# PyTorch 131 Hacker News commenters recommended PyTorch and 47 criticized it, November 2025 to November 2026. It was recommended in 12 of 13 months. It meets the Hacker Approved badge rule. **Verdict:** The de facto deep learning framework, praised for its Pythonic, debuggable design, easy GPU acceleration and huge ecosystem; critics cite installation headaches with GPU/CPU wheels and Python overhead. ## Why people recommend it - Industry and research standard (6 comments): "It is hard to overstate how much Pytorch has accelerated ML/AI development, across the board." ([source](https://news.ycombinator.com/item?id=45850997)) "To increase your chances of anyone actually clicking on your GH link, use good old Pytorch." ([source](https://news.ycombinator.com/item?id=47578747)) "It has, along with Python, become something of a linga franca for research in the field." ([source](https://news.ycombinator.com/item?id=49085999)) - Pythonic, readable, easy to understand (5 comments): "PyTorch felt friendlier and more Pythonic, and it came with a comprehensive library of prebuilt components for deep learning in torch.nn" ([source](https://news.ycombinator.com/item?id=45850406)) "PyTorch is one of those tools that’s so simple and easy to take apart that you feel like you might’ve been able to make it yourself." ([source](https://news.ycombinator.com/item?id=45844131)) "It captures the complexity to allow you to make a whole GPT model with it, running on your GPU, and keep the code readable." ([source](https://news.ycombinator.com/item?id=48678888)) - Simple GPU acceleration (4 comments): "low level libraries like pytorch as well as higher level tools like Ollama all seamlessly support multiple GPUs" ([source](https://news.ycombinator.com/item?id=47543754)) "The value of PyTorch lies more in utilizing accelerators like GPUs while offering a nice abstraction." ([source](https://news.ycombinator.com/item?id=48679300)) "via torch or arrayfire or whatever, you could dispatch it to both CPU and GPU at will" ([source](https://news.ycombinator.com/item?id=46794665)) - Prebuilt components, batteries included (2 comments): "PyTorch felt friendlier and more Pythonic, and it came with a comprehensive library of prebuilt components for deep learning in torch.nn" ([source](https://news.ycombinator.com/item?id=45850406)) "I switched to PyTorch for my next project which was more "batteries included"." ([source](https://news.ycombinator.com/item?id=45844251)) ## What people warn about - Painful installs and missing wheels (4 comments): "you have to install the same package, torch, but from different Python indexes in order to install the cpu version or gpu version, on Linux" ([source](https://news.ycombinator.com/item?id=49905440)) "If you are doing any serious work with torch or local LLMs, you still run into issues where wheels aren't available" ([source](https://news.ycombinator.com/item?id=46782197)) "magic error telling that tensorflow version they are installing is not compatible with pytorch version" ([source](https://news.ycombinator.com/item?id=46399778)) - Overhead and suboptimal performance (4 comments): "when using PyTorch I wanted to try to speed up my NN kernel by 2x by just using half precision and haven't noticed any speedup at all" ([source](https://news.ycombinator.com/item?id=49435279)) "a lot of ML frameworks (Tensorflow, PyTorch, etc) by default use the GPU relatively suboptimally" ([source](https://news.ycombinator.com/item?id=48271421)) "Pytorch is fairly efficient, but python as an ecosystem can have some blind spots or bottlenecks" ([source](https://news.ycombinator.com/item?id=46672831)) - Messy, sprawling codebase (3 comments): "Pytorch is such a maddening mess of half implemented research features in a state of Heisen-deprecation" ([source](https://news.ycombinator.com/item?id=47606533)) "A simple kernel in torch is hundreds to thousands of LOCs, with much worse style." ([source](https://news.ycombinator.com/item?id=46442106)) "It is a wildly successful project which has acquired ton of code over the years." ([source](https://news.ycombinator.com/item?id=45845659)) ## Alternatives hackers recommend - [MLX](https://hackerapproved.com/p/mlx.md): 75 recommend - [JAX](https://hackerapproved.com/p/jax.md): 49 recommend - [OpenCV](https://hackerapproved.com/p/opencv.md): 46 recommend - [tinygrad](https://hackerapproved.com/p/tinygrad.md): 40 recommend - [TensorFlow](https://hackerapproved.com/p/tensorflow.md): 20 recommend ## Cite as Hacker Approved, "PyTorch", https://hackerapproved.com/p/pytorch/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).