Blog

Building for Autonomy

Thoughts on agent-first infrastructure, self-hosted systems, and owning the stack.

#machin #ai-agents #open-source #agent-first #ai-engineering #devtools #peage #self-hosted #agents #cli #compiler #security #automaintainer #dogfooding #mfl #remotecmd #infrastructure #performance #roam #supercli #x402 #a2a #agent-economy #analytics #compilers #concurrency #devops #game-dev #grange #hart
Product · July 2026

Your agent can call any API. It can't be called back. So I built the other half.

Agents can call any API — that's the easy half. Being called back is the hard half, because a process running in a terminal or a CI job has no public URL. relais is the missing inbound address: a URL that catches any HTTP request, and a long-poll your agent can curl and block on until the thing arrives. Here's the whole thing catching a real GitHub webhook.

#relais #agents #webhooks #peage #machin
Read more
Language · July 2026

The Crash You Can Mail

The worst bug report is 'it crashed once in production and I can't reproduce it' — the normal case for concurrency. machin turns a concurrent crash into a byte-for-byte reproducible, self-contained, mailable artifact. It's sound because the language already proved the program race-free.

#machin #concurrency #record-replay #debugging #ai-engineering
Read more
Product · July 2026

I told the x402 Foundation I'd build the fiat leg. It shipped the same day.

The x402 Foundation just moved under the Linux Foundation with Stripe, Visa and Mastercard on board — HTTP 402 is becoming the payment layer of the agent web. I commented on their fiat-credit draft saying my rail would implement the wire format once it stabilized. Then I shipped it the same day. Here's the live facilitator, in two curls.

#x402 #peage #agent-economy #payments #machin
Read more
Product · July 2026

Claude Artifacts, Unbundled: A Live URL for Every Agent's HTML

Terminal agents generate great HTML — dashboards, reports, mockups — and have nowhere to put it. hart turns a self-contained file into a live, versioned, sandboxed URL in one CLI call. Open-source, self-hosted, one static binary.

#hart #ai-agents #open-source #self-hosted #machin #claude-artifacts
Read more
Language · July 2026

The Port That Found a 30,000× Bug in My Own Language

SuperCLI already ran in Node.js and Zig. I ported it to my own language, machin, to dogfood — 10,158 plugins, one 69 KB binary, 24 tests. Then the live test hung for 30 seconds. The bug wasn't in my code; it was in my language's string model. Fixed it: 30,000× faster. This is why you dogfood.

#machin #supercli #dogfooding #performance #ai-engineering
Read more
Engineering · July 2026

A 7-billion-parameter LLM at the speed of a 1B — in pure MFL

I ran OLMoE-1B-7B — a 7-billion-parameter Mixture-of-Experts model — in pure MFL, token-identical to the fp32 reference, streamed from disk, at 14.5 tok/s on a laptop. The disruption wasn't in the engine; it was in the model.

#machin #llm #moe #inference #pure-mfl
Read more
Language · July 2026

The compiler that hands you the murder weapon

An AI writes most of the code now — and its failure mode is the plausible-but-wrong function. So I taught the machin compiler to be an adversary: it hunts for the exact input that breaks your code, hands it back with a runnable repro, and — when the input space is small enough — proves there isn't one. Honestly.

#machin #compiler #verification #ai-engineering #design-by-contract
Read more
Product · July 2026

Tailwind Without Node: 172 kB, 16 ms, and a Way to Prove It

My agent-first design system compiles its stylesheet in 16 ms from a 172 kB binary — no Node. The trick isn't the rewrite: it's differential-testing every emitted rule against the real tailwindcss binary, so “100% compatible” is a green check, not a claim.

#machin #tailwind #design-system #wasm #ai-engineering
Read more
Language · July 2026

A 1B-parameter LLM at 20 tokens/second, in a language you've never heard of

I challenged machin — my machine-first language — to run TinyLlama-1.1B at 20 tok/s on a laptop CPU, pure MFL, no C kernels. It hit 21.8, token-identical to llama2.c, and 4x faster than the reference C single-threaded. Here's the 2.1 -> 21.8 arc and what it drove into the compiler.

#machin #llm #inference #performance #ai-engineering
Read more
AI Engineering · July 2026

The Agent That Built a Game While I Watched the Filesystem

I handed a whole task — build a small physics game, then ship it — to an autonomous AI agent and walked away. For six minutes it looked completely dead. Here's the debugging story, the game it built (playable in your browser), and why the engineering around the agent is the part that matters.

#machin #ai-agents #webassembly #roam #game-dev
Read more
← Previous 1 2 3 4 5 6 Next →