Nosana (NOS) — the revision ledger

Generated 2026-09-02T18:47:00Z from live ClickHouse queries against this estate's ledger. Point-in-time archive held for Nosana: 9,933 observations · 43 distinct series · 5 upstream sources · first row 2026-08-09 17:15:20 · newest row 2026-09-02 18:38:03 · 9,933 of 9,933 rows carry a verify_url · 39 series refreshed in the last 24h (896 fresh rows). Nothing on this page is cached, estimated, or remembered.

Revision ledger — 0 backward revision(s) recorded for NOS

Queried live at generation time: default.depin_revision_ledger holds zero rows for NOS (raw detector output for this network: 0 rows). /v1/revisions?symbol=NOS cross-check passed: the served API reports 0 row(s), matching the deduped ledger count rendered on this page.

What we monitor for NOS — live monotonicity of every held series

The detector only judges series that are declared cumulative-by-construction: a counter that cannot legitimately decrease. A gauge (price, market cap, staked supply, staker counts, queue depths, holder concentration) legitimately falls when the market or the fleet moves — that is not a revision and is never recorded as one. For every daily-snapshot series we hold for NOS (50 metrics, 800 rows, 27 snapshot dates), this is its measured state across the full archive, queried live at generation. The 2 detector-judged counter series show 0 backward moves in total — that is the measured content of the zero above. The 48 gauge-class series move both ways by nature and are excluded from the detector by design.

Classification is imported from the detector at generation (propintel/revision_ledger.py, its CUMULATIVE and NOT_TRULY_CUMULATIVE sets) — this page does not maintain its own opinion. Resolution rule: latest observation wins for current-state reads; the superseded value is retained, never deleted — retaining the before-value is the only reason a restatement is detectable at all.

Same-stamp rewrite check on the raw stream

The daily-snapshot detector is not the only rewrite vector. The raw stream (default.depin_onchain) was queried for duplicate (metric, as_of) rows carrying two different values — the same-timestamp rewrite pattern that hit Arweave's BLOCKS counter (1,988,930 → 1,702,206 between the 2026-08-27 and 2026-08-28 snapshots). For NOS the live count at generation is 0: no same-stamp rewrite exists in the held archive.

The ledger, live — every restatement recorded estate-wide (18 deduped events)

NOS is currently the clean case. These networks are not: upstream sources for AR, AUKI, EIGEN, FLOCK, GRT, OVPP, R1, SOGNI, STORJ replaced already-published values after this estate recorded the original. Deduped on revision identity exactly as /v1/revisions dedupes it; rendered live from ClickHouse at generation, not copied from a snapshot. The superseded value is retained, never deleted; that retention is the only reason a rewrite is detectable, and it is the part every overwriting vendor discards.

How detection works — and what it cannot see

The detector (propintel/revision_ledger.py) diffs every daily snapshot this estate holds of every declared cumulative-by-construction series — counters that cannot decrease unless upstream rewrote them (alltime fees, token transfers, completed jobs, rewards). For NOS the declared set is JOBS_COMPLETED, JOBS_USD_REWARD. A row enters default.depin_revision_ledger iff two snapshots disagree for an already-observed date; zero human judgement. Resolution rule: latest observation wins for current-state reads, the superseded value is retained, never deleted. Population gauges (node counts, stakers, prices, queue depths, utilization) legitimately move both ways and are deliberately excluded — a drop there is a real population change, not a rewrite. Honest boundary: this catches value-history rewrites ex-post; identical-schema semantic drift with plausible values is what the hash-pinned derivation_replay layer is for. The detector's full classification is inspectable at /v1/revisions and frozen daily at /revision-ledger/REVISIONS_20260902.json.

Verify any cell yourself — the exact commands

No account, no key, no trust required. Commands 1–2 hit this estate's public API (the same rows this page queried — every row carries source / as_of / verify_url); command 3 hits Nosana's own dashboard stats endpoint, the first-party source of the staking and jobs family; command 4 is the estate's graded network row; command 5 is the frozen daily snapshot of the ledger. One polite call per endpoint, no hammering.

# 1) The served revision feed for NOS — the same rows this page queried, live:
curl -s "https://kairossignal.com/v1/revisions?symbol=NOS" | python3 -m json.tool
# 2) The estate-wide feed (every network, both sides of every diff):
curl -s "https://kairossignal.com/v1/revisions" | python3 -m json.tool | head -80
# 3) Nosana's own dashboard stats endpoint (the first-party source of the staking/jobs family):
curl -s -H "User-Agent: kairos-check" "https://dashboard.k8s.prd.nos.ci/api/stats" | python3 -m json.tool | head -40
# 4) This estate's public API row for NOS (grades + provenance):
curl -s "https://kairossignal.com/v1/network/NOS" | python3 -m json.tool | head -60
# 5) The public daily snapshot of the ledger (frozen copy):
curl -s -o /dev/null -w "%{http_code}\n" "https://kairossignal.com/revision-ledger/REVISIONS_20260902.json"

Method & honesty notes. All counts are queries against the live ledger at generation time, not restamped copy — this page is regenerated by gen_nos_revision_page.py and the hourly claim_drift_check.py run grades its claims against the live API like every other published surface. The NOS revision count is the deduped live result of SELECT ... FROM default.depin_revision_ledger WHERE network='NOS'; the page refuses to ship if the served /v1/revisions?symbol=NOS disagrees with it. The 18 estate-wide events shown are deduped on revision identity exactly as the API dedupes them; detected_at is the first detection and never moves. The zero-restatement claim is backed on this page by the per-series monotonicity table (queried live) and the same-stamp rewrite check. Series absent from a section are absent, not zero. Upstream probes ran serially by this generator immediately before rendering: dashboard.k8s.prd.nos.ci/api/stats (first-party) HTTP 200; estate /v1/revisions?symbol=NOS HTTP 200; estate /v1/network/NOS HTTP 200.