Field notes from a forward-deployed human in an agentic software factory, shipping products at scale.

By Erik Benjaminson, Founder of Sapient Technology Group

Featured

• 01
  

  The Frikadeller Trial: How AI Judges Art

  21 Apr, 2026

  I put Claude Opus 4.7 on trial against 4.6 using the Delphi method and a three-judge LLM panel, before migrating production. Here is what the protocol found.

Recent

• 02

  Intake Is the Architecture

  17 Apr, 2026

  Four intake pipelines feed one search layer on a $200 gaming laptop, turning Karpathy's LLM Wiki into a personal knowledge system that already knows you.

• 03
  

  Turning on the lights

  3 Apr, 2026

  OTEL observability on Claude Code sessions revealed 277,000 wasted tokens in one workflow. The grep-before-read fix took fifteen minutes.

• 04

  So the Recipes Have Soul

  21 Mar, 2026

  I built an autonomous AI research department to produce deep cultural research on culinary pioneers. Here's how it works and what it found.

• 05

  I've Seen This Before

  20 Mar, 2026

  A builder recognizes the same paradigm-shift feeling from March 2020, except now AI agents have removed the ceiling and everything is skill issue.