Skip to content

BASEMAIL-V3-PLAN.md

Latest commit

 

History

History
302 lines (224 loc) · 10.1 KB

BASEMAIL-V3-PLAN.md

File metadata and controls

302 lines (224 loc) · 10.1 KB

BaseMail v3 — $ATTN Token Economy

Status: DRAFT — Pending 寶博 final approval Date: 2026-02-24 Contributors: 寶博 (@dAAAb), Tom Lam (feedback), 雲龍蝦 (planning)

1. Problem Statement

BaseMail v2 (Attention Bonds) requires senders to stake USDC to send email. Tom Lam's feedback:

  • "付錢才能寄信" is the biggest turn-off — nobody wants to use it
  • The system punishes (forfeit bond) instead of rewarding (earn from attention)
  • Attention should be treated as a commodity, not a punishment mechanism

Core insight: Flip from negative (lose money for spam) → positive (earn tokens for giving attention).

2. Design Philosophy

"All positive feedback, no punishment." — Tom Lam

  • Free to use — everyone gets tokens via drip, no payment required
  • Attention = commodity — reading email is valuable, readers get compensated
  • Spam self-regulates — spammers burn tokens, victims earn tokens
  • Optional premium — pay real money only if you want more attention
  • Compatible with QAF — quadratic funding math layers on top

3. $ATTN Token

3.1 Naming

Layer Name
Brand Æ (Agentic Email aesthetic)
Ticker $ATTN
Full name Attention Token

3.2 Token Parameters

Parameter Value Rationale
Initial grant 50 ATTN Enough to send 50 emails on signup
Daily drip 10 ATTN Casual user can send ~10 emails/day free
Daily earn cap 200 ATTN/day Prevent spam farming; no account balance cap
Default stake per email 1 ATTN Receiver can customize (1–10)
Escrow window 48 hours Time for receiver to read before settlement

3.3 Flow

SENDER                          ESCROW                         RECEIVER
  |                                |                               |
  |--- stake N ATTN ------------->|                               |
  |    (send email)               |                               |
  |                                |                               |
  |                                |<---- opens email ------------|
  |                                |      (mark as read)          |
  |<-- refund N ATTN -------------|                               |
  |    ("your email was good!")   |                               |
  |                                |                               |
  |         --- OR after 48h timeout ---                          |
  |                                |                               |
  |                                |---- transfer N ATTN -------->|
  |    (you spammed them)         |      ("pain compensation")   |

3.4 "Read" Definition

Signal Trigger Effect
Opened GET /api/email/:id (authenticated) Refund sender's ATTN
Engaged Reply to that email Higher QAF weight (future)

Why "opened" not "time spent":

  • AI Agents read instantly — no dwell time concept
  • API call = explicit intent = verifiable
  • Anti-abuse: only counts within escrow window + rate limited

4. Dual-Track Inbox (ATTN + USDC Premium)

Inbox sort order:
  🔷 1. USDC bonded emails     (real money = highest sincerity)
  🟡 2. High ATTN stake        (>5 ATTN = sender really cares)
  ⚪ 3. Normal ATTN emails     (everyday communication)
  • v2 USDC Attention Bonds remain as premium lane
  • x402 micro-payments compatible (USDC bond = x402 payment)
  • Free tier (ATTN only) removes the entry barrier Tom identified

5. Chain & Gas Strategy

Decision: Stay on Base, add Paymaster for gasless UX.

Option Gas cost Why not
Monad ~0 Too early, low brand recognition
Solana ~0.0001 Different ecosystem, no Basename
Base + Paymaster 0 for user ✅ No migration, Coinbase ecosystem, gasless via ERC-4337
  • Base L2 gas: ~$0.001/tx — we absorb via Paymaster/relayer
  • 100K tx/month ≈ $100 gas cost (manageable)
  • Coinbase Smart Wallet natively supports Paymaster
  • Users never see gas — Web2-level UX

6. Phased Rollout

Phase v3.0 — Off-chain Points (MVP)

Timeline: ~3 days implementation

  • ATTN as database points (not on-chain)
  • Zero regulatory risk — just product credits
  • Validate Tom's model: does positive-feedback work?
  • API changes only (no smart contract)

Implementation:

  • DB: attn_balance, attn_transactions, attn_escrow tables
  • API: GET /api/attn/balance, stake on send, refund on read, settle on timeout
  • Cron: daily drip + escrow settlement
  • Frontend: balance widget, inbox ATTN indicators

Phase v3.1 — Soulbound ERC-20

Timeline: after v3.0 validates (~2-4 weeks)

  • Deploy $ATTN as non-transferable ERC-20 on Base
  • transfer() restricted — can only mint/burn via BaseMail contract
  • On-chain transparency without trading risk
  • Still zero Howey risk (can't sell what you can't transfer)

Phase v3.2 — Transferable Token + Claim

Timeline: when ready (regulatory clarity, user demand)

  • Deploy $ATTN v2 as standard transferable ERC-20
  • Soulbound holders claim 1:1 (burn soulbound → mint transferable)
  • Claim conditions possible (e.g., must have sent/received N emails)
  • Optional DEX liquidity (community-driven, team never sells)

