this is super cool as an art form but ASCII art is an accessibility nightmare so please don't use it for docs unless you know what you're doing and have made it accessible in some other way

Is there any way to make images accessible other than the alt tag? I'm sure SVGs are more machine-readable, but how would that help vision-impaired folks?

im not an expert on a11y best practices, but id say an <img> tag with alt text is enough

for inline SVG, i read somewhere you can use <title> and <desc> to describe the SVG document, but i don't know how they work in practice

Is it still true with llms being so good at interpreting it? I just tried all the examples on the home page, it works perfectly. In the past couple months I've moved almost entirely back to the terminal because I can just ask my coding agent to "have a look at this tmux session".

Good reminder to keep in mind.

At the same time, I don't think one should necessarily limit your expression based on constraints like accessibility.

they definitely are. im not sure why they did that, still pictures are okay too and much greener

Aren't procedural macros amd build.rs arbitrary code being executed at build time?

Pretty much, yes. And they don’t have much as far as isolation goes. It’s a bit frightening honestly.

It does unlock some interesting things to be sure, like sqlx’ macros that check the query at compile time by connecting to the database and checking the query against it. If this sounds like the compiler connecting to a database, well, it’s because it is.

i like the points the article makes, but i really wish it used looping videos instead of actual GIFs

i don't really see any reason to use GIFs here; any widely available video codec like H.264, VP8/VP9 or AV1 will result in significantly smaller file sizes, look better, and will allow enabling controls for seeking and play/pause

See also: https://vert.sh/

Judging by the comments: was this project built by an LLM, posted here by an LLM, the replies to comments written by an LLM?

Is there even a human at any step or is this an agent running in the background with "go to HN front page" as its ultimate goal?

And it seems there are a bunch of broken stuff that anyone with a lick of CSS skills would notice and fix right away. For example, the first 4 elements of each card (number, title, external link icon, and tag) seem like they should've been in on the same line. The pill-like "Hot" doesn't normally take a whole line by itself...

This os trivially fixed by wrapping them in a flex container. In tailwind that would take 2 seconds.

So we have:

- OP didn't even look at this project before posting it, or...

- they looked and didn't find anything wrong (no eye for quality), or....

- they looked, saw it looked off, but didn't bother to fix it, either due to lack of necessary skill to do that or maybe they didn't want to invoke the agent a second time???

I concur with your conclusion that it seems no human was involved in any step.

Also, evidently, thing doesn't even work given it shows years-old repos that didn't actually get any traction recently. It's AI slop in product form

Even if an LLM came up with the idea for using an LLM to come up with project ideas, then implementing those project ideas, then posting those projects on HN and representing the project on HN. Someone set up the first LLM. So there's always a human...

First I thought you're surely kidding, but after looking into it I realized you could be right. RIP HN.

It's a bit sad that this is using Svelte 3 instead of Svelte 5, which changed its reactivity model to signals!

Displaying comments is cheap because the AppView (the "cache" layer that consumes the aggregated feed of records from the relay) can maintain a relational database of the social web/graph.

When it receives a post, it creates a row in the Post table. When it receives a comment, it creates a row in the Comment table, which has a column pointing to the Post it's replying to. Then, getting all comments under a post is easy SQL.

About your second question: yes, you're correct. What handle you use and where your data is hosted are completely separate matters.

used to be, in the early days, but nowadays runtimes optimized the function call overhead between WASM and JS to near zero

https://hacks.mozilla.org/2018/10/calls-between-javascript-a...

It did improved a lot, but unfortunately not near zero enough.

It is managable if you avoid js wasm round trips, but if you assume the cost is near zero you will be in for a unpleasant surprise.

When I have to do a lot of different calls into my wasm blob, I am way way faster batching them. Meaning making one cal into wasm, that then gets all the data I want and returns it.

I don't know if it's just me but in my case the problem isn't really the crypto bs, but rather Brendan Eich himself.

As much as Mozilla and Firefox can be criticized for both technical and non-technical reasons, at least I share the same core values. I don't seem to share any core value with Brave or its creator. Plus, yeah, still smells like Google :)

Brendan Eich’s contributions to computing are immeasurable. Your opinions of his social or political views don’t change that.

Do you not use Linux because you don’t like Linus? He’s quite a controversial figure. And before you say Linux is not Linus, the same can be said about Brave and Brendan.

Many other people work on these projects than just the leader.

Do you not use JavaScript because you don’t like Brendan?

> No moral choice is possible unless all your choices are moral.

https://knowyourmeme.com/memes/we-should-improve-society-som...