"Zig with TypeScript bindings and React components" for a TUI is definitely the wildest, weirdest stack for a given job I've seen.
Ha, mine's weirder, I think? I wrote a React-like renderer in Lua that renders into Neovim buffers[1], then built a pile of apps on top, including an (unfinished/experimental) HN reader[2].<p>[1] <a href="https://github.com/jrop/morph.nvim" rel="nofollow">https://github.com/jrop/morph.nvim</a><p>[2] <a href="https://github.com/jrop/tuis.nvim" rel="nofollow">https://github.com/jrop/tuis.nvim</a>
Where are you seeing any mention of Zig? All I see is TypeScript?
That is a wild stack to use for a RSS feed.
I'd recommend saving actual HN item IDs rather than feed positions. Since the front page reorders constantly, relying on item IDs makes returning to a saved thread completely predictable.
I've been using <a href="https://github.com/bensadeh/circumflex" rel="nofollow">https://github.com/bensadeh/circumflex</a> for a while now, I like it a lot.
No dependency, node.js in the next line.
Piping through readability before rendering is what made mine usable, raw html to markdown on news sites is mostly nav junk.
Without Claude code demo it would be impossible to see how it works, nicely done
It would be nice to read the articles in the terminal as well, maybe just as an markdown output.
Does it support account login ?
Thank you, I've been looking for a meaningful way to browse HN on my cyberdeck and this looks perfect.
As an aside, 'meaningful' is an obtuse adjective to use there; and it's typical of text generated by Claude.[0] I seem to see it a lot more frequently than pre-AI (~2022). Maybe my next project will be statistically analyzing word frequencies in HN comments pre-AI and post-AI.<p>[0]: <a href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7031138" rel="nofollow">https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7031138</a>
Thanks for you concern, I typed that message out by hand and genuinely meant it. The terseness stems from the use of a mobile phone, not Claude. As it happens, I don't have an Anthropic account, nor do I possess an OpenAI one. If you look at my comment history, you'll notice I've been around for a while (as I noticed, have you).
I think a lot of the focus on AI output poisoning future models' training has left their effects on our reasoning/ vocubulary neglected, which is why I brought it up.<p>How we define words, what connotations they evince and the qualities they exude, simultaneously limits and enables thought.
why not use ratatui
maybe, the author isnot interested in low-level langs<p>Edit: sorry, i was wrong. <a href="https://github.com/holo-q/ratatui-ts" rel="nofollow">https://github.com/holo-q/ratatui-ts</a>
Can you make an interactive ssh instead?
How many fps can it do?
just why
thanks for sharing! gonna try it out
Loved it, thanks for sharing! If you can add logins and posting it would be great!
Made me think of a cool way to burn tokens: browser that reports websites you visit to an agent, which then writes a TUI for that site, ultimately leading you to a collection of low-resource UIs for the five or so websites you actually visit. Then you stop needing to use a browser and your 16GB RAM with a browser using 64GB of RAM becomes a system with free RAM for the first time in years.<p>Think: Lynx but pretty.<p>(Also bonus, bot checks that flag you as one means you have more free time to not get advertised to or sucked into Internet drama.)
That's a lot of deps to read HN.<p>Maybe your clanker can do it without.
And this is why people aren’t going out of the loop any time soon. A normie asks the bot to make a project, it produces nonsensical complexity like this.
Are you comparing with `curl | lynx` or with Chrome?
In the old days, you would telnet into a host machine that just serves this up to you. No need to install anything.
YaHnTui