Latest notes
-
Why Scaling Infrastructure Fails Before Throughput Does
Most infrastructure doesn't fall over because it ran out of capacity. It falls over because the number of things you have to keep alive grew faster than your ability to reason about them.
-
Building a Local-First Autonomous Browser on a Patched Chromium
A local-first autonomous browser platform built on a pinned Chromium downstream, driving the engine over CDP and refusing to fall back to stock Chrome.
-
Building a Permissionless Bonding-Curve Launchpad on Robinhood Chain
A Stock-native meme-token launchpad where fixed-supply ERC-20s trade on a constant-product curve quoted in a Robinhood Stock Token, then graduate into a locked Uniswap V4 pool.
-
Building a USB HID That the Operating System Believes
What it takes to make an ESP32-S3 present as a keyboard and mouse indistinguishable from real hardware, and why "indistinguishable" is a systems problem, not a soldering one.
-
Proving a Document Existed Without Storing the Document
ProofGate anchors a SHA-256 hash of your file or text to Solana via the Memo Program, giving verifiable proof of existence without ever storing the content.
-
Shipping the Moonshine Labs Site With a Bun Contact API
The Moonshine Labs studio site is a static front end paired with a tiny Bun-powered contact API and a repeatable deploy path.
What I work on
AI Agent Infrastructure Audits
Find where an agent system will fail operationally before it does.
Distributed Systems Architecture
Design coordination and state that survives partition and load.
Real-Time Multiplayer & Collaboration
Shared-state systems that stay consistent under concurrency.
Validator & Protocol Infrastructure
Stabilize and operate protocol nodes and validator fleets.
Productionization of AI Systems
Take AI from demo to something that runs unattended.
Reliability & Operational Recovery
Stop the bleeding, then make the failure impossible to repeat.