Issue #16: Big Yellow Taxi

🎵 Don’t it always seem to go that you don’t know what you got ‘til it’s gone? They paved paradise to put up a parking lot… 🎵

The software world changes constantly, and we all working in the industry may say that people make these changes with good intentions. But what happens eventually and whether the results bring good or bad consequences — only time will tell. In this issue of the Metaframeworks Weekly, we will try to analyze if the ecosystem news we see in tech resources are brought for real developer value or just to make some quick money on the current hype.

The Good

After the recent weird divide in the Redwood project (not sure how to call it now — RedwoodJS, RedwoodSDK, Redwood GraphQL, or something else, but there should be some kind of forest there anyway for these trees), the guys have started to come up with lots of insights on full-stack web development, tools, and the philosophy behind it. While all of that might sound a bit chaotic, there’s some decent core of wisdom in their recent blog posts (authored by Herman Stander, though sometimes I felt some strong Hunter S. Thompson presence influence).

In the piece on the web evolution, the team sets the scene for the current RedwoodSDK project developments, making them sound like a logical prolongation of the history of web software development.

The next article is called “Fullstack in the true sense,” hinting at Redwood’s positioning as some kind of reimagined full-stack answer to all the problems of the aforementioned evolution (with a strong Cloudflare scent).

And the most recent blog post actually demonstrates some practices and opinions related to implementing these answers, like so-called full-stack co-location. While not being something new for many web frameworks (Hi, Angular!), this approach sounds like some revelation for the neo-React world, which is probably just an illusion caused by the author’s message tone, but still. Everything old is new again.

Anyway, even if you won’t learn a lot new from this Redwood’s thought leadership marathon, it still is a good direction and brilliant format for creating a good wider developer tribe around the [pretty narrow] technology, so kudos to the team.

The Bad

Another interesting opinion about the good and (rather) the bad parts of Next.js came from the team at Hardcover (who do something around books in 2025, so there’s no reason not to trust them). The article called “How We Fell Out of Love with Next.js and Back in Love with Ruby on Rails & Inertia.js” starts some kind of series dedicated to the topic and tangentially related to the web evolutions described in one of the Redwood’s articles mentioned before, but in a slightly backward direction. There’s a lot to digest and multiple edge-case prerequisites, but all in all, among the sane opinions and pains, one can find a good bunch of advice on assessing one’s tech stack in a thoughtful and reasonable way, even if you are not going to give up on Next.js and metaframeworks as a whole.

The Noteworthy

The last week brought some notable new releases and events worth mentioning, as usual, and the ground of it all is of course the release of Node v24 with npm v11. Time to update your .nvmrc and some GitHub actions, isn’t it?

But even if it’s not, and you’re too busy with your agentic AI project, there’s something interesting for you too. Not only has Redwood gone full Cloudflare way, but there are also some efforts from the Nuxt team around that, specifically the NuxtHub. They have released hubAutoRag for creating fully-managed RAG pipelines, potentially allowing the religious Nuxt fans to stop using Vercel’s AI SDK for that purpose. Need to admit, the Nuxt team is doing great work around the projects from different sides of the stack, and TIL, there’s even a Raycast extension for auxiliary Nuxt tasks.

There haven’t been too many other metaframeworks-related releases recently, but React Router got its new minor which always means some goodies for traditional React (and Remix now!) development teams these days. Moreover, the team did some decent bugsquashing with lots of resulting important fixes making the tooling even more stable and trustworthy.

Last but not least, Barcelona seems to have had a flood of Svelte developers last week, hosting the Svelte Summit. There were lots of insightful talks, as usual, both from the core team members and the ecosystem representatives, so if you’re biting your elbows for not visiting it, you can still watch the recording just for forty bucks (instead of a couple of months wasted on trying to make Cursor subscription work, for instance). And if you want to feel the vibe of it, check out this cool Bluesky thread from Storybook’s Jeppe Reinhold.

I admit, much of the stuff mentioned today sounded a bit sarcastic or sharp, like Joni Mitchell songs or something. But after all, that’s how we, questioning everything, move forward in the search for better solutions and viable outcomes. I hope it helps in some way too, as a sane view of the problem from different points always does. And if not, at least I have reminded you about the good ol’ song, as usual. And that’s already something.

👋