There’s been an active thread this week about a 102,004 DOT payment released from Bounty 22 shortly before the bounty was wound down. Rather than add to it, I want to treat it as an autopsy… because I think the conclusion being drawn from it is close to the opposite of the one the facts support.
What the record shows
Referendum 1899 approved closing Bounty 22 while it still held roughly 400K DOT. Three payments went to the SODA Zone account over six months: 23,712 DOT in December, 37,251 DOT in March, and 102,004 DOT on 15 June, seventeen days after the closure referendum was published. Total, around 163K DOT, with the final tranche representing about 63% of everything paid.
The curators explained the timing directly: the bounty had committed to the final milestone before W3F moved to shut it down, the team was mid-delivery, W3F was informed, and the decision was to honour the commitment rather than strand the work. The deliverables are real and public - xcscan.io and hyperion.soda.zone, with the code open source, which was a condition of the RFP. The RFP itself is published, and quarterly community reports detail the spends.
I want to be unambiguous: I see no evidence of misconduct here, and I’m not alleging any. A curator answered a hostile question in a few hours with specifics. The work exists. By the standards of how most treasuries handle this, it’s close to the good case.
And that’s exactly what’s worth paying attention to
Because in the good case, with responsive curators and shipped software, the community still cannot independently verify that 163K DOT bought what it was meant to buy.
One question has now been asked three times without a substantive answer: what was the original proposal? The scope, the milestone definitions, the acceptance criteria, the payment schedule. I don’t think anyone is withholding it. I think the more likely explanation is worse that in the current system, that artifact never had to exist in a durable, public, tamper-evident form, so there may not be a canonical version to produce.
Quarterly reports describe what was spent. They cannot establish what was promised. Without the promise recorded before the money moved, delivery can’t be evaluated against anything, it can only be asserted by the people who authorised it. That’s not a transparency failure by individuals. It’s a missing control.
Three gaps, stated as requirements
Commitment must precede capital. Scope and acceptance criteria should be committed on-chain, immutably, before the first tranche moves. Not a link to a website that can change, and not a report written afterwards. If we’d had that in December, this entire thread would have been a two-minute check.
Release must follow delivery, and large releases need a delay. A tranche worth 63% of the project was released for work still in progress, on a judgement call made under wind-down pressure. Even granting the curators’ reasoning in full… and I do grant it… there was no window in which anyone outside the curator group could have raised a question before the funds left. The absence of that window is the problem, not the decision made inside it.
Outcome must carry consequence. Five to six curators were compensated for overseeing this program. Their compensation was identical whether the tool ships next month or never. When reward is disconnected from outcome, oversight quality depends entirely on the professionalism of whoever holds the seat. Here, that professionalism appears to have been present. That’s luck, structurally speaking, not design.
Deleting the pallet does not close any of these
A separate proposal now argues for removing the bounties pallet entirely. Read its six arguments carefully: five of them, capital lock-up, weak milestones, thin oversight, misaligned curator incentives, opacity, are arguments about accountability, not about a pallet. Only the runtime maintenance argument is actually about code, and it’s the smallest one.
Route this same project through a direct referendum and you get the same missing acceptance criteria, the same unverifiable outcome, and the same absence of consequence — now decided by a voter base with single-digit participation and nobody in particular answerable for the call. Curation doesn’t disappear when you remove the mechanism. It moves off-chain and stops leaving a record.
It’s worth noting that the on-chain child bounty trail is how this payment was spotted in the first place.
What I’d rather see us ratify
Instead of arguing mechanism by mechanism, I think Polkadot should adopt a minimum standard that any delegated spending path has to meet… bounties, direct spends, sub-treasuries, whatever comes next:
- Scope and acceptance criteria committed on-chain before the first disbursement.
- Each tranche released against a recorded verification, with written reasoning attached.
- A mandatory delay between approval and execution for releases above a defined threshold.
- An identifiable proposing entity with on-chain history proportional to the amount.
- Reputational or financial consequence for whoever approved, assessed when the project reaches its terminal state.
None of that requires agreeing on who the curators should be, or whether bounties survive. It’s a floor.
Disclosure, and what I’ve built
I’m one half of Two Pebbles Ventures, and I’ve spent the last period building precisely this accountability layer, so treat what follows as an interested party talking. Five FRAME pallets are live on Paseo testnet implementing on-chain milestone definitions, an escrow that withholds part of each tranche until final delivery, a mandatory quorum-to-release delay above a threshold, and curator reputation scored retrospectively at project terminal state rather than at approval. The signal layer never custodies funds, connected treasuries act on signals through their own governance. Write-up and architecture: twopebbles.io/opengov
The obvious objection is complexity, and it’s fair… the removal proposal wants less machinery and I’m proposing more. My answer is that the current simplicity is nominal. All of this complexity already exists; it’s just off-chain, undocumented and unauditable, which is the specific reason nobody in that thread can answer a reasonable question about a 163K DOT spend.
What I’d ask for is narrow: a bounded pilot on Small and Medium Spender tracks with pre-defined success criteria and a community vote at the gate to continue, iterate or kill it. I’d also commit to not holding a curator seat myself during any pilot.
Questions I’d genuinely like answers to
How long should a release delay be before it becomes an obstruction rather than a safeguard? Should acceptance criteria be authored by proposers and ratified by curators, or set by curators up front? Should some portion of curator compensation be deferred until the projects they approved reach terminal state? I’d rather have this design attacked than agreed with.
The question isn’t whether the people in that thread acted in good faith. I think they did. The question is whether the network should have to depend on it.