Drift is the observatory's core claim: a model's answer to a fixed question, tracked across
time and across releases. Because every result records the exact model id it came from, a
version bump is distinguishable from same-version drift.
First observation recorded. The series so far contains
1 run. Drift is a diff between two readings, so
it begins with run #2 — the next scheduled weekly measurement. Once a second run lands, this
page will show exactly which models changed their minds on which questions, and by how much.
The instrument is designed to make this honest: nothing is extrapolated from a single reading.