Interactive research
When can a local model take over?
How the task and cost of mistakes change which local Qwen size is good enough. Compare the measurements and explore your own cost assumptions.
Explore the resultsWriting
Essays on what to build, how to evaluate it, and what changes when the cost of doing the work falls.
Companion API study
GPT-6 Luna Decisions: fast choices, uneven results
The Decisions API compared with Responses and archived Jev results. Where it worked, where dependent choices failed, and what remains untested.
Read the API studyEssay
When Is a Cheaper Model Actually Cheaper?
Sol versus Luna: compare token fees, the cost of mistakes, and the work required to find out.
Read the essayEssay
How to Be Freakishly Decisive
Choose now, name what could beat your choice, research when it pays, and update. Six yes/no questions to run on your next decision.
Read the essayWhen a Free Model Can Replace the Frontier (and How to Prove It)
An experiment in deciding when a local model can take over a classification task, accounting for errors in the frontier model used to evaluate it.
The Eighth Ledger: When Intellectual Labor Becomes a Capital Asset
Published draft, still being revised
What earlier technology transitions suggest about turning recurring knowledge work into owned software pipelines.
Any Sufficiently Advanced Technology: A Taxonomy of Magic
Every sufficiently advanced technology collapses one friction and reveals the new scarcity behind it. LLMs collapsed the cost of language, putting commodity intellectual output in the blast radius. A 6-step protocol, calibrated on 700 years of transitions from the printing press to the internet, for finding the new scarcity and moving first.
What AI Actually Made Scarce
AI made production cheap and revealed the constraint it was hiding: selection. Green dashboards stay green while quality quietly decays. A 30-minute protocol to find your new scarcity.