diff --git a/ideas/effects-growth-flywheel.org b/ideas/effects-growth-flywheel.org
index 8a8e735..8ae0cf5 100644
--- a/ideas/effects-growth-flywheel.org
+++ b/ideas/effects-growth-flywheel.org
@@ -36,7 +36,7 @@ Key observation: the verified API gateway decouples the effect from triad adopti
 ** Phase 1 (10² → 10⁴ instances, months–years)
 
 | Instance count | Effect that starts | Growth driver generated |
-+---------------+-------------------+------------------------
+|-------+-------------------+------------------------|
 | 100–500 | /Regulation as code:/ first regulator encodes a rule as a gate | All regulated entities under that regulator must adopt Passepartout — step function in demand |
 | 500–2K | /AI safety shift:/ gate rule verification becomes expected in enterprise AI procurement | Every company buying AI services requires a proof log — API gateway demand explodes |
 | 2K–10K | /Proof library compounding:/ the collective regression suite has enough edge cases to be qualitatively better than any solo library | Competitive advantage for adopters — those not on the network fall behind on verification coverage |
@@ -46,7 +46,7 @@ Key observation: regulation-as-code creates a /step function/ in demand. Before
 ** Phase 2 (10⁴ → 10⁶ instances, years)
 
 | Instance count | Effect that starts | Growth driver generated |
-+---------------+-------------------+------------------------
+|-------+-------------------+------------------------|
 | 10K–50K | /Computational trust:/ PDS model makes surveillance advertising visibly obsolete | Consumer demand for PDS — "why does my bank still own my data?" |
 | 50K–200K | /Verification cachet:/ /I verify/ becomes a resume signal | Developer adoption accelerates — not from enterprise mandate but from peer pressure and cultural norm |
 | 200K–1M | /Attestation marketplace:/ verifiable reputation has enough data to be reliable | Insurance products become viable — insurers price unverified code higher |
@@ -56,7 +56,7 @@ Key observation: the shift from enterprise adoption to consumer adoption is cult
 ** Phase 3 (10⁶ → 10⁹ instances, years–generations)
 
 | Instance count | Effect that starts | Growth driver generated |
-+---------------+-------------------+------------------------
+|-------+-------------------+------------------------|
 | 1M–10M | /Insurance loop closes:/ premiums for unverified code are 10× verified | Economic necessity drives adoption — not engineering preference, not regulation, but /cost of doing business/ |
 | 10M–100M | /Verification monopoly:/ regulator references the early player's gate library | New entrants cannot compete with the installed proof base — the moat compounds with every new instance |
 | 100M–1B | /Compute as geopolitical asset:/ nations run triad instances for digital sovereignty | Nation-state procurement — 100M to 1B happens via government mandate, not organic adoption |