🎉 Introducing sqlite-js-stream: Litestream, rewritten in TypeScript for @nodejs.
Same LTX format; runs inside your process as an npm package. No sidecar, no extra moving parts.
Written by an AI agent, reviewed by a human.
blog.platformatic.dev/sqlite…
One Agent Is Easy. Ten Is a Management Problem
Everyone is running AI agents. But how do you coordinate them smoothly?
Matteo is out of office for the first time in Banter history, so naturally we replaced him with a team of agents.
The biggest surprise from our research? We didn't need to invent new fundamentals.
Durable execution already runs on three rules. One writer per run. Append-only journal. Deterministic replay.
Hypercore, the foundation of the Pear stack, follows the same principles. Engines use a lease row in the database to enforce a single writer. Hypercore achieves this with a keypair.
Tampering takes on a new meaning here. If you rewrite history in a database, it goes unnoticed. In a core, rewriting creates two signed blocks at the same sequence number. That is proof, not just suspicion.
Look at your stack. What else is decentralized by design, but centralized only in its storage location?
Vendors promise "no lock-in" for durable execution. That holds for your code, but not for your live workflows.
If a workflow is paused for two weeks waiting on approvals, its state sits locked in a single operator’s journal. Want to switch providers? Every in-flight run must be drained first. For long-running workflows, that means you’re stuck… migration never finishes.
You can move your code, but your state is anchored. That’s real lock-in. It’s not about bad intentions; it’s built into the architecture. We face the same reality.
How many of your workflows are mid-flight right now? If you needed to move, could any of them actually go with you?
Teams everywhere are deploying model-generated code. But almost no one is rethinking what that code can actually access.
Most reviews stop at "is the output correct?"
The real question is "what authority does this code run with?" Only one of these gets caught in tests. The other defines your blast radius.
It does not matter how good your model is. Even flawless generated code, if it runs with full app permissions, exposes your entire system. The risk lives in the execution contract.
Ask yourself: who on your team can say, right now, exactly what your generated code is allowed to touch?
His view? Rewriting a dependency is now one prompt away... ⬇️ so every project should carry its own code.
@matteocollina , who co-wrote the @nodejs fix, and @lucamaraschi push back on the idea that open-source dependencies are finished.
November 11th, 8:00 AM | 5:00 PM (CEST)
📢 𝐃𝐞𝐩𝐞𝐧𝐝𝐞𝐧𝐜𝐢𝐞𝐬 𝐀𝐫𝐞 𝐍𝐨𝐭 𝐭𝐡𝐞 𝐏𝐫𝐨𝐛𝐥𝐞𝐦
The same Brotli decoder bug hit Node, Bun, .NET, Envoy, and Istio at once.
@jarredsumner of @bun reads that as proof that dependencies are over.
streamyard.com/watch/rEv8sSX…
The most effective security shift I made this year is surprisingly simple: give your code a verb, not a toolbox.
When code needs to act, the default move is to hand it a client; maybe a database connection, a fetch, or an SDK. But that quietly hands over every permission you meant to protect. A general client can go anywhere, and suddenly your boundaries are gone.
Expose the operation, not the client. Instead of 𝘧𝘦𝘵𝘤𝘩(𝘶𝘳𝘭), offer 𝘳𝘦𝘤𝘰𝘳𝘥𝘴.𝘧𝘪𝘯𝘥(𝘪𝘥). Your guest receives a focused, promise-returning function. You keep control of credentials, authorization, and validation.
Strong security boundaries and great API design are the same discipline.
Where in your stack are you handing out a client when a single, well-defined function would be safer and smarter?
Next Week Banter!
From OpenAPI to AI Agent: Building a Production MCP Server, End to End
@lucamaraschi & @matteocollina are joined by a Special guest! Roberto Bianchi, Staff Software Engineer @Spendesk
Wed. November 4th, 8:00 AM (PT) | 4:00 PM (CEST)
streamyard.com/watch/PURDcva…
AI assistants generate bursty, exploratory traffic.
@Spendesk 's MCP tools run inside their @fastifyjs app, trusted internal workflows reuse handlers via mcpCallTool with no loopback HTTP, and Redis-backed sessions survive a switch between production instances.