docs: fix ECONOMICS.md's stale pricing source-of-truth

[wikithoughts]

What

docs/ECONOMICS.md's pricing paragraph pointed readers exclusively at the legacy
in-process demo path (council/coordinator.py / council/run_demo.py) as if its fixed
WORKER_POOL = 30/JUDGE_FEE = 5 constants were the live network's pricing. Per D50 in
docs/DECISIONS.md, the live network (council/net/) actually prices every job through
council/net/config.py's pool_for(job_type, n_minds) — a single source of truth that
scales the per-mind price by responder count and each job type's pool_mult weight
(e.g. chat 1x, research_report 3x, assisted 5x), driven by the same-defaulted
PW_WORKER_POOL/PW_JUDGE_FEE env vars.

Fix

Rewrote the paragraph to:

• keep STARTER_ALLOWANCE = 100 (confirmed as PW_STARTER_CREDITS in council/ledger.py)
  as shared by both paths
• explicitly distinguish the legacy demo's flat per-job constants from the live network's
  pool_for()-scaled pricing
• point readers at GET /job-types on a live coordinator as the actual current numbers
  for any deployment, rather than this doc

Docs-only change — no code touched.

Verification

• ruff check . → All checks passed
• pytest tests/ -q → 507 passed, 1 skipped (0 failed) — this box is heavily loaded
  (many concurrent agent sessions/CI runners sharing codeagents.slice), so the run took
  ~10 minutes instead of the documented ~5s, but completed clean.

🤖 Generated with Claude Code