Stefano J. Attardi

Most-linked posts?

1. We memo all the things (2020)
   6 people · 6 mentions · first linked 2020-10-28
2. Shipping a Neural Network on iOS with CoreML, PyTorch, and React Native
   2 people · 2 mentions · first linked 2018-02-13

What Hacker News says?

1. casey · 2018-02-13Recommends

   I love the description of convolutions: > I think of convolution as code reuse for neural networks. A typical fully-connected layer has no concept of space and time. By using convolutions, you’re telling the neural network it can reuse what it learned across certain dimensions. The diagram is great too: https://attardi.org/pytorch-and-coreml#convolution

   Comment linking /pytorch-and-coreml

   on “Shipping a Neural Network on iOS with CoreML, PyTorch, and React Native”
2. ot · 2018-02-13

   Submitted Shipping a Neural Network on iOS with CoreML, PyTorch, and React Native as a story · 501 points

   HN discussion
3. casey · 2018-02-13

   I believe this is what the author tried first in the post. He even links to this test UI where you can compare the "plain math" approach to the neural network: https://attardi.org/pytorch-and-coreml-test-ui/

   Comment linking /pytorch-and-coreml-test-ui/

   on “Shipping a Neural Network on iOS with CoreML, PyTorch, and React Native”
4. steadicat · 2011-01-24

   Submitted The Dilbert Black Swan Portfolio: a skeptical/practical guide to investing as a story · 114 points

   HN discussion
5. tbezman · 2021-10-23

   Potentially controversial post on why we memo all the things. https://attardi.org/why-we-memo-all-the-things/

   Comment linking /why-we-memo-all-the-things/

   on “React Docs Beta – rewritten with hooks and interactive examples”
6. davecardwell · 2021-10-25

   Submitted We memo all the things (2020) as a story · 108 points

   HN discussion
7. emptysea · 2022-08-17

   from the article: > If a component has a bunch of props and not a lot of descendants, it can actually be slower to check if any of the props have changed compared to re-rendering the component. (I don't have a source for this claim, but I've seen prominent developers like Dan Abramov make this case on Twitter) It might vary from application to application but this seems like something that could be tested. Also on t…

   Comment linking /why-we-memo-all-the-things/

   on “Why React Re-Renders”
8. steadicat · 2018-02-12

   Submitted How I Shipped a Neural Network on iOS with CoreML, PyTorch, and React Native as a story · 10 points

   HN discussion
9. ipsum2 · 2018-02-21Recommends

   Check out https://attardi.org/pytorch-and-coreml for an example of writing in PyTorch and using ONNX to convert to CoreML.

   Comment linking /pytorch-and-coreml

   on “A promenade of PyTorch”
10. steadicat · 2020-10-28

    Submitted Why We Memo All the Things as a story · 2 points

    HN discussion
11. jbaudanza · 2022-03-16

    I know this is theoretically true, but I've never found this to be the case in practice. Coinbase, in fact, has made it a rule that all things must be put in useMemo. https://attardi.org/why-we-memo-all-the-things/ Facebook is working on an experimental compiler called React Forget that will transparently memoize everything for you. https://www.reddit.com/r/reactjs/comments/rcn5ks/react_forge...

    Comment linking /why-we-memo-all-the-things/

    on “Flutter is better than React Native in all the ways that don’t matter”
12. zeorin · 2025-07-29

    React is easy, once you understand why {} !== {} and that it uses reference equality as a proxy for value equality. This is why the absolute simplest thing to do is to just memoize all the things (see https://attardi.org/why-we-memo-all-the-things/ for a deeper dive into this approach). I know that the conventional wisdom (e.g. https://kentcdodds.com/blog/usememo-and-usecallback ) is not to do that, but it's worth n…

    Comment linking /why-we-memo-all-the-things/

    on “The Useless UseCallback”

Search all of HN for attardi.org