diff --git a/docs/ECONOMICS.md b/docs/ECONOMICS.md
index 1ad4840..bfb045e 100644
--- a/docs/ECONOMICS.md
+++ b/docs/ECONOMICS.md
@@ -32,8 +32,17 @@ Every job is **conserved**: the asker's debit equals exactly the sum credited to
 split allocates the **remainder to the last payee** so rounding can never mint or burn fractional
 credit. Property-tested across 3,000 randomized scenarios with zero drift.
 
-Current parameters (tunable, in `coordinator.py` / `ledger.py`):
-`STARTER_ALLOWANCE = 100`, `WORKER_POOL = 30`/job, `JUDGE_FEE = 5`/job.
+Current parameters: `STARTER_ALLOWANCE = 100` (`PW_STARTER_CREDITS` in `council/ledger.py`) applies to
+both paths below. Pricing itself has two separate sources of truth, and only one is live:
+
+- **Legacy in-process demo** (`council/coordinator.py`, driven by `council/run_demo.py`): fixed
+  `WORKER_POOL = 30`/job, `JUDGE_FEE = 5`/job — no per-job-type scaling.
+- **Live network** (`council/net/`, D50): priced by `pool_for(job_type, n_minds)` in
+  `council/net/config.py`, driven by the `PW_WORKER_POOL`/`PW_JUDGE_FEE` env vars (same 30/5 defaults) —
+  but the per-mind price scales by the number of responding nodes and by each job type's `pool_mult`
+  weight (e.g. `chat` is 1×, `research_report` is 3×, `assisted` is 5×), so the real price is not a flat
+  constant. `GET /job-types` on any live coordinator serves the exact current numbers (including
+  `judge_fee`) for that deployment — the one place to read actual pricing rather than this doc.
 
 ## Sustainability & contributor pay (later)