1,431 karma · joined November 22, 2017
Since Opus 4.6, LLMs have been pretty clever at using fancy types with libraries like Servant and Beam. The expressiveness of the types, combined with feedback from the compiler, means that agents converge quickly to something that works. I don't think I've noticed agents having to run the compiler so often that compilation speed is an issue.
[0]: https://discourse.haskell.org/t/after-7-years-in-production-...
I'm quite interested though!
Also, what's the context of "Built in Canada"? What's the point of mentioning this? Should I expect the results to be better for Canadian content?
Our metrics are also showing that Haskell usage is actually growing. So overall, can't complain!
For example, one thing that is a friction point at my own job is knowledge of tooling for performance tuning and optimization. It's not that Haskell doesn't have this tooling, it's just not well-known.
Is there anything in particular you would need?
Not mentioned in the update is that the Haskell Foundation is moving towards a model that already exists within the OCaml Foundation. The HF and OF have the same challenge: a shallow pool of (relatively) large sponsors.
Of course big successful firms have more money to contribute, and we are thankful for them. This move, whereby the HF will deploy more of its financial resources on technical challenges directly, is meant to attract the other 99.9% of firms using Haskell in production. Many firms, including my own employer, have a more tit-for-tat view of sponsorship.
If you have thoughts, or would like to get involved, do not hesitate to reach out to me! E-mail in bio
My "evidence" is that trading firms that kept raking in the money, and that benefit from maximum productivity of their employees, still generally have a hybrid work culture.
[0]: https://thenewstack.io/github-will-prioritize-migrating-to-a...
From the outside, I would love to participate in semiconductor manufacturing.
Edit: I was just able to pull again
Since you probably don't have industry experience, I strongly suggest you build a good Github profile. Build your own software in public, contribute to projects. That's an excellent hiring signal
I'm particularly excited by the prospect of bringing together the ecosystem of data science tools, with Cloud Haskell[0]. There's no reason the Haskell community can't have something like Dask or Ray, but with Haskell's stellar type system.
[0]: https://github.com/haskell-distributed/distributed-process
You might be interested in reading the Monday Morning Haskell blog[0] series, which presents examples of how to do certain tasks in Haskell. See [1] for an example.
[0]: https://mmhaskell.com/blog
[1]: https://mmhaskell.com/blog/2025/5/19/comparing-code-leetcode...
I still use VSCode when navigating complex codebase changes, just because of the visual file tree. I hope to fully switch over soon.
I might try to learn this instead of neovim for a while.
[0]: https://github.com/ghc-proposals/ghc-proposals/blob/master/p...
[1]: https://blog.poisson.chat/posts/2023-01-02-del-cont-examples...
I've been using Pandoc for my static site, which I also used for my PhD thesis. I wrote pandoc-plot[0] to render figures in documents like the author does, inspired by Matplotlib's plot-directive[1]. I feel like this feature is a must for technical writing, but getting the details right is tricky!
[0]: https://github.com/LaurentRDC/pandoc-plot
[1]: https://matplotlib.org/stable/api/sphinxext_plot_directive_a...
Even if tools to programmatically create maps were available, I wouldn't have known how to program. So this is more about my lack of knowledge of tools of the time.
4-bit is small enough that you can build it manually, without the use of external tools (which I don't think existed at the time anyways).
Highly recommended for children interesting in computing!