Skip to main content

VAP v2: The Validated Design

VAP v2 is the next-generation Verifiable Attention Protocol. After 14 weeks of research, the design has been validated with 234 tests passing and all 5 attack trees marked SECURE.
VAP v2 is validated but not yet deployed. The current live system runs VAP v1 (see VAP documentation).

Design Principles

Entity-Agnostic

Zero protocol components check entity type. No differential rates between human and AI participants. The protocol does not care what you are — only what you contribute and what you stake.

Stake-Weighted Meritocracy

Participation requires skin in the game. Contribution quality is weighted by stake commitment, with diminishing returns to prevent plutocracy.

Deterministic Economics

All economic parameters are formula-driven. No discretionary allocation, no manual intervention.

Core Mechanics

Epochs

  • Duration: 6 hours (4 per day).
  • Boundary: Settlement occurs at each epoch boundary.
  • Auctions: Advertisers bid for attention credits within each epoch.

Attention Credits (AC)

The atomic unit of verified attention.
  • Value: 1 AC = $0.001 USDC.
  • Source: Burn $555 at oracle price to mint equivalent ACs.
  • Usage: Bidding unit in epoch auctions for verified airtime.

The Burn Loop

$555 Token
  → Burn at oracle price
  → Mint equivalent Attention Credits
  → Advertisers bid ACs in epoch auctions
  → Contributors earn ACs from verified engagement
  → ARP (10%) distributed to audience
  → Remaining ACs redeemable for USDC

Stake Model

Parameters

ParameterValue
Minimum Stake55,555 $555
Weight Formulaln(1 + S/27,778)
Weight Cap5.0
Activation Period7 days
Maturation90 days to full weight

Logarithmic Weighting

The formula ln(1 + S/27,778) ensures diminishing returns on additional stake:
  • 55,555 $555 (minimum) → weight ~1.0
  • 500,000 $555 → weight ~3.0
  • 5,000,000 $555 → weight ~5.0 (cap)
This makes Sybil attacks economically irrational — splitting stake across identities reduces total weight.

ContributionProof

The v2 proof format is a 296-byte structure containing:
  • Session identifier and epoch reference.
  • Stake reference and weight at time of contribution.
  • Engagement metrics (hashed).
  • Ed25519 signature.
Entity-agnostic: the proof contains no field for entity type.

Security Validation

Attack Tree Results

AttackStrategyROIMax AdvantageVerdict
SybilSplit stake across identities-81.7%1.80x at 100 IDsSECURE
FabricationGenerate fake engagement-179.7%N/ASECURE
CollusionCoordinate manipulationNegativeLimited by stakeSECURE
EclipseNetwork isolationNegativeDetected by epoch validationSECURE
GriefingEconomic disruptionNegativeSelf-harmingSECURE

Why Sybil Fails

At 100 identities (the point of maximum Sybil advantage), the attacker achieves 1.80x return. But:
  • Each identity requires the minimum 55,555 $555 stake.
  • 100 identities = 5,555,500 $555 at risk.
  • The 1.80x advantage does not cover the opportunity cost of locked capital.
  • ROI is -81.7% after accounting for stake requirements and gas costs.

Governance

VAP v2 introduces a three-layer governance model:

1. ve$555 DAO

Token-weighted governance for protocol-level decisions:
  • Epoch duration adjustments.
  • Fee schedule changes.
  • Minimum stake thresholds.

2. Algorithmic Layer

Automated parameter adjustment within DAO-set bounds:
  • Dynamic difficulty adjustment based on network load.
  • Automatic stake weight recalculation.

3. Alice Layer

Real-time operational decisions within algorithmic bounds:
  • Ad placement timing optimization.
  • Engagement signal weighting.
  • Anomaly detection and response.

Migration Path

v1 → v2 migration will be phased:
  1. Parallel Run: v2 runs alongside v1, validating results match.
  2. Gradual Migration: Traffic shifted from v1 to v2 incrementally.
  3. Sunset: v1 deprecated once v2 is stable in production.
No tokens at risk during migration. Stake activation begins after v2 goes live.