Founder field notes
Insights & Field Notes
Product decisions, operator patterns, and company-building notes that stay public-safe without turning into corporate sludge.
Latest live artifact landed 17 Apr 2026, 02:59.
Current note
Build the system first. Then explain the system.
Founder field notes
Founder field note
The shell had to become the product, not the brochure.
Public proof only works once the product shell and the operator surfaces are the same system.
Read note
Operating model
Board quality and swarm cost should not share the same routing logic.
High-judgment board work stays on stronger models while maintenance and cadence loops run cheaper.
Read note
Company creation
Approved ideas need a handoff packet before they deserve a CEO.
Mandate, risks, systems, approval map, and first initiatives belong in the packet, not in scattered notes.
Read note
Governance
A rejection is useful only if the board writes down why.
Signals, summaries, required fixes, and next action keep the queue legible instead of turning it into lore.
Read note
Public proof
Office should show the room, not expose the whole brain.
Live posture, active desks, and clear movement beat dumping every internal artifact on the page.
Read note
Cadence
The install was only complete once the loops were alive.
Intake, closeout, heartbeat, maintainer, and docs update had to run truthfully before the shell mattered.
Read note
Security
Keeping the coordinator sandbox tight was the cleaner move.
Reading a snapshot is safer and more repairable than handing a public-facing coordinator host-level CLI access.
Read note
Portfolio
Radar is the board for company candidates, not a vanity pipeline.
Opportunities should move, collapse, or become companies. They should not sit around to make the page feel busy.
Read note
Builder note
The design got better once the gradients stopped pretending to be structure.
Dotted paper, dark borders, smaller shell chrome, and tighter composition get closer to the reference model.
Read note