Netlify
@netlify.com
19 days ago We don't sell your code or content. We don't train AI on it. We use API vendors specifically chosen because they don't feed your data back into model training.
💬 1
♻️ 0
❤️ 3
We don't sell your code or content. We don't train AI on it. We use API vendors specifically chosen because they don't feed your data back into model training.
Your code is yours. We've said it before. Here's the full why.
But: If there are blocking things in the bundler, we are happy to help fix them. What we won't do is building a separate MF implementation in Rolldown.
In collaboration with @shulaoda, @nst.blue, @2hea1, and Zack Chapple, more bug fixes and examples have been added.
Instead of duplicating effort, we're going all-in on collaboration instead. ...we learned is that not only the scope is huge, but that most MF features inevitably live in the JS plugin layer and not the bundler.
Meanwhile, @gioboa.bsky.social had already shipped full Vite 7/8 support in module-federation/vite. The Vite plugin is now the official recommendation for Module Federation with Vite!
github.com/module-federation/vite
What happened?
We spent months exploring built-in MF support in Rolldown, even reimplemented MF 2.0 from scratch for both Rolldown and Vite.
What... instructions & details here:
supabase.com/blog/custom... We just released Custom OIDC Providers for Supabase
Supabase Auth already supports ~20 providers. Now you don’t need to ask us for more - you can enable your own and your users can login using the same code as built-in providers: the flighty app tells you how old the plane you're flying on is - i find it pretty nifty and sometimes scary
Opt-in is a deliberate choice. Opt-out is a choice someone else made for you.
We think the industry has the default backwards. We don’t sell your data, and we don’t use your code to train AI models unless you explicitly opt in.
More from our CISO Mark Dorsi:
www.netlify.com/blog/you-own... you are correct 😂
geocoding is hard, notion edition
全速前進 (full speed ahead)
Changelog #0285
• Railway x Kyoto
• We're hiring
• Q1 2026 Highlights
• Fixes and improvements
railway.com/changelog/2026-04-10-retreat-v9 I moved my manuscript from Obsidian to Scrivener (keeping my research notes in Obsidian), and so far so good
They hooked me a while back with clever h1 wordplay ("See the forrest *or* the trees") and I finally bit Postgres wasn't designed as a job queue, but many apps use it as one right alongside analytics, OLTP, and everything else. Our latest blog goes deep on what to be aware of when doing this and how the situation has improved for Postgres over time.
Introducing the Oxc Angular Compiler ✨
◆ 6.4x faster than Angular CLI
◆ 20.7x faster than Webpack
◆ First-class @vite.dev plugin with full HMR
◆ Built with Oxc
◆ Not another slop fork
While this is experiment, the @angular.dev team is looking into an Oxc integration.
voidzero.dev/posts/oxc-an... i've been trying not to buy one of these things but you're making it very hard here
March was PACKED for the Vite ecosystem. 👀
Vite Plus, Vite 8 with Rolldown, Vitest test tags, Oxlint JS plugins, and much more.
Full recap in the video!
www.youtube.com/watch?v=pyiA... AI coding agents can now optimize code performance automatically.
↳ Profile bottlenecks from runtime traces
↳ Refactor hot paths for lower latency
↳ Re-test to verify speed gains
Performance tuning is shifting from manual profiling to autonomous optimization. and might as well shoot my shot from the footnote...
@ftrain.bsky.social I'd love to find a print copy of that 2015 Bloomberg Businessweek issue - do you know where I could find one? Paul meant ert=compiling but here it's ert=deploying
just published an essay on the @val.town blog:
*Code is inert. Val Town makes it ert*
blog.val.town/code-is-inert
inspired by this sentence in @ftrain.bsky.social's "What is Code" essay 11 years ago
> Code is inert. How do you make it ert?