On Fri, Feb 6, 2026 at 10:23 AM David Cournapeau via NumPy-Discussion <
[email protected]> wrote:
> Hi,
>
> I know there has been discussions in the past on AI-generated
> contributions. Is there a current policy for NumPy ? E.g. do we request
> that contributors are the "sole contributor" to the written code, or do we
> allow code written by AI as long as it follows the usual quality
> requirements ?
>
> Context of my question: ~18 months ago I started in some spare time
> writing an ARPACK-replacement in numpy/scipy during scipy sprint. At that
> time, I used ChatGPT for the "research part" only: literature review,
> explain to me some existing BSD implementation in Julia for points I could
> not understand. I implemented the python code myself. There is still quite
> a bit of work needed to be a viable replacement for ARPACK.
>
> Seeing the progress of the AI tooling in my team at work, and how I myself
> use those tools for other hobby projects, I believe I could finish that
> replacement very quickly with those tools today. But I don't want to
> "taint" the work if this would risk the chances of integration into scipy
> proper.
>
> Thanks,
> David
>
>
The following ranking of project difficulty for AI is interesting
(generated by Grok):
SWE-bench Verified draws from 12 popular Python repositories. These cover a
nice spread of complexity levels:
- Lower-to-mid complexity (easier end of the benchmark):
- django (web framework) — large but very modular, lots of high-level
logic
- sympy (symbolic math) — math-heavy, but often localized changes
- pandas (data frames) — complex internals, but many issues are
API/UX or performance tweaks
- scikit-learn (ML) — algorithms + estimators, moderately tricky
- Mid-to-high complexity:
- matplotlib (plotting) — mixes Python + some C extensions, rendering
logic
- sphinx (docs) — build system heavy
- High complexity (harder end):
- numpy — core numerical array library with heavy C/C++/Fortran
extensions, low-level memory management, ufuncs, dtype system,
broadcasting
rules, BLAS/LAPACK integration
- scipy (scientific computing) — builds on NumPy, dense linear
algebra, sparse matrices, optimization, signal processing — very
math-intense and performance-critical
- pytest (testing framework) — meta, plugin system, very tricky edge
cases around fixtures, parametrization, async
NumPy (and scipy) consistently rank among the harder repositories in the
SWE-bench collection for both humans and AI agents.I was surprised by
pytest. PyTorch and Jax rank higher for complexity, Jax in particular is
difficult for its functional programming style, no side effects. The very
high complexity projects are not listed -- I only asked about NumPy.Chuck
_______________________________________________
NumPy-Discussion mailing list -- [email protected]
To unsubscribe send an email to [email protected]
https://mail.python.org/mailman3//lists/numpy-discussion.python.org
Member address: [email protected]
