NumPy
Python library for fast N-dimensional arrays and numerical computing.
HN commenters value NumPy as the array foundation of Python's open source data science stack12, a less painful alternative to MATLAB for scientific work34, and a fast way to do brute-force vector and matrix math.567 Commenters note that importing it takes about a second on an older laptop8 and that it offers no easy path for non-array structures like trees.9
Why people recommend it
Ecosystem foundation · 3
Seen as an essential core of the Python data science stack alongside SciPy, pandas and matplotlib.
“numpy/scipy/matplotlib/pandas/plotly/polars/pyarrow/jupyter, which is essential to my work”
“the open source ecosystem is great and matlab would be my go-to tool if they would not exist”
Show 1 more
“I love how you stuck with pure Python libraries - tkinter, numpy, and sounddevice.”
Numeric speed · 3
Brute-force NumPy computation beat a vector database and was fast for averaging embeddings.
“The brute-force NumPy solution is a significant improvement, making it faster than other calculations like clustering.”
“I have used numpy as well as built in postgres functions to average”
Show 1 more
“I use numpy automatically when simple arrays are involved”
MATLAB alternative · 3
Former MATLAB users found it less awful for lab code and graphing.
“Numpy was less awful than Matlab”
“the open source ecosystem is great and matlab would be my go-to tool if they would not exist”
Show 1 more
“I was using Mathematica or NumPy/Sympy quite a bit.”
Well liked · 2
Commenters, including a past contributor, say they simply love it.
“this is cool and I love numpy anyway lol”
“I love how you stuck with pure Python libraries - tkinter, numpy, and sounddevice.”
Caveats and dissent
Slow import · 1
Importing NumPy takes about a full second on an older laptop.
“They're both about a full second on my old laptop.”
Arrays only · 1
Helps with simple arrays but offers no easy path for tree traversal.
“I use numpy automatically when simple arrays are involved”
C templates · 1
A contributor found the codebase's C template system surprising and wished it used C++.
“this is cool and I love numpy anyway lol”
PyTorch preferred · 1
At least one commenter prefers PyTorch over NumPy.
Often compared with
Who recommends it in their own words
Threads where it came up
- Tracing np.add, all the way down 1 mention
- Python's pre-declared constants are kinda weird 1 mention
- Show HN: CLI tool for detecting non-exact code duplication with embedding models 1 mention
- Cleve Moler (Matlab, MathWorks) passed away on May 20, 2026 1 mention
- Python: The Optimization Ladder 1 mention
- Inside the M4 Apple Neural Engine, Part 1: Reverse Engineering 1 mention
Data: numpy.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 13
“this is cool and I love numpy anyway lol”
“Numpy was less awful than Matlab”
“The brute-force NumPy solution is a significant improvement, making it faster than other calculations like clustering.”
“the open source ecosystem is great and matlab would be my go-to tool if they would not exist”
“I use numpy automatically when simple arrays are involved”
“I also use numpy for large correlation matrices, covariance etc.”
“my very first attempt used numpy.searchsorted”
“I love how you stuck with pure Python libraries - tkinter, numpy, and sounddevice.”
“I was using Mathematica or NumPy/Sympy quite a bit.”
“They're both about a full second on my old laptop.”
“I prefer PyTorch over NumPy”
“numpy/scipy/matplotlib/pandas/plotly/polars/pyarrow/jupyter, which is essential to my work”
“I have used numpy as well as built in postgres functions to average”