SapientTech.dev 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
-
I built a machine for settling arguments. Then I asked it where to fish.
A multi-agent research council that settles architecture arguments at work, pointed at a personal question: which river, which stretch, which day. The brief is the whole job.
-
How Culinary Explorer Is Built
A field guide to one pipeline: cuisine councils, deep chef research, persona authoring, collection planning, two-pass curation, schema-validated generation, and end-to-end observability.
-
The Frikadeller Trial: How AI Judges Art
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
-
Opus 5 Failure Patterns: An Executive Accounting
A source-bounded executive accounting of ten documented Opus 5 failure patterns, observed over five days across four project contexts, with one comparative transcript study.
-
Intake Is the Architecture
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.
-
Turning on the lights
OTEL observability on Claude Code sessions revealed 277,000 wasted tokens in one workflow. The grep-before-read fix took fifteen minutes.
-
So the Recipes Have Soul
I built an autonomous AI research department to produce deep cultural research on culinary pioneers. Here's how it works and what it found.