If they built a WASM runtime, why do they keep on talking about using wasmtime, which is a WASM runtime.

Yeah they phrase it more honestly on the actual GitHub repo as a WASM execution platform which makes more sense

I appreciate the write up.

It comes off very much as a summary of days of vibe-coding from the AI itself, but it's still nice to see the condensed decisions, approach and frankly the whole concept of embedding a wasm runtime.

I would treat bwrap processes as single-use across tenants, since host-side state (allocator, FDs, namespace bits) accumulates and you can not really prove it clean.

“I got an LLM to write a summary of the work it did while I occasionally checked in on Cursor.”

I’m getting so tired of these posts. Why waste our time?

We need a button for [I carefully read and evaluated this post (I kid, I asked an LLM to summarize it for me)]".

This way, we could return the same level of effort. Most of these vibe coded tools will gather dust almost immediately, such a monumental waste of resources.