The default answer to "I need more from my AI" is still a bigger model on a bigger GPU. But every piece of the other answer, horizontal scaling for inference, already exists in the research. Here's how small, specialized, cache-friendly workers make AI cheap instead of scarce.
When a system runs out of headroom, you've got two options: buy a bigger box, or run more boxes. We settled that question for the whole internet a long time ago; the web runs on adding cheap…
Bolting AI onto an organization built for walking pace doesn't make it faster. The bottleneck was never the typing. It was coordination, and AI only shone a light on it.
There's a genre of AI blog post that I skip as soon as I recognize it: the company retrospective where they handed everyone Copilot or Claude Code, waited a couple of quarters, and came away…
Why coding agents need a sandbox the agent cannot reach, an audit log the agent cannot write, and signed instructions the agent cannot rewrite.
Every coding-agent sandbox I've set up ends in the same tired loop. I lock it down, the agent stops being useful, I loosen it, and I end up exactly where I started: YOLO mode. At best there is a…
Trans Day of Visibility, 2026. Being myself is apparently political now.
I built an open-source pipeline where 7 specialized Claude agents research, build, validate, and ship side project ideas, coordinated through plain files on disk, with human approval gates between phases.
I have a side project graveyard. You probably do too. The pattern is always the same: idea hits at 11pm, I start building immediately, skip the "does anyone actually want this" step, burn a weekend…
I built a WoW-inspired raid frame on a 64x64 LED panel to monitor my AI coding agents. When an agent needs approval, it catches fire.
I run up to ten AI coding agents at a time. I wanted to see what they were all doing. So obviously I built a WoW raid frame on a $25 LED panel. Each agent gets an 8x8 sprite. When it's working, you…
Prompt engineering is what happens in the moment. Context engineering is what you do over time to keep your agents from drifting off a cliff.
Large language models are trained on massive amounts of text. During training, the model learns patterns and stores them as weights. Not rules, not lookup tables. Weights: billions of numerical…
I spent two weeks using GPT and Codex for my daily work. Between fabricated technical explanations, a model that insisted it could hear birds, and tooling that couldn't keep up with my workflows, I learned a lot about trust. Here's what happened.
I was at a friend's apartment, testing GPT-4o's voice mode. She had lovebirds, and they were chirping away. "Can you hear the birds?" I asked. "Yes, I can hear them," GPT said. This surprised me. I'd…
How a self-sabotaging navigator taught me that agent identity is performed, not declared - and why Judith Butler's philosophy maps onto multi-agent systems better than I expected.
Last week, Claude Code was about to open a PR for a feature I never asked for. Not a slightly-wrong implementation. A completely different feature. I'd just spent an hour planning the task with my…
Text files resembling a knowledge base combined with agentic interfaces are a simple, but powerful, tool. They can update themselves and can be easily engineered by humans as well. Here's how Navigators work and why they matter.
I've been using Navigators for my platform engineering work and they've completely changed how I interact with AI. Keeping track of work, planning implementations, and learning over time and across…
Your game saves in this browser.
Too small to play
Make this window bigger to carry on, or go back to the site (Esc).