HN user

kyleyeats

1,150 karma
Posts17
Comments363
View on HN
2024 YR4 1 year ago

The story of this asteroid having a ~3% chance to hit Earth was huge, but I never heard about it being ruled out shortly afterward. It still has a 4% to hit the Moon.

Atomic CSS is great. It does abuse class="" in the same way style="" was abused. It's counter-intuitive but the obvious code smells produced here are better than the non-obvious code smells created by other style architecture approaches.

I've always now-and-then packaged up and open-sourced the pattern I use for CSS. The projects have gotten smaller and smaller. This reflects well on CSS as a technology.

I think class="" has more to offer in an information density sense. There's more potential there than style="" had. Instantly lumping them together was my first response too, but I was wrong. The in-the-HTML shorthand of frameworks like Bootstrap/Tailwind/CASS is insanely useful in a way that inline styles never were.

Yep, it's weird and controversial. That's why I wrote a book explaining it and used a pseudonym.

Now: ~60% of devs know some CSS, rest feel guilty

YSAC's pitch: ~10% of devs know lots of CSS, rest avoid it without guilt (they suck and that's okay)

I guess 0% would mean a dead CSS, so it's closer to dead, sure.

If you get a more powerful saw, you can either cut down more trees or you can spend less time cutting down trees. CSS is a more powerful saw now. I'm suggesting spending less time cutting down trees instead of cutting down more trees.

The argument I make in the book is that the last 5% of CSS/design should be written by people who can write CSS. Nobody else should even be writing CSS because it turns into a huge mess when everyone jumps in.

I mothballed this project because people were so incredibly cruel about it (a CSS project!). Remember that people who work on this stuff are people, and we're just trying to make things better. Also, you can pry .vertical-center from my cold, dead hands.

Thanks! Lists were the bane of my existence with this project. I could never make all the browsers happy.

This is a #wontfix (sorry), but I might fork it into a new, LLM-oriented CSS project. Fonts and lists will be the first things I look at.

The predecessor to this project was a Bootstrap 3 theme: https://rriepe.github.io/1pxdeep/

The book has a lot of content over what to make pretty and what not to make pretty. I think knowing what not to bother with is an underrated skill. A lot of what inspired me to write it was backenders handing off markup that they tried to make semi-passable. Unstyled HTML, please!

The perfection should be downstream of the project's coded standards, not downstream of the faithful implementation of the designer's work. Some designers are really good about maintaining standards-- but the book would argue all that effort should be spent elsewhere.

It's basically "the arbitrary padding the designer liked in the moment" vs. "the standard padding that's everywhere in the project." This book argues you should always use the standard padding. Your product should be pixel-perfect, just not in the PSD-to-HTML sense.

Edited the year into the title. CASS (the CSS library) and YSAC (this book) have been a huge marketing mess, to be honest. And yeah the substack was part of it-- I think the book came first though. Kyle Yeats (rhymes with stylesheets) was going to be a Youtube persona but I could never really get the feel of it right.

The book morphed into being more about project management. I think there's a lot of value in it still, in that respect, so I'm putting it all online for free.

As someone who made a competing product along these lines (that got no attention or traction): Tailwind gives CSS a "place" in the codebase. It benefits orgs, not necessarily apps. I didn't get it at first either. But it's very useful to the people to whom it's useful.

The seven offenses in question: distributing narcotics, distributing narcotics by means of the Internet, conspiring to distribute narcotics, engaging in a continuing criminal enterprise, conspiring to commit computer hacking, conspiring to traffic in false identity documents, and conspiring to commit money laundering

I was eating lunch at a place I'm applying to, and volunteering at, and someone casually dropped a hateful comment towards "straight white males." At work! Nobody called her out on it. It boggles the mind. So I stopped volunteering and applying there. Good luck with the hate, that's my thought.

This is a propaganda piece by the state-funded BBC. Notice that the focus of the article is whether the technology is accurate enough, not whether they should be doing it at all in the first place. It's an Overton nudge, a softball position that's easy to take-- because of course the tech will get better.

Personally, I'm in favor of nuking the Moon. The Moon is moving away from the Earth at a steady rate. I consider this an act of secession.

Not only would nuking the Moon teach it a lesson, it might also keep it in our orbit. All we have to do is nuke the far side of it enough times to stop the outward movement. If we do it too much, it's an easy fix-- just nuke the near side.

Anyone who opposes this idea is for a moonless Earth, and all the chaos that would bring. I call them anti-tiders. They're the most despicable people on the planet.

But why are you pushing errors to production? You know you're allowed to fix the LLM's code output, right?

If a robot could paint your house, but made three small errors, would you refuse to use it? Or would you just fix the three small errors by painting over them?

There's some kind of John Henry complex going on in this AI discussion.