Re:Show HN: Sumi.css – CSS Framework
by xd1936 on Wednesday October 21, 2026 @02:15PM
(#50185228)
This site seems very slopped together.
Several of their new components are just classes, defeating the original purpose and contradicting the vision of Pico.
Why are the first one or two letters in headings wrapped in <em>, like the _Bl_ in "Blades"?
Why are they calling their Pico fork "anyblades/pico", when the original Pico author's post says "The Pico CSS name and logo are not covered by the MIT license. Please give your fork its own name, so users don’t confuse it with the original."?
Re:Show HN: Sumi.css – CSS Framework
by anydigital on Sunday October 25, 2026 @02:29PM
(#50234541)
No fake stats or LLM seeding. The initial percentages were calculated against respondents who actively use CSS frameworks (~4.8k devs). They have been updated to reflect total State of CSS respondents (~9k–10k), including vanilla/CSS-in-JS users.
Re:Show HN: Sumi.css – CSS Framework
by anydigital on Sunday October 25, 2026 @02:28PM
(#50234536)
Appreciate the honest feedback, so here are direct answers and fixes:
Utility classes: Offloaded to a separate "anyblades/ninja" repo. Core Blades stays true to original Pico (which btw used some classes like .overflow-auto).
Fork naming: The main project is Blades at https://github.com/anyblades/blades. The "anyblades/pico" repo was just an early bugfix fork created months before Pico announced deprecation.
CSS Frameworks stats: The initial percentages were calculated against respondents who actively use CSS frameworks (~4.8k devs). They have been updated to reflect total State of CSS respondents (~9k–10k), including vanilla/CSS-in-JS users.
Re:Show HN: Sumi.css – CSS Framework
by elicash on Wednesday October 21, 2026 @11:13AM
(#50182400)
I used a classless CSS framework once, for a project where I wanted a non-technical coworker to update an HTML file repeatedly throughout the week and then upload it to AWS/s3. I wanted it to be as simple for them as possible, to minimize errors, and it worked great.
I'm not sure what other use cases are, but it's also just true that limitations can be fun.
Re:Show HN: Sumi.css – CSS Framework
by codazoda on Wednesday October 21, 2026 @01:27PM
(#50184548)
Looks good!
I created a similar alternative, Neat CSS, a while back. I use it for tons of projects. It could probably use a fresh look from me to see if there's anything I want to update but it's been a stable "framework" for me. I didn't intend it to be classless, though it's mostly semantic, but I instead focused on keeping it small and tight. It's only 3KB.
Re:Show HN: Sumi.css – CSS Framework
by oliwarner on Thursday October 22, 2026 @06:46AM
(#50193622)
Plenty of interfaces mimic a form process. If that's the case, and it's really just a page positional thing (not state), I see no problem making button-based navigation.
We don't scream and shout that page nav links all look like content links, after all. Well, maybe you do, but it's not 1996 any more.
Re:Show HN: Sumi.css – CSS Framework
by delbronski on Thursday October 22, 2026 @03:21AM
(#50192252)
Honest question. Not trying to offend any css fans here.
Why would someone use this nowadays instead of just using AI? To me the purpose of these small css frameworks was to have a nice consistent UI quickly. What is the value of these frameworks now?
Re:Show HN: Sumi.css – CSS Framework
by SimonLarsen on Thursday October 22, 2026 @03:29AM
(#50192324)
Two answers I can think of:
1. Believe it or not: some people are still not using AI.
2. AI generated websites tend to look very similar. Having a clean starting point and then tweaking the style to your liking avoids the obvious AI-stink.
Re:Show HN: Sumi.css – CSS Framework
by jszymborski on Thursday October 22, 2026 @03:34AM
(#50192361)
Sometimes I don't want to take out the stand mixer and just mix things by hand. I prefer to walk to the grocery store rather than take the car. I don't always use a drill where a screwdriver will do.
https://picocss.com/docs/maintenance
https://blades.dev/
Several of their new components are just classes, defeating the original purpose and contradicting the vision of Pico.
Why are the first one or two letters in headings wrapped in <em>, like the _Bl_ in "Blades"?
Why are they calling their Pico fork "anyblades/pico", when the original Pico author's post says "The Pico CSS name and logo are not covered by the MIT license. Please give your fork its own name, so users don’t confuse it with the original."?
The worst offender: What is going on with this nonsense: https://blades.dev/css/frameworks/
Utility classes: Offloaded to a separate "anyblades/ninja" repo. Core Blades stays true to original Pico (which btw used some classes like .overflow-auto).
Fork naming: The main project is Blades at https://github.com/anyblades/blades. The "anyblades/pico" repo was just an early bugfix fork created months before Pico announced deprecation.
CSS Frameworks stats: The initial percentages were calculated against respondents who actively use CSS frameworks (~4.8k devs). They have been updated to reflect total State of CSS respondents (~9k–10k), including vanilla/CSS-in-JS users.
1. https://github.com/oxalorg/sakura
It's not a direct replacement; I fixed a lot of annoyances from PicoCSS.
I'm not sure what other use cases are, but it's also just true that limitations can be fun.
https://dohliam.github.io/dropin-minimal-css/
Small bug, opened an issue here: https://github.com/vladocar/Sumi.css/issues/1
I created a similar alternative, Neat CSS, a while back. I use it for tons of projects. It could probably use a fresh look from me to see if there's anything I want to update but it's been a stable "framework" for me. I didn't intend it to be classless, though it's mostly semantic, but I instead focused on keeping it small and tight. It's only 3KB.
Anyway, thought I'd share it here too.
https://neat.joeldare.com
Please, just use a link. This has a destination, so making it masquerade as a button gives so many false affordances.
We don't scream and shout that page nav links all look like content links, after all. Well, maybe you do, but it's not 1996 any more.
Why would someone use this nowadays instead of just using AI? To me the purpose of these small css frameworks was to have a nice consistent UI quickly. What is the value of these frameworks now?
1. Believe it or not: some people are still not using AI.
2. AI generated websites tend to look very similar. Having a clean starting point and then tweaking the style to your liking avoids the obvious AI-stink.
Same reasoning.