I used to think everyone should just get onboard with English, because everything is in EN; it'd make things easier in the long term anyway. But now, I have known better. I18N isn't about being fixated in one's own language, it's about accessibility.
Posts by quang ๐๐
I'm honestly surprised by the fact that no VI translation was proposed yet, while ~30 langs are now available. Considering Vietnam is one of the popular destination for offshore software development, this may say something about how little open source is talked about here.
npmx.dev/translation-...
Submitted a VI translation to npmx.dev at github.com/npmx-dev/npm.... Any vietnamese in the house? feedback is welcome.
For NeoVim, add `classFunctions = { `tw`, `clsx` }` to where you define your config.
Ref: github.com/neovim/nvim-...
For VSCode, just add `tailwindCSS.classFunctions": ["tw", "clsx"]` to your `settings.json`
github.com/tailwindlabs...
Do yourself a favor and add `"tailwindCSS.classFunctions": ["tw", "clsx"]` to your LSP configuration to get hint & autocomplete everywhere.
Should be documented no? I didn't know remote functions can be cached on the server too like this. That's game changing
Slow and steady wins the race. May luck and the whole web ecosystem be on your side ๐ค
A quick post describing some custom text objects in Neovim for Svelte that I've been using to make it easier to select the template blocks (if, each, snippet, ...).
Hope this is helpful. Anyone out there still typing code in 2026?
www.sveltevietnam.dev/en/blog/2026...
`tick` is generally not available in e2e test execution environment. But if even it is, `tick` is Svelte-specific and can only be used after hydration, which defeats the purpose of awaiting for hydration.
Will do, thanks for the reminder Kev
Thanks for the suggestion. I didn't know there is an existing pattern in the SvelteKit test suite. Will take a look and amend some references to the post. Fixture sounds like a good idea too.
Thanks for the write-up! Iโll add your article to my testing resource collection fubits.dev/notes/2026-0...
Hibernated from social media (including bluesky) for a few months. Highly recommended. My mental health has been much better
Back now but barely. And hoping to keep it that way.
Thanks, an amazing collection you have there!
Thanks, didn't want to disturb the man haha
Awaiting for hydration in Playwright+SvelteKit tests has improved test repeatability for me. I wrote about it here.
www.sveltevietnam.dev/en/blog/2026...
May Furgus find many furiends and happiness in furadise.
This makes me miss my old friend, he was so adventurous but we never gave him enough chances to explore the world...
PSA: Our roadmap for 2026:
I am baffled. Why the f is there a need for a dedicated unit test framework just for AI???
github.com/paralleldriv...
I wonder what had been wrong with us until we needed "cognitive amplifier".
I personally have always been in awe with human and our ability to make art / do science.
You use SlopAI all you want. All I am asking is leave place in the world for those who don't want to. Please and thank you.
I can confirm that I have seen this error too and it is in fact impossible... to not retry the vite command and hope for the best haha
Tried Coolify on homeserver today, reminisced about DevOps and DNS settings and such, got too tired and uninstalled everything... Gonna be just docker-composing for now.
Should I look into self hosting a docker registry?
Sorry to hear what you have to deal with. Building AI to fight against AI is a dangerous route I personally think. It gives AI companies more excuse to keep pumping in more resources. But at this point I don't even know what the alternative may be. I feel much distress, more everyday.
Am I making an insignificant correlation? I don't know. Do I think energy management takes an active role in instances like this, yes. Do I think it will get worse if we need to build more energy plants, regardless of being hydropower, fossil fuel, geothermal, etc? 100%.
Please think twice before prompting!
Right this moment, my hometown is experiencing historical extreme flooding. Hydroplants are still releasing water at record level.
Meanwhile, billion of dollars of AI data center investments is coming in waves, demanding more and more energy.
Your actions matter!
I don't mind outage / service / server going down (Cloudflare this time). But when it keeps happening, one company after another. Something else is going on...
I believe small models are indeed more productive. Would love to see more embedded models that run locally and offline. Centralized coding agents are probably what I despise.
Reasons I refuse to use L_L_M: is not deterministic (enough), often includes closed source data & biases, steals creative work from independent creators, is energy hungry and likely to cause energy crisis, esp. in the global south. Am I wrong?
And If I become obsolete in this industry, so be it!