Technical
Checking My Dark Factory's Receipts After 4,000 Runs
Eight days after pruning 65% of Leaf's prompts and adding Dream telemetry: 655 planning rooms cost 40% less, 91% of bugs self-healed, and 1 metric lied.
Recent Entries
Technical
Why the Best AI Architect Isn't the Best Coding Builder
An accidental 1,199-room randomized trial in Leaf showed Opus 5.5 halves downstream rework as an Architect, GPT-6.1-Sol wins builds, and overrides lose.
Technical
40 Billion Tokens In: Why 28% Agent Rework Is a Feature
At 39.9B tokens and 4,044 runs in Leaf, catching a lying rework denominator proved why 28.3% Tech Lead rejections make 99.6% unattended merges work.
Technical
Six Agentic Development Rules That Survived 2026
Six agentic development rules from a year of building software factories, 15.8B tokens, and deleting harnesses when the data pushed back.
Technical
Inside my Dark Factory hardware stack: 48 Agent Pods on 2019 Mini-PCs
Inside my dark factory hardware fleet: 6 mini-PCs, 48 pod slots, why a 2.6x faster Zen 4 CPU only runs CI 1.17x faster, and why AWS costs $20k-$33k/year.