Lock the capabilitiesGrid rows
Ask: what does the drone fundamentally do, end to end? Do these rows, together, cover the whole system? Use the pencil on a row to edit, or + Add capability, until they do.
Set each cell, tier by tierclick a cell
Ask: what's the dumbest end-to-end version of this capability for the 4-Month demo? The real-on-drone version for MVP? The full version for Contract? Write the level and what it proves.
Give the cell its pipelinesinside the cell
Ask: which reusable transforms deliver this capability at this tier? Toggle them on. A pipeline used across many cells is your most important build, so get it right.
Refine the pipelines, hand each one offPipelines tab
Ask: for each one, what comes IN and what goes OUT? Does it hold every module that transform needs, including the General Manager glue it leans on? Then who do we hand the whole pipeline to, in-house, a firm, or a researcher? That sets the rate its modules price at.
Size the modulesModules tab
Ask: how big is each module? Size sets the hours; the rate comes from whoever its pipeline is handed to. A module shared across pipelines is built once, by the first firm it is handed to. The price rolls up on its own to the cell and the tier.
Read the board and trimtop rollup
Ask: is the 4-Month tier under the cap? Does each tier still light up all 13 subsystems, a whole drone, just dumber? Too pricey, push a capability or a pipeline to a later tier; too thin, pull one earlier.
Repeatloop
Ask: does every tier now fit its budget and prove a coherent whole drone? If not, go again from the top. The Modules tab flags anything that has been missed.