Posts
All the articles I've posted.
Beads - Memory for your Agent and The Best Damn Issue Tracker Your're Not Using
Posted on:January 24, 2026A practical guide to Steve Yegge's git-native task system for AI coding agents.
In 2026, Technical Skills Are No Longer Enough
Posted on:January 14, 2026The skills that differentiated engineers for decades are becoming commoditized. What's scarce now are the skills they were never taught.
In 2026, Planning Is Becoming the Bottleneck
Posted on:January 12, 2026When AI makes code cheap, planning and decision-making become the real bottleneck.
2026 Prediction - The Year Agents Get S#&t Done
Posted on:January 7, 20262026 is the year orchestration becomes the dominant abstraction for finishing tasks.
Vibe Decoding
Posted on:January 5, 2026Vibe Decoding is a personal experiment in using events, Kafka, and AI agents to turn everyday signals into timely, contextual insight.
In 2026, Code Reviews Are the Wrong Abstraction
Posted on:January 3, 2026Code reviews made sense when code was scarce; in 2026, the real bottleneck is judgment, not syntax.
From AI-Enabled to AI-Native
Posted on:December 29, 2025Why 2026 Will Be the Year of Real-Time, Agent-Driven Work
Towards an AI-Native Development Workflow (Using Jujutsu as the Backbone)
Posted on:December 9, 2025A practical AI native development workflow that uses Jujutsu, temporary intent files, and structured planning to keep AI generated code fast, controlled, and meaningfully organized.
How Much Can You Ask an LLM to Track? Finding the Working Memory Cliff
Posted on:December 8, 2025A short, practical look at how far you can push an LLM's working memory before accuracy falls off a cliff.
Why Embedding a JavaScript Runtime Inside an LLM Is a Big Deal
Posted on:December 5, 2025This post explains why giving LLMs a built-in JavaScript runtime unlocks far more accurate, scalable, and flexible computation than traditional tool-calling approaches.