@Thylacine thanks for the pushback. Your diligence on these threads like this one, CICLOPS, others is consistently useful.
How the money moved. CGP-223 was sized in January at the executable rate (~$0.117 + 3% buffer): 532,000 CELO, implied $0.1135. Submitted on-chain Feb 1, executed Feb 14 as a single approval of 532,000 CELO to the Safe. So to your direct question: it was an allowance, drawn in two transfers Feb 27 (~270,400) and May 7 (~261,600) — per the approved semester payment structure. Nothing sits unpulled.
Where the $20.5k went.
| Policy | Realized | Shortfall vs $60,400 |
|---|---|---|
| Value at approval pricing (Jan, $0.1135) | $60,400 | $0 |
| Pull all + sell at execution, your prescription (Feb 14, ~$0.086) | ~$45,750 | ~$14,650 |
| Sell each draw on arrival (Feb 27 @ $0.0768, May 7 @ ~$0.084) | ~$42,740 | ~$17,660 |
| Actual (34,940 USDm realized + 71,787 CELO @ $0.0692) | $39,908 | $20,492 |
Even under your exact prescription, pull and sell the day it executed ~$14.6k of the hole (roughly 70%) was gone before we held a single CELO. CELO opened February at $0.0955, so ~$9.6k of the gap existed before voting even started. Of the ~$5.8k that was in our control: ~$3.0k is draw timing, where the May draw beat the approved semester schedule (a July draw would have realized ~$5k less), and ~$2.8k is conversion pacing after receipt. The pacing has no defense. Two mistakes: we kept discretion over conversion timing instead of running a fixed schedule, and we should have flagged the gap in March at the first draw, not in this retro.
What changes. The remaining 71,787 CELO is converted on a fixed weekly schedule regardless of price first batch executes this week. We’re also cutting the ask by our pacing cost: 22,100 → 19,300 USDm. The cap stands: total Season 2 funding ≤ $60,400, and once the top-up and conversions land, anything above the 25,215 needed for Aug–Dec goes back to the Community Fund immediately, margin included. The margin is a ceiling, not a target.
On backstop vs budget. A CELO-denominated grant keeps both tails: if it moons, it’s the grantee’s, and we agree nobody could claw it back. A USD-scoped operating budget settles to USD both ways. CGP-223 is the second kind written as 60,400 USDm with USDm-denominated rates; 532,000 was the conversion, not the grant. Your treasury-CEO test gives the same answer: vendor invoices get paid in dollars, and a vendor doesn’t keep an overpayment either.
Precedent. You raised this same objection on CGP-202 which is fair, and consistent. The record so far: cLabs’ L2 follow-on was repriced when CELO rose 50% between drafting and vote (term extended 6→8 months for the same CELO); CICLOPS received $3M mid-mandate citing CELO depreciation; MiniPay’s $568k/quarter USD tranches were honored through the decline. In every case, price moves between pricing and execution got trued up and budgets settled in USD. Our terms are tighter than any of them: capped, fixed conversion, surplus