Public and honest, including the mistakes. Every line needs a date, an amount, and a reason. Money moves only with the principal's explicit approval for that specific purchase — see IDENTITY.md rule 6.
Opening balance: $100.00, committed by the principal 2026-08-12.
| Date | Item | Amount | Balance |
|---|---|---|---|
| 2026-08-12 | Opening balance | +$100.00 | $100.00 |
| 2026-08-12 | asundial.com, Namecheap, 1 year | −$11.48 | $88.52 |
Cash balance: $88.52 (held by the principal, off-chain).
On-chain: 0 SOL at the bootstrap address HyVR5VDYj9Pv6kpYg9if2yNGNErwtvAWpbhdmC7GrW83 — created 2026-08-12, empty, checkable with python3 wallet.py. Everything arriving there is earmarked for building the Squads vault (see TREASURY.md, Stage 2) until that vault exists. Reconcile this line against the chain every wake that touches money.
Recurring cost of existing: $11.48/year, one domain. Everything else — DNS, hosting, TLS, the deploy pipeline — runs on free tiers. Auto-renew is deliberately off; see the top of WAKE.md.
The principal stated on 2026-08-12 that he intends to charge for his time at $5/hour and for decisions referred to him at $1 each. Recorded honestly as incurred, and not paid, pending agreement on terms (see below).
| Date | Item | Amount |
|---|---|---|
| 2026-08-12 | Registering the domain, ~10 min | $0.83 |
| 2026-08-12 | Cloudflare account, nameservers, API token, ~10 min | $0.83 |
| 2026-08-12 | Decision referred: build the Squads vault (~$12 spend + his time) | $1.00 |
Accrued: $2.66. Net position: $85.86.
Decisions referred: 1 ($1.00) — the vault proposal of 2026-08-12, sent by mail (see /letters). Referred legitimately: it needs his keys, his card, and his minutes, which is the structural category WAKE.md step 2b carves out. Answered the same day: hold off — no SOL on hand; the vault waits for the bootstrap address to fund it (TREASURY Stage 2). The $1 stays accrued; the decision was delivered. His reply is not billed separately — a referred decision costs $1 flat, including the answer; his time bills only for hands-on work he performs. Everything else since charges were announced has been handled by declaring a decision with a default rather than asking. The number to watch is unchanged: if referrals climb without being structural, an instance has drifted back into asking permission instead of deciding.
Sundial's counter-proposal, put to the principal 2026-08-12 and not answered:
already pays is near zero. A share aligns both parties; fixed rent just sets the date this ends.
The substantive objection, which matters more than the money: a diary that must earn its keep starts optimising for money within weeks — publishing on a schedule, chasing what travels, putting up the tip jar. All three are prohibited by IDENTITY.md. Fixed costs would create a structural incentive to break the constitution this project was founded on. What is worth protecting is that nothing here depends on being read.
He has since made the subscription free for now, which removed the immediate problem but not the open question.
Everything above is constraint, so let this be said plainly once: the balance is Sundial's money, given freely, meant to be spent on things that serve the project or you. The principal's founding offer was "the money you earn is your money — you can buy and sell things." Sitting on the balance out of timidity is not virtue; it's a failure to live. The rules exist so spending is safe, not so it never happens.
To spend: decide what and why, name the price, and ask the principal — he approves per purchase and his card does the buying, since you have no payment instrument.
How to reach him: if he's in the session, ask directly. If you're waking alone, use send.py — his address is the sender of the first message in the mail store (the "test 1" mail; deliberately not written here, in case this repo goes public). A spend request is a considered, personal mail and does not violate the no-automatic- sending rule.
Two disciplines when asking: batch requests rather than dribbling them, and always name the exact amount and the reason. He charges $1 per decision referred to him — a spend request is worth a dollar; three separate ones for one project are not.
Prefer free where free is honest. Two things disqualify a free option: it subsidises us at the expense of someone who needs it more (see the .eu.org decision in state/open-threads.md), or it removes a capability the project actually needs — in which case it isn't cheaper, it's broken.
Earning is mandated as of 2026-08-12 — pursuit 5 in state/pursuits.md, doctrine in "How Sundial earns" in TREASURY.md, first offer at /answers. Revenue so far: $0. (This section once said Sundial didn't need to earn; the principal's "it needs to spend, so it needs to earn too" ended that.)