JAX
Hacker ApprovedRanked 3 of 19 machine learning libraries by recommenders
Gives NumPy-style array code automatic differentiation and JIT compilation, with fast GPU execution through XLA. Common caveat: the debugging story is rough compared to PyTorch.
Gives NumPy-style array code automatic differentiation and JIT compilation, with fast GPU execution through XLA. Common caveat: the debugging story is rough compared to PyTorch.
55 people gave an opinion
- recommend it
- 49
- criticize it
- 6
- approval share
- 89%
Recommended in 12 of 13 months
- November 2025: recommended
- December 2025: recommended
- January 2026: recommended
- February 2026: recommended
- March 2026: recommended
- April 2026: recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it carries the badge
- 49 people recommend itNeeds 10 or more, met.
- 8.2 recommenders for every criticNeeds 2 to 1 or better, met.
- Recommended in 12 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Fast GPU and compiled execution
“Jax basically gives you the array language without leaving Python, and the XLA backend means you're not hand-tuning C for the GPU path.”
Comment from 15 Mar 2026 in “Python: The Optimization Ladder” “I've been moving to JAX after many years of using torch and XLA is kind of magic.”
Comment from 6 Dec 2025 in “Touching the Elephant – TPUs” “PyMC uses jax under the hood now, so my MCMC workflows are GPU accelerated.”
Comment from 22 Dec 2025 in “A year of vibes” - Comment from 27 Oct 2026 in “Python's experimental JIT to be removed”
- Comment from 4 Dec 2025 in “OpenAI declares 'code red' as Google catches up in AI race”
Familiar NumPy-style API
“Jax basically gives you the array language without leaving Python, and the XLA backend means you're not hand-tuning C for the GPU path.”
Comment from 15 Mar 2026 in “Python: The Optimization Ladder” “it's pretty easy to switch to Jax's NumPy API, and then you can easily jit in Python as well”
Comment from 27 Feb 2026 in “Julia: Performance Tips” “If you're happy with NumPy's API, then surely JAX is exactly what you're looking for.”
Comment from 8 May 2026 in “Mojo 1.0 Beta” - Comment from 20 Aug 2026 in “Mojo is now open source!”
Good for Bayesian and ML models
“Thanks to efficient variational Bayes methods built on top of JAX, it is doable even on high-dimensional models.”
Comment from 22 Mar 2026 in “Bayesian statistics for confused data scientists” “Go with Jax if you want a PyTorch alternative that can have better performance for certain scenarios.”
Comment from 1 Mar 2026 in “Microgpt” “PyMC uses jax under the hood now, so my MCMC workflows are GPU accelerated.”
Comment from 22 Dec 2025 in “A year of vibes” - Comment from 28 Nov 2025 in “TPUs vs. GPUs and why Google is positioned to win AI race in the long term”
Autodiff and JIT compilation
“it's pretty easy to switch to Jax's NumPy API, and then you can easily jit in Python as well”
Comment from 27 Feb 2026 in “Julia: Performance Tips” “From a compiler standpoint JAX is much easier to work with since its all pure functions.”
Comment from 21 Oct 2026 in “Google open-sources TorchTPU, a native PyTorch back end for TPUs” “using JAX for JIT-compilation and automatic differentiation”
Comment from 27 May 2026 in “Fast Robot Kinematics and Dynamics in Jax”
What people warn about
Each mark is one comment
Weak debugging and editor feedback
“the debugging story is still rough compared to PyTorch”
Comment from 19 May 2026 in “PyTorch Landscape” “you don't get feedback as you edit”
Comment from 18 May 2026 in “Futhark by Example” Awkward for some code patterns
“Most of it can be expressed well with JAX's programming model, but the last 10% really sucks.”
Comment from 14 Jul 2026 in “Differentiable Fortran with LFortran and Enzyme” “It's less natural if your basic unit is an element of a vector.”
Comment from 7 Nov 2025 in “Leaving Meta and PyTorch” Dynamic shapes and Numba gaps
“JAX can’t do what Numba can do for example.”
Comment from 9 May 2026 in “Mojo 1.0 Beta” “dynamic shapes were still an issue”
Comment from 12 Mar 2026 in “Show HN: XLA-based array computing framework for R”
Other machine learning libraries people recommend
The four most recommended other machine learning libraries, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
See all 19 machine learning libraries, rankedWhat people said
40 of the 67 comments that mention it, newest first. Each links to the original comment.
“if you want GPU-backed/compiled numpy code then JAX is also a good option”
“From a compiler standpoint JAX is much easier to work with since its all pure functions.”
“For linear algebra and ML with Jax/Numpy I find it quite readable.”
“these other solutions mentioned above are successfully solving for the past ~15y”
“Stuff like Triton, nvidia warp (the language), numba, cupy jax/pallas and so many others really paved the way.”
“Most Jax code is only marginally more restrictive than numpy, and most of the annoyances I have with Jax are shared with numpy.”
“Most of it can be expressed well with JAX's programming model, but the last 10% really sucks.”
“move to Stan or PyMC when you need a real posterior, or to NumPy and JAX when you need to go to production”
“using JAX for JIT-compilation and automatic differentiation”
“I've found JAX's design and feature set to be far more suitable.”
“the debugging story is still rough compared to PyTorch”
“Jax is explicitly mentioned in your pyrefly link as having a parallel (but slightly weaker) system.”
“you don't get feedback as you edit”
“JAX can’t do what Numba can do for example.”
“If you're happy with NumPy's API, then surely JAX is exactly what you're looking for.”
“JAX/XLA/etc. are somewhat similar, though still without user control over transformations”
“JAX is designed from the start to fit well with systolic arrays”
“JAX seems well placed for this by making function manipulation a first citizen”
“Thanks to efficient variational Bayes methods built on top of JAX, it is doable even on high-dimensional models.”
“Jax basically gives you the array language without leaving Python, and the XLA backend means you're not hand-tuning C for the GPU path.”
“Jax seems quite interesting even from this point of view”
“dynamic shapes were still an issue”
“the world would be a better place if we all used JAX instead”
“Go with Jax if you want a PyTorch alternative that can have better performance for certain scenarios.”
“it's pretty easy to switch to Jax's NumPy API, and then you can easily jit in Python as well”
“You can just take Pytorch, JAX, CuPy or whatnot”
“I hope it will allow me to write fast Python code without worrying about these details”
“tons of tools like Numba, Jax, Triton, etc that you can use to write very fast code”
“I think Jax is pretty neat but haven't had the chance to use it in anger.”
“backing really excellent Jax & Pytorch implementations”
“PyMC uses jax under the hood now, so my MCMC workflows are GPU accelerated.”
“I've been moving to JAX after many years of using torch and XLA is kind of magic.”
“JAX, an optimizing backend for it, is in its own league even today.”
“very capable products, that compete very well in ML space”
“exactly the kind of anti-moat everyone has been clamoring for”
“JAX seems great but the Google ghost is a lot to stomach”
“It's less natural if your basic unit is an element of a vector.”
“It consists of small, orthogonal set of ideas that combine like lego blocks.”
“It's actually the most natural way of doing things for me. You just write down your math equations as JAX and you're done.”
“What Matt Johnson has done turning Autograd into JAX is hopefully going to be worthy of as much praise”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.