The Field Exports Instruments and Never Imports the Question
Pattern · geography_cartography · hypothesis · register: public · reviewed: tish 2026-07-13
Geography’s trade balance runs one direction. The exports — instruments, algorithms, data — leave
constantly: earth-observation methods get lifted into the rest of machine learning as a matter of
routine. The imports — the representational questions ML could send back — never clear customs.
Exhibit A is LiDAR: geography invented it for forestry, and every other industry now uses it in a
more novel way — self-driving cars, building information modeling, an entire consumer scanning
ecosystem — while the average GIS technician has never processed a point cloud. The field invents
the instrument, loses the frontier, and doesn’t retain operating competence; the export includes
the expertise. Exhibit B is the missing patron: there is no Microsoft or Google footing the bill
for spatial foundation models, and the would-be patron runs extraction-shaped business at
mom-and-pop competence — all of C.lawful-evil’s appetite, none of the efficiency that lets
lawful evil launder itself.
Sibling of P.spatial-stats-got-there-first (held), pointed the other way: the field that
anticipated ML’s math now can’t hold its own instruments, because the exceptionalism that
guards the gate (P.spatial-is-special-is-the-problem) also keeps the gatekeepers from
leaving the booth.
Connections
- →
P.spatial-stats-got-there-first(held) · is_sibling_of - → P.spatial-is-special-is-the-problem · shares_cause_with
- → C.lawful-evil · inverts
- → C.corporations-as-paywalls · conceptually_related_to
source: live code session (2026-07-03/04) · public register; the internal record is unchanged · synthesis: Claude, 2026-07-12