Agents

MAY 29, 2026

Skill Distillation

How a personal AI agent built on markdown skills lets a frontier model teach smaller, local models to do real work, without retraining.

Skill Distillation
MAY 27, 2026

Software After AI

Software is no longer about UX & data. It is about the harness, the layer that turns an LLM into a reliable agent. Seven components define the new stack.

Software After AI
MAY 11, 2026

Localmaxxing

About half of agent tasks can run on a local 35B model. The real advantage isn't cost or privacy — it's latency. 2.1x faster means more iteration cycles per session.

Localmaxxing
FEBRUARY 26, 2026

Is AI Doing Less & Less?

Analysis of 14 production agent workflows reveals the optimal balance between deterministic code & LLM inference. 65% of nodes run as pure code, saving tokens & improving reliability.

Is AI Doing Less & Less?
FEBRUARY 19, 2026

9 Observations from Building with AI Agents

Practical lessons from a year building AI agent systems : prototype with frontier models, fine-tune for stable tasks, use static typing, & run competing agents as critics.

9 Observations from Building with AI Agents
JANUARY 22, 2026

AI Managing AI

The most capable AI models now spend their cycles orchestrating other tools. Specialization creates startup opportunity : build the specialists the executives call first.

AI Managing AI
JANUARY 16, 2026

Software That Debugs Itself While I Sleep

Rather than manually debugging AI failures, I built a Ralph Wiggum loop that pushes the model against its failures each night until it dreams a correct solution.

Software That Debugs Itself While I Sleep