HN user
Software engineer from Switzerland.
Like big data and machine learning.
You can get a refund, from the article:
If you hit unexpected limits or these changes just don’t work for you, you can cancel your Pro or Pro+ subscription and you will not be charged for April usage. Please reach out to GitHub support between April 20 and May 20 for a refund.
You can also bootstrap your initial schema with LLMs with the excalidraw MCP "app" [0]. But MCP "apps"[1] are quite new and not very well supported yet.
[0] -- https://github.com/excalidraw/excalidraw-mcp
[1] -- https://modelcontextprotocol.io/extensions/apps/overview
We could simply make the "view file" tool not able to see .env. Same for other "grep-like" tools.
In the improved error message [0] how are they able to tell nested attributes without having a big impact in performance? Or maybe this has a big impact on performance, then using exceptions for control flow is deprecated?
...
print(container.area)
[0] -- https://docs.python.org/3.15/whatsnew/3.15.html#improved-err...
All submissions must come from verified individuals or organizations. Inside the OpenAI Platform Dashboard general settings, we provide a way to confirm your identity and affiliation with any business you wish to publish on behalf of. Misrepresentation, hidden behavior, or attempts to game the system may result in removal from the program.
They really want your ID
Unlike openai, you can use the latest grok models without verifying your organization and giving your ID.
The logo is further on the left than the other icons.
But why the logo of the website/app should be aligned with the icon of the actions?
The icons are thin, compared to the text, which is bold.
Why this is an issue?
I can somewhat agree with the other points, but I wouldn't call this "bad design." Just because the information can sometimes be presented better doesn't mean the previous way was bad.
DNS tunnels with iodine works well, it's easy to setup and work in a lot of place.
You can also connect to some random corporate wifi and it's very likely that this will work (not necessary in "direct" mode).
I hope you will get further sponsorships so you can continue your cool work in the FOSS space.
I also totally agree that uv makes shipping python software smoother. Prior to uv I hesitated to use Textual to build cool TUIs for internal tools. But now, it’s a totally different story. I wish I had an internal tool to write with Textual right now.
Maybe by the time Toad is officially out, I will have a project ready to dive into.
I also prefer the mental model of immediate mode, but when I played with Dioxus[0] for a rust fullstack hobby project[1], I was able to adapt.
I liked the DX with the tools and the `rsx!` macro. The use of `#[cfg(feature = "server")]` to define server-side code is interesting, it lets you keep a shared codebase for frontend and backend, while still controlling what gets compiled to WASM for the client.
[0] -- https://dioxuslabs.com/
[1] -- https://blazingboard.ch/ (not mobile friendly, sorry)
I agree, the latest models are not bad at Rust. Most issues I have when writing LLM-assisted Rust code are related to dependencies.
There are a lot of v0.x.x crates in Rust with frequent updates and breaking changes. As LLMs "learn" from code in the wild, they are exposed to different way to use the same crate and thus fail to produce working code "from memory".
But as soon as you explain how to use it with the latest syntax, they are able to fix the code.
You can also install it "globally" for your user with:
uv tool install ty
ty checkThey are using the "normal" dependencies section in the pyproject.toml, so you should be able to install with pip or other tools if you want.
It's only for dev-specifics dependencies that they are using the uv-specifics section.
You can try uv in 2 min and see if you like it or not.
Some "micro cheating" are really easy to develop. When I was younger and a bit bored, I wrote my own "micro-cheat" in AutoIt[0] with less than 10 lines of code.
This was for the game counter-strike (I don't remember which version, either Source or early CSGO). The logic of the cheat was:
- I manually aim, with the sniper, close to the wall of an intersection
- I press a special key, then when the pixel at the center of the screen change, simulate input mouse click to "fire"
I prefer mypy but sometimes pyright supports new PEPs before mypy, so if you like experimenting with cutting-edge python, you may have to switch time to time.
disclaimer: I don't work on big codebases.
Pylance with pyright[0] while developing (with strict mode) and mypy[1] with pre-commit and CI.
Previously, I had to rely on pyright in pre-commit and CI for a while because mypy didn’t support PEP 695 until its 1.11 release in July.
After this month, Python 3.8 will be end-of-life, so maybe you can push internally to upgrade python.
With the 3.13 TypeIs[0] and the 3.10 TypeGuard[1], we can achieve some of Rust's power (such as the 'if let' pattern) without runtime guarantees.
This is a win for the DX, but this is not yet widely used. For example, "TypeGuard[" appears in only 8k Python files on GitHub.[2]
[0] -- https://docs.python.org/3.13/library/typing.html#typing.Type...
[1] -- https://docs.python.org/3.13/library/typing.html#typing.Type...
[2] -- https://github.com/search?q=%22TypeGuard%5B%22+path%3A*.py&t...
I agree, you can also use Rust with PyO3[0] and Maturin[1]. It's far easier than what the author is doing is the article.
[0] -- https://github.com/PyO3/pyo3
Most people in tech will roll their eyes when they hear this, but they'll keep their mouths shut.
You can use conventional pre-commit[0] to validate that commit messages follow the convention and use only the types that the team agreed upon.
This still requires discipline to choose the "right" type and scope.
[0] -- https://github.com/compilerla/conventional-pre-commit
Similar(identical?) concept with typestate pattern[0]
[0] -- https://zerotomastery.io/blog/rust-typestate-patterns/
py2wasm converts your Python programs to WebAssembly, running them at 3x faster speeds
This is clearly written in the article, but I hope that the impatient readers will understand that this is 3 times faster than the CPython wasm, not the native CPython.
In C++, you can use C++20 concepts[0].
[0] -- https://en.cppreference.com/w/cpp/language/constraints
Look nice! I like how you use the scope to remove the animation, this is very convenient.
Yes, we are aware of typst. I think it’s cool, but C++ hasn’t replaced C, Rust hasn’t replaced C++, Typst is unlikely to replace LaTeX. Likewise, many are aware of LuaTeX, but, again, the entrenching of a 40-odd year system is not to be underestimated. I am rooting for typst, anyway, and hope it finds its place. A good place to start would be to provide a compilation toolchain from typst to TeX, if they really want to replace TeX.
Pandoc[0] can convert Typst to LaTeX.
IMO If you are able to write in Typst, write in Typst, it's so much better and readable. Your final LaTeX3 macro are hard to read and difficult to parse with the eyes... Also Typst is easier to learn.
[0] -- https://pandoc.org/try/
Soon the 2017 April joke "GeForce GTX G-Assist" from Nvidia will be doable.
I don't see any direct dependencies to polars. But in the comment, they wrote that some part are "taken from", "influenced by", "adapted from" and "based on" polars.
[0] -- https://github.com/search?q=repo%3AEventual-Inc%2FDaft%20pol...
It's because they are comparing to distributed solutions only. Polars is one of the fastest solution for single computer workflow, but it doesn't support distributed workflow.
No Pay
From this job posting [0]
Compensated Interview Tasks: We value your time and effort; candidates will be compensated for completing interview tasks.
[0] -- https://join.com/companies/qdrant/10275180-core-rust-enginee...