Articles

Thoughts on software development, AI, and technology. Sometimes other things too.

2026

  • Newsletter •

    What Outlives the Code

    Joseph Campbell s monomyth says every hero story is the same story wearing different clothes. The hero descends, transforms, returns. Gilgamesh, Odysseus, Luke Skywalker. Different languages, same...

  • Newsletter •

    The Simplest Feedback Loop

    The fastest way to improve your AI workflow isn t a better model, a bigger context window, or a fancier orchestration pattern. It s a three-step loop that takes five minutes to set up.Make a skill....

  • Newsletter •

    How the Claude Code Team Works

    Notes from Boris Cherny (creator of Claude Code) on the Y Combinator Light Cone podcast.Source: Inside Claude Code With Its Creator Boris ChernyI feel like there are so many nuggets in this interview...

  • Newsletter •

    Harness Engineering Has a Name Now

    September 2025. I m building what I called an AI Goalie: an agent that reads production logs on a cron job, compares them against existing GitHub issues, updates those issues with new context if...

  • Newsletter •

    The Agency Multiplier

    There s a version of the AI hiring conversation that goes: Does the candidate know how to use the tools? That s the wrong question.The right question is older. It just matters more now.Agency. The...

  • Newsletter •

    When Can You Run Claude Code Overnight, On Your Own Hardware?

    Every night around midnight I kick off parallel Claude Code agents against acceptance criteria files. By 7am there s a PR waiting. Sometimes three. The model does the work while I sleep, and most...

  • Newsletter •

    Teaching Agentic Coding to a Team

    This week I paired with one of our fastest engineers on a webhook refactor. Quick and dirty job: break a monolith endpoint into handlers, keep the abstraction thin. I pulled up Claude Code and...

  • Newsletter •

    When Your Harness Becomes Your BI Team for a Day

    Yesterday I needed to answer a question that should have been simple: How much does it cost us to serve each client? Engineering costs are scattered across seven services. Anthropic, OpenAI, MongoDB...

  • Newsletter •

    Ralph Wiggum for Teams

    This weekend I watched Ralph chew through a Telegram agent I was building for myself. Similar to my Gastown experiments but different. Seven PRs yesterday at work. Each one reviewed and merged. The...

  • Newsletter •

    Zero Trust Launch Day: Search Over Prediction

    Last month we switched our first client team from our legacy VPN to Cloudflare Zero Trust Enterprise. We knew the launch would break things. We just couldn t predict what.Zero Trust is aggressive by...

  • Newsletter •

    Orchestrate to Survive the Speed

    I ve been running 3-5 Claude Code instances in parallel for months. They re all productive. They re all stuck waiting on me.The speed isn t the problem. The speed is almost too easy to get. The...

2025

  • Newsletter •

    Build vs Buy

    Looking at a platform invoice last month. $10 per seat. 2,000 employees. $240k/year.Half the features we never touch. The modules for our Philippines and LATAM teams are thin or missing entirely. We...

  • Newsletter •

    The Asymmetry Between Construction and Destruction

    April 2019. Notre Dame burning. Centuries of construction, hours of destruction. Master craftsmen, generations of accumulated technique, irreplaceable artifacts. Gone in an evening. The restoration...

  • Newsletter •

    The Only Eval That Matters

    GPT Image 1.5 launched this week and immediately took #1 on every leaderboard that matters. LMArena: 1277. Design Arena: 1344. Artificial Analysis: 1272. Clean sweep.Reddit, Twitter, and Discord all...

  • Newsletter •

    The Rounding Error

    spoiler-alert: This post contains spoilers for Consider Phlebas by Iain M. Banks. If you haven t read it, consider stopping here and picking it up. It s a fantastic sci-fi novel and the first in the...

  • Newsletter •

    Building a Slack Bot That Actually Knows Your Stuff

    I wanted to add Google Drive to my Slack bot. Simple enough, right? Just find an MCP server and plug it in.Except there isn t one. Not an official one, anyway.There are community implementations...

  • Newsletter •

    Claude 4.5 Opus in Claude Code

    Been using this all morning and the reports seem true. Anthropic reported Opus 4.5 reduced token usage Anthropic s new Tool Search Tool cuts context bloat by 85% for agentic systems. Potentially that...

  • Newsletter •

    Orc Arch

    Building with AI agents in production, you hit the same wall every time: context rot. You give OpenAI/Claude/Gemini detailed specs, it starts implementing, and halfway through it barely resembles...

  • Newsletter •

    Building Safety Nets for Dangerous Ideas

    Ship faster. The only acceptable speed is faster.Kaz, CEO of OpenDoorI wrote about this in Many Discoveries are Accidents, how iteration velocity creates discovery surface area. But velocity without...

  • Newsletter •

    Notes from The Agents Economy Backbone

    Latent Space Podcast - 1h 37m with Emily Glassberg Sands, Head of Data AI at StripeKey TopicsStripe s Domain-Specific Foundation ModelBuilt payments foundation model processing 50K...

  • Research •

    Toward a New Global Economic Treaty: From Bretton Woods to an AI/Energy-Based Order

    Explore the potential for a new global economic treaty based on AI and energy, and its implications for geopolitics and technology development.

  • Article •

    SvelteKit with Cursor AI

    SvelteKit with Cursor AI: Boosting Development Productivity 2-4x

  • Article •

    Mongo with Large Datasets

    Mongo with Large Datasets: Understanding the Performance Challenges and Optimizing for Scale

  • Article •

    Vibe Coding is the Future

    Vibe Coding is the Future: Why Agency and Taste are More Important than Ever for SWEs and AI Engineers

  • Research •

    Mongo Vector Date Filtering

    Mongo Vector Date Filtering. How to filter vectors by date in MongoDB Atlas Vector Search, failing with DeepResearch, and succeeding with Cursor.

  • Research •

    Comparing LLM-Based vs Traditional Clustering for Support Conversations

    Comparing LLM-Based vs Traditional Clustering for Support Conversations

  • Note •

    Sharing Research

    I've decided it may be useful to share some of the research I've done in the past.

  • Article •

    Attention is Not All You Need

    Attention is Not All You Need. You actually need quite a lot.

  • Research •

    AI in Software Development: Accelerating Towards the Next Bottleneck

    Explore how AI in software development is accelerating towards the next bottleneck.

2024

2023

2022