Blog
Writing
Showing 33 posts
Always Protect the King - Notes from a Founder Who Stopped Searching
On starting young, surviving betrayal, and the creed that carried me to 40 - protect the person in front of you, and be the King in the moment.
Deploying React Native/Expo Apps to the Apple App Store
A comprehensive walkthrough of shipping a React Native or Expo app to Apple's App Store — from developer account setup and code signing, through App Store Connect metadata, submission, and post-launch maintenance.
Deploying React Native/Expo Apps to the Google Play Store
A step-by-step guide to publishing a React Native or Expo app on the Google Play Store — Play Console setup, AAB vs APK, Google Play App Signing, store listing metadata, and ongoing release management.
LLMs and Their Integration with AI Harnesses: A Comprehensive Guide
An overview of Large Language Models and the harness infrastructure that makes them production-ready — covering model management, API routing, caching, observability, security, and integration patterns from direct API to self-hosted and event-driven approaches.
The Hard Truth About GitHub Actions and Continuous Deployment
Introduction
Every tutorial makes it look effortless. A few lines of YAML, a git push, and your application magically deploys to production. If only it were that simple.
The reality of GitHub Actions and continuous deployment is far messier. After years of building, breaking, and rebuilding deployment pipelines, I can tell you this: CI/CD is not a solved problem. It is a continuous negotiation between your code, your infrastructure, your tooling, and the infinite ways they can fail silently until 2 AM.
This post is not a tutorial. It is a field report from the trenches — an honest look at what actually goes wrong, why it goes wrong, and what you need to understand before you treat deployment as an afterthought.
Building null-agent: A Multi-Provider Coding Assistant with 49 Built-in Tools
Introduction
Most coding assistants lock you into a single provider, a single interface, and a black-box you can't extend. If you want OpenAI for one project, Anthropic for another, and a local model for sensitive work, you're juggling three different tools with three different CLIs.
null-agent is an open-source TypeScript library that solves this by unifying multi-provider LLM support, a rich tool system, a terminal UI, an HTTP API, and even a developer accountability tracker — all behind one configurable interface.
It ships with 49 built-in tools, 5 LLM providers, 4 interfaces, and a plugin system that lets you extend it without forking the codebase.
Stay Updated
Get the latest posts delivered straight to your inbox.