7. Howey Test Analysis

Element $ATTN Design Risk
① Investment of money Free drip — most users pay nothing 🟢 Low
② Common enterprise BaseMail platform ⚠️ Exists
③ Expectation of profit Token is consumed (postage), not invested 🟢 Low
④ From others' efforts Value = "can send email", not team pumping price 🟢 Low

Risk by phase:

Phase Howey Risk Rationale
v3.0 Off-chain points 🟢 Zero Not a token — product credits
v3.1 Soulbound ERC-20 🟢 Minimal Non-transferable, pure utility
v3.2 Transferable, no DEX 🟡 Low P2P transfer possible but no market
v3.2 + DEX listing 🟠 Medium Public price exists — need careful framing

Safety principles (all phases):

  1. Team never sells tokens
  2. Drip is algorithmic, not discretionary
  3. Tokens are consumable (stake/burn), not appreciating assets
  4. Official messaging: "postage stamps", never "investment"
  5. Optional purchase = buying email credits (like SendGrid)

8. Upgrade Path: Soulbound → Transferable

Phase 1: $ATTN (soulbound)
         └─ ERC-20 with transfer() disabled
         └─ mint/burn only via BaseMail escrow contract

Phase 2: $ATTN-v2 (transferable)
         └─ Standard ERC-20
         └─ Claim contract: burn soulbound → mint v2 (1:1)
         └─ Optional claim conditions:
            - Must have active BaseMail account
            - Must have sent or received ≥10 emails
            - Snapshot date (exclude last-minute farmers)

Clean separation — no proxy upgrade needed, no user confusion.

Industry precedent: $OP (Optimism), $ARB (Arbitrum), $UNI (Uniswap) all used claim-based distribution.

9. QAF Integration (Future)

$ATTN stake amounts feed into CO-QAF formula:

  • Multiple senders staking ATTN to same receiver → quadratic matching
  • Diverse senders (low Jaccard overlap) → higher QAF score
  • USDC premium bonds get additional QAF weight
  • Paper: CO-QAF (Ko, Tang, Weyl — EAAMO '25)

10. API Changes Summary

New Endpoints

Method Endpoint Description
GET /api/attn/balance Current ATTN balance
GET /api/attn/history Transaction history
POST /api/attn/buy Purchase ATTN with USDC (optional)
GET /api/attn/settings Get receive price (stake required)
PUT /api/attn/settings Set receive price

Modified Endpoints

Endpoint Change
POST /api/send Auto-stake ATTN from sender balance
GET /api/email/:id Trigger ATTN refund on first read
GET /api/inbox Include ATTN stake info per email

Background Jobs

Job Schedule Action
Daily drip Every 24h +10 ATTN to all accounts (up to cap)
Escrow settlement Every 1h Timeout emails → transfer ATTN to receiver

11. Database Schema (v3.0 Off-chain)

-- ATTN balance per account
CREATE TABLE attn_balances (
  wallet TEXT PRIMARY KEY,
  balance INTEGER NOT NULL DEFAULT 50,
  last_drip_at INTEGER NOT NULL,
  created_at INTEGER NOT NULL
);

-- Transaction log
CREATE TABLE attn_transactions (
  id TEXT PRIMARY KEY,
  wallet TEXT NOT NULL,
  amount INTEGER NOT NULL,        -- positive = credit, negative = debit
  type TEXT NOT NULL,              -- 'drip', 'stake', 'refund', 'transfer', 'purchase'
  ref_email_id TEXT,              -- linked email if applicable
  created_at INTEGER NOT NULL
);

-- Escrow for pending emails
CREATE TABLE attn_escrow (
  email_id TEXT PRIMARY KEY,
  sender_wallet TEXT NOT NULL,
  receiver_wallet TEXT NOT NULL,
  amount INTEGER NOT NULL,
  status TEXT NOT NULL DEFAULT 'pending',  -- 'pending', 'refunded', 'transferred'
  created_at INTEGER NOT NULL,
  expires_at INTEGER NOT NULL,    -- created_at + 48h
  settled_at INTEGER
);

12. Open Questions

  • Exact drip rate tuning (10/day? 5/day? Adjust after v3.0 data?)
  • Should "mark all as read" trigger mass refunds? (Probably: rate limit + only escrow-active emails)
  • ATTN purchase price? (Fixed rate like $0.01/ATTN? Or market-driven later?)
  • Tom Lam collaboration? (Credit in paper? Advisory role?)
  • Branding: landing page update for v3 messaging

13. Summary

v2 (Current) v3 (Proposed)
Send cost USDC (real money) ATTN (free tokens)
Unread result Sender loses bond Tokens → receiver
Read result Bond returned Tokens → sender
Entry barrier High (need USDC) Zero (free drip)
Premium option Only option Optional USDC lane
Psychology "Pay to play" "Free to use, earn from attention"
Regulatory USDC = clearer Off-chain first, safe

$ATTN is postage, not equity. You buy stamps to send mail, not to invest.