameisash »

MattGaiser wrote: post #8

My last job was at an ML company. Most ML people there cannot build large robust systems and some struggled with the non-algorithmic bits of software. I am sure that some can out there in the world, but for the most part our ML people were very good at creating models and not very good at the development part, especially as the program grew (part of the motivation to hire devs like me in the first place). Python gets…

Your description is pretty much spot-on.

I'm a data engineer at FAANG. I love the data scientists I work with. They are, generally speaking, crazy smart and simultaneously humble about their (in)ability to write code - they're highly specialized in ML, not so much SWE. I therefore have generally good job security working in operationalizing and optimizing their code. (Recently tweaked a script a scientist wrote and dropped runtime from 5hrs to 5min.)

If I never saw another line of Python again, I'd probably be quite happy. But I think they - and even some ML engineers I work with - love them some Python precisely because they can mash the keyboard a bit and take the shortcut to the finish line. And I don't begrudge them this at all - good on them that they can get their job done quickly! But it's a pain to make things stable and efficient.

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#33

Post by auntienomen »

Python's a lingua franca in AI/NN because it was already a dominant language in scientific computing. Its dominance in scientific computing grew steadily through the 1990s and 2000s, for a few reasons:

1) Python -- specifically CPython -- made it easy to wrap existing, thoroughly tested high performance libraries in Python APIs. So, you got easy access to things like GSL and BLAS and LAPACK, but you get to call numpy.linalg.svd instead of GESDD.

2) Python was a general purpose language, unlike R or MATLAB, so you could extend existing systems to do more without running into a wall.

3) Python was a heck of a lot less effort to use than C++.

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#34

Post by danwills »

dragonwriter wrote: post #25
perth wrote: post #13

Earlier quoted context omitted.

The most insane thing about Python is how you can override single methods in classes and use the class like normal. One time I was working on getting FIFO working on Windows, and none of the Python built-ins were set up to handle any random process writing to a named pipe that wasn't within the same Python instance. So what I did was I took the closest implementation Python offered, which was in the multiprocessing m…

> The most insane thing about Python is how you can override single methods in classes and use the class like normal. Isn't that true of basically every language supporting class-based OOP and inheritance?

It's called 'monkey patching' and python does make it particularly easy, simply:

class.methodName = newMethod

.. kinda thing, future callers now get your method instead of the original.

This does seem a fair bit easier than other languages make it to do?

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#35

Post by kragen »

chubot wrote: post #20
chubot wrote: post #6

Because Python has NumPy, which implements vectorized math on arrays and matrices. Machine learning algorithms are implemented naturally and efficiently with those primitives. PyTorch, TensorFlow, and I think every other machine learning framework in Python all use NumPy. JavaScript, Ruby, and Perl either don't have this abstraction at all, or they have much weaker versions of it, and many fewer scientific libraries.…

I should also add a cultural / social reason why Python is used in scientific computing and machine learning much more than JS/Ruby/Perl: Python was the only one of those languages (partially) funded by government research agencies. Guido was a research programmer in the Netherlands at CWI, and then he moved to the US when he was hired by CNRI, a research agency headed by Bob Kahn (loosely connected with DARPA as far…

Perl was originally written at JPL, which is the epitome of government-funded research, and for the first many years of its life most of its numerous contributors were at one or another government-funded research institution, because people who weren't didn't have internet access.

Lua does support operator overloading:

    $ luajit
    LuaJIT 2.1.0-beta3 -- Copyright (C) 2005-2017 Mike Pall. http://luajit.org/
    JIT: ON SSE2 SSE3 SSE4.1 AMD fold cse dce fwd dse narrow loop abc sink fuse
    > x = setmetatable({}, {__add=function() return 37 end})
    > print(x+5)
    37
Not sure if that was true 20 years ago.

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#36

Post by cameldrv »

cgb223 wrote: post #28
bbulkow wrote: post #22

Another way to analyze the problem: what other language would it have been, given the moment ml hit? You say compared to other scripting languages'. Let's list them. Ruby: no numeric support Go: unnecessary typing, modest numeric support, shitty generics Bash: ha ha ha Scala, java, c, cpp: not a scripting language, complex Tcl, php: out of favor Rust: hadn't happened yet R: in memory bias, not as simple Other languag…

Non ML/AI coder here: Why does ML/AI work need to be written in a scripting language? Why can’t it be something like C++ etc instead?

One reason is that you usually need to try a lot of things before you get something to work. Language productivity is at a premium. You really want some sort of interactive shell where you can do calculations and pull up plots etc. This used to be done with IPython, which evolved into Jupyter.

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#37

Post by friedman23 »

perth wrote: post #13
gonab wrote: post #9

Earlier quoted context omitted.

I don't think this is a fair assessment of "most ML people" Some of the biggest distributed systems built today are used for statistical inference or scientific computation Most "ML people" I know are highly versatile in software, networks and deep hardware knowledge, i.e., essentially they have a very good understanding of what a computer is and what is capable from Its very naive to think that you can assemble mach…

The most insane thing about Python is how you can override single methods in classes and use the class like normal. One time I was working on getting FIFO working on Windows, and none of the Python built-ins were set up to handle any random process writing to a named pipe that wasn't within the same Python instance. So what I did was I took the closest implementation Python offered, which was in the multiprocessing m…

Monkey patching is a terrible practice outside of unit testing and can lead to extremely difficult to debug bugs.

Also monkey patching isn't unique to python.

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#38

Post by perth »

danwills wrote: post #34
dragonwriter wrote: post #25

Earlier quoted context omitted.

> The most insane thing about Python is how you can override single methods in classes and use the class like normal. Isn't that true of basically every language supporting class-based OOP and inheritance?

It's called 'monkey patching' and python does make it particularly easy, simply: class.methodName = newMethod .. kinda thing, future callers now get your method instead of the original. This does seem a fair bit easier than other languages make it to do?

You see exactly what I mean!

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#39

Post by kragen »

threeseed wrote: post #5

I would put it mostly down to Spark. Originally, it was only available in Scala/Java but then they added Python support courtesy of Py4J. And since Python was massively simpler than Scala it exploded in popularity very quickly becoming the default language. So then you had Data Scientists who were already writing a lot of data transformations in Spark looking around at the rest of the Python ecosystem finding librari…

By the time SPARK was born (02010?) Python had already eclipsed the non-JS alternatives (Scheme, Perl, Tcl, Ruby, awk, BASIC, Lush). I'd put the crossover point around 02002. IPython notebooks and pandas came even later than SPARK.

Re: Ask HN: How did Python become the lingua franca of ML/AI?

#40

Post by specproc »

bbulkow wrote: post #22

Another way to analyze the problem: what other language would it have been, given the moment ml hit? You say compared to other scripting languages'. Let's list them. Ruby: no numeric support Go: unnecessary typing, modest numeric support, shitty generics Bash: ha ha ha Scala, java, c, cpp: not a scripting language, complex Tcl, php: out of favor Rust: hadn't happened yet R: in memory bias, not as simple Other languag…

R is definitely a good language for quant work. In some ways it could have been the natural choice, and there are still places where it's a better choice than Python.

It's just far to fragmented, only really good for numeric work (and thus harder to integrate with production systems), and full of the weirdest gotchas.

https://www.burns-stat.com/pages/Tutor/R_inferno.pdf

Post reply on HN