Along with the blog post there's a recording of the interview. Listen to the full interview OR the last 10 minutes (linked).
www.youtube.com/watch?v=ZbWx...

I’m so stocked that Codemotion made me as a Magic card. 🥹

See you tomorrow at conferences.codemotion.com/roma/

The DCO Debate: Who Is Responsible for AI-Generated Code? Exploring the reality, benefits, challenges and accountability of AI-assisted development in the open-source community.

Another awesome post by @nodeland.dev
This express 100% my vision for AI-assisted coding.

I took a look deeply and couldn’t find a good reason. I would need to do quit expensive profiling.

From the look of it.. it seems some issue with cluster.

done, good catch!

Read the full story:

blog.platformatic.dev/ssr-framewor...

All code is open source. All data is public. Check it yourself.

github.com/platformatic...

Same eCommerce app. Same data. Same AWS EKS infrastructure. Same 6 CPU cores. Same 1,000 req/s load.

No compression on any framework. Pure SSR throughput, nothing else.

Readers spotted that compression was not applied consistently in our React SSR benchmarks. TanStack Start had it disabled while the others had it on.

We disabled compression everywhere and re-ran the tests. Here are the corrected results.

Node.js VFS: Unlocking Dynamic Code, Testing, and Multi-Tenancy Node.js delivers high-performance I/O, moving data fast between network and disk. But today’s cloud-native, serverless, and AI-driven workloads hit a wall: the filesystem isn’t virtual. That means…

PR #61478 is live on nodejs/node. 20,000 lines across 100 files.

In this episode, @lucamaraschi and I go deep on the architecture, the use cases, and why virtualizing the filesystem unlocks an entirely new class of Node.js applications.

Watch it here: streamyard.com/watch/qjiaVX...

Why does this belong in Node.js core?

The userland package works. But it carries 960+ lines of duplicated module resolution logic. It cannot hook into native addons. It cannot integrate with Workers properly.

In core, VFS hooks directly into the existing resolver and fs layer.

Real-world use cases we cover in the episode:
📦 Single executable applications without binary bloat
🧪 Fully isolated test suites that never touch disk
🤖 AI agents and code-generation pipelines importing dynamic modules

The architecture is provider-based.

🧠 MemoryProvider for in-memory testing and ephemeral workloads
💾 SqliteProvider for persistent virtual filesystems backed by node:sqlite
🪞 Overlay mode: intercept only virtual files, let everything else fall through to the real fs

Mount, unmount, compose.

Virtual File System for Node.js. From @platformatic/vfs in userland today, to node:vfs landing in core.

Write a file to memory. Mount it. require() and import just work. fs.readFileSync just works. Third-party code sees your virtual files without knowing they are virtual.

Think about what you cannot do today in Node.js:

You cannot import a module from memory. You cannot sandbox tenants with isolated file trees. You cannot bundle assets into a single executable without bloating the binary.

Every workaround is a hack. Monkey-patching core is fragile.

🔥 NEW BANTER: "Node.js VFS: Unlocking Dynamic Code, Testing, and Multi-Tenancy"

🔗 streamyard.com/watch/qjiaVX...

I’m stoked by the amount of private feedbacks I’m receiving for leading the AI contributions discussion so far.

Thank you all ❤️.

Read the full article at blog.platformatic.dev/durable-work...

Try it out: github.com/platformatic/platformatic-world

It's a drop-in replacement for Vercel World. If your workflows run on Vercel today, they'll work on your cluster with Platformatic World.

The ICC dashboard gives you full observability: real-time run status, step-level trace waterfalls, and a directed graph view of your workflow structure.

You can replay completed runs, cancel running ones, and inspect hooks and events.

For local dev, just start PostgreSQL and the workflow service:

npx @platformatic/workflow postgresql://user:pass@localhost:5432/workflow

In production with ICC, setup is two env vars and three pod labels. That's it.

Your workflow code doesn't change. Same 'use workflow' and 'use step' directives.

The fix: a dedicated service that owns the data.

@platformatic/workflow is a Watt app backed by PostgreSQL that manages all state and queue routing. It knows about active runs, pending hooks, sleeping workflows, and queued messages.

ICC only decommissions a version when all counts hit zero.

We tried extending our Intelligent Command Center (ICC) directly, but it couldn't tell when a version had no work left.

A workflow can register a webhook and suspend. No memory use, no heartbeat. ICC sees nothing and kills the version. Hours later someone clicks the link and the pods are gone.

Platformatic World pins every workflow run to the code version that created it.

Queue messages route to the correct versioned pods. Old versions stay alive until all in-flight runs complete. No manual config needed.

Vercel handles this automatically. On self-hosted Kubernetes, you're on your own.

Until now.

The problem: Workflow DevKit uses deterministic replay. When a workflow resumes, it reruns from the start, matching steps to cached results in the same order.

Deploy a new version while runs are in-flight? The step order changes, correlation IDs break, and you end up with silent data corruption.

We just launched @Platformatic World 🚀

It brings durable workflow orchestration to Kubernetes with full version safety.

If you use @WorkflowDevKit, your workflows now run on your own cluster with the same guarantees you get on @Vercel.

Hey I wrote a personal blog post... www.jasnell.me/posts/ai-in-os

I didn’t chose to focus it on the DCO. Others did it for me.