AMA about indie web dev (November 2025 edition)

That is a fabulous question.

The biggest dilemma by a long mile working with designers on a handoff is when real-world content needs clash with the pristine beauty of design mockups. And as someone who has worked on the design side myself over the years, I can totally screw this up as well.

A few examples:

• A blog layout with glorious (fake) imagery for each post. A few months later: OMG now we are locked into having to find a good image for each post.
• A sidebar with a nice little box to highlight a news update. Only problem: the one sentence news update in the design mockup turns into several paragraphs because some content writer doesn’t know how to be concise, and some words are long blowing out the width on a smaller screen or looking weird with lots of hyphens.
• Alternatively, lots of lorem ipsum for sections of a landing page, which look nice with alternating backgrounds. Then the real landing page gets worked on and there’s not much to say for some sections so the backgrounds are scrunched up on a single viewport and looks annoying.

I could go on and on. The solution to a lot of this stuff is to recognize there will need to be some trial-and-error between work on content/UX and design, and the more you can anticipate these problems ahead of time, the more likely you can avoid the obvious footguns.

Edit: I realized I didn’t actually answer your main question. A little HTML/CSS knowledge is good, but I don’t mind working with Figma or any other design tool really. I worked on a big design system project a while back and all the component design work was done in Figma. But it helped a ton that the design tokens for colors/fonts/spacing/etc. in Figma matched up with tokens we implemented in the CSS as variables. As long as those sync up reasonably well, it makes the process go smoothly.

I’m not shy about saying I’m not a big fan of Bluesky, and I think ATProto has made lots of big promises it has been very slow to deliver. But given a choice between Bluesky/ATP and proprietary corporate silos, obviously the more open systems are preferable.

I’m bullish on fedi long-term, because I think the incentives between operators of instances (many of them not large) and the people using them are far better aligned. I honestly have no clue how Bluesky will make enough money over the next few years. Enshittification seems almost inevitable. 🤷🏻‍♂️

I think the first place to start would be to learn about some of the Signals implementations which are available. Preact Signals (by the Preact team, you don’t need to use it with Preact) is arguably the OG of standalone implementations which got a lot of buzz, but Alien Signals is also a good one. Here’s one of their code examples:

import { signal, computed, effect } from 'alien-signals';

const count = signal(1);
const doubleCount = computed(() => count() * 2);

effect(() => {
  console.log(`Count is: ${count()}`);
}); // Console: Count is: 1

console.log(doubleCount()); // 2

count(2); // Console: Count is: 2

console.log(doubleCount()); // 4

From there, the reactivity in web components piece would come from writing a render() method of some kind in your HTMLElement subclass which references signals you’ve set up and modifies component elements in the DOM. While this is certainly doable in pure JS, you can end up with a degree of boilerplate. To make that easier, I wrote a small library which provides some connective tissue between signals implementations and web components: Reciprocate

Thank you, I appreciate hearing that! CSS Nouveau was a labor of love for sure, and while I did get a little return on investment, I know that another course would also be mostly a labor of love. I don’t think course writing is my strong suit, it feels more worthwhile in the long run to keep working on OSS projects. But it’s definitely a priority for me to continue writing mini-tutorials for That HTML Blog, whether that’s for projects I work on or other libraries/frameworks/web APIs.

That’s a hard question…I don’t really have a precise answer, but the stuff that’s just content static that crosses my path I try to ignore as much as possible. In the cases where it’s direct “communication” from people you know, I would gently push back via asking as many questions as possible where they need to actually answer you. Like “I’m wondering about this part of the code, why did you use that particular structure/style/naming/etc.?” or “I didn’t quite get the point of your message, can we hop on a chat for a few minutes to go over it?” Just try hard not to sign off on anything which feels merely generated.

For inspiration I’m enjoying listening to music or watching TV where the creators specifically point out that it’s AI-free. Seeing that more and more now, which is awesome!