Posts

Most writing about the technical leadership path is noise shaped like signal: clickbait ladders, ten-point checklists, motivational platitudes sold as frameworks. This post is the opposite. It distills the field’s harder-won lessons into one working model for the stretch from Senior Software Engineer to Tech Lead to Staff or Principal, with no fluff and no name-dropping. It answers one question: what actually changes as you move up the technical track, and what do you do differently on Monday? ...

There is a Listen button on top of every post here now. Press it and the browser reads the article aloud, highlights and scrolls to the paragraph it is speaking, shows progress in a status bar, and lets you jump to any paragraph by clicking it, or step through them with the previous and next buttons in the player. It costs nothing per listener, stores no files, and shipped as three files and a config flag. Here is exactly how it works and what I had to dodge to make it survive the real world. ...

Most AI PR bots do the same thing. They read a diff, emit soft prose, and vanish. No memory of the last review on that repo. No package you can open later and see which tools the model ran. No honest score when the code is wrong. I built Luffy for the opposite job: a comment-triggered review control plane. You type @luffy review this pr. GitHub Actions runs Hermes Agent through OpenRouter. The agent can shell into the workspace, repro failures, and post a structured Markdown review. Then it stores a redacted trace and grows hub memory so the next run on that repo is not starting from zero. ...

Memori: The Recursive Design of Agent Memory Memori is not primarily a database for notes. It is an attempt to turn agent memory into an engineered retrieval system. The thesis is simple. An agent fails across sessions because state is lost. Lost state only matters if previously learned information cannot be recovered when needed. So “memory” is really three problems: Capture useful state. Preserve it cheaply and portably. Retrieve the right item at the right moment. Everything in the code follows from that split. This post is about the concepts underneath the system – what each design decision reduces to, and why the project is recursive at its core. ...

Building Persistent Memory for AI Agents Every AI coding agent session starts from zero. The agent doesn’t remember that you spent three hours last Tuesday debugging a SQLite WAL-mode deadlock, or that you decided against a particular abstraction pattern after trying it twice, or that your project has a non-obvious deployment constraint that every new session will trip over unless you paste it into the prompt again. That tax – re-establishing context at the start of every session – compounds invisibly. You stop noticing it the same way you stop noticing a slow network: individually small, collectively enormous. ...

Hard Constraints Belong in Code LLMs are probabilistic reasoners. They draw on model priors to interpret ambiguous inputs, frame hypotheses, and synthesize trade-offs across a large solution space. That’s exactly what makes them useful for complex decisions - and exactly what makes them unreliable for enforcing hard rules. Deterministic code is the opposite. It has no priors, no judgment, no ability to explore. But it guarantees correctness on anything reducible to math or a lookup. ...

Designing CLI Tools for AI Agents Most CLI tools are designed for humans who can google error messages, read man pages, and ask Stack Overflow. AI agents can’t do any of that. They get one shot at parsing your output, one chance to recover from an error, and they pay per token for everything you print. I built Memori – a persistent memory system for Claude Code – and discovered that designing for agents forced me to make a better tool for everyone. Here are the principles I extracted, grounded in exact code from the repository. ...

The AI-Augmented Developer Playbook A comprehensive framework for leveraging abstract thinking, pattern recognition, and AI orchestration in modern software development. Core Principles The five foundational truths: 1. Code is Artifact, Not Work Product The work product is correct system behavior. Code is just how you get there. 2. Abstraction is Leverage Higher abstraction = more leverage. Line-level: 1x Function: 10x Component: 100x Pattern: 1000x Architecture: 10000x 3. Patterns are Hypotheses, Not Facts Every pattern needs validation: ...

Forging AI’s Lasting Memory: Learning from First Principles and Asimov’s Robots 1. Why AI Agents Today Still “Forget” Despite advances in large language models (LLMs), interacting with most AI chatbots feels like talking to a tide: each new wave of conversation washes away earlier details. Contrast this with Isaac Asimov’s positronic robots, Daneel Olivaw or R. Giskard, who carry knowledge and context across years (or even centuries), never losing sight of a user’s core preferences or a mission’s essential facts. ...

My Reading List for the Next 10 Years and Why Over the past few years, I’ve found myself at the intersection of philosophy, cognitive science, and AI engineering. As a firm believer in first-principles thinking, I recognized the need for a coherent, decade-long roadmap to guide my intellectual growth and practical skill-building. Instead of random selections, I asked myself: What are the irreducible concepts I must master in epistemology, symbolic systems, cognition, memory, AI, cognitive systems, human-AI augmentation, and large-scale engineering? The result became this curated reading list. ...