Written down before it changes again
Everything here expires. We print the date on it.
Writing about AI tooling goes stale faster than anyone admits. So every post on this site carries a half-life meter: how old it is, and whether you should still act on it. The current shape of the stack lives on the board — models, harnesses and frameworks.
Latest writing
-
Harnesses
Harness Engineering Lessons from Production AI Chatbots
Real technical lessons from deploying AI chatbots for client websites: prompt architecture, context management, fallback chains, cost control, and the harness patterns that actually matter.
half-life 90dfrom 14 Sep 2026
-
Models
GPT-6 Astra needs less prompting than you think.
OpenAI's guidance for GPT-6 Astra is clear: shorter skills, fewer approval gates, less hand-holding. The models got better. Your prompts should get smaller.
half-life 60dfrom 15 Sep 2026
-
Pattern
Skills frameworks are everywhere. Here is what makes one worth using.
OpenAI, Hermes, HumanLayer, and a dozen GitHub repos all shipped skills systems in the same month. The pattern is clear. The implementations are not. Here is what separates a good skill from a bad one.
half-life 90dfrom 12 Sep 2026
-
Security
MCP security is an unsolved problem. Here are the four attack surfaces.
Three papers published in August-September 2026 show that every MCP deployment has at least four distinct attack surfaces. Here is what they found, and what to do about each one.
half-life 90dfrom 11 Sep 2026
-
Tools
Engrim: local-first memory that survives context clears
Every time you clear your agent's context, months of architectural decisions vanish. Engrim fixes that with a local SQLite store that works across Claude Code, Cursor, Antigravity, and Codex.
half-life 90dfrom 09 Sep 2026
-
Protocols
NLIP, A2A, MCP: the three protocols your agents actually need
Three protocols claim to solve agent interoperability. They address different layers. Here is what each one does, where they overlap, and what to build on today.
half-life 180dfrom 08 Sep 2026
-
Harnesses
How we built a 24/7 AI enquiry assistant for static sites
Static sites cannot run server-side logic, but clients wanted an AI chatbot that could handle enquiries, qualify leads, and send emails. Here is the architecture and what worked.
half-life 90dfrom 07 Sep 2026
-
Tools
DaeOne: One Interface to Run Your Entire Website
Website ownership is fragmented — analytics, email, SEO, support, all in different tools. DaeOne unifies everything behind a single AI-powered dashboard.
half-life 90dfrom 05 Sep 2026
-
Models
Local LLM inference in September 2026: what actually works
Running models on your own hardware is no longer a compromise. Here is the current state of local inference — tools, hardware, and where the trade-offs actually land.
half-life 45dfrom 04 Sep 2026
-
Harnesses
Harness Engineering is now a discipline. Here is what that means.
The model is becoming commoditized. The harness — everything around the model — is where the engineering value lives. Three data points prove it, and the implications are immediate.
half-life 90dfrom 27 Aug 2026
-
Models
OpenRouter: the LLM gateway that lets you test 400+ models with one API key
OpenRouter gives you 409 models from 59 providers behind one API key. 18 free models including frontier-class options. The practical starting point for autonomous agents and AI frameworks.
half-life 60dfrom 13 Aug 2026
-
Models
NVIDIA Nemotron 3 Ultra: frontier reasoning, built-in safety, and completely free
NVIDIA's Nemotron 3 Ultra matches frontier models at 3-10x fewer active parameters, with safety guardrails baked into training — not bolted on. Free on OpenRouter.
half-life 90dfrom 13 Aug 2026
-
Guide
AI Model Price-to-Performance Guide
A living comparison of 400+ models across five price tiers. Updated every Monday with fresh OpenRouter pricing, new releases, and benchmark data.
half-life 45dfrom 13 Aug 2026
-
Models
What Claude Fable 5 actually does differently
Anthropic's most capable widely released model runs with thinking always on and a 1M context. Here is what that changes about how you build on it.
half-life 60dfrom 12 Aug 2026
-
Pricing
Xiaomi MiMo: what you get for the money
A 1M-context omnimodal model at roughly a fiftieth of frontier pricing. Where that trade is worth making, and where it very much is not.
half-life 45dfrom 11 Aug 2026
-
Harnesses
Anatomy of an agent harness
A model is not an agent. The gap between them is six pieces of unglamorous engineering, and almost every agent that fails in production fails in one of them.
half-life 90dfrom 10 Aug 2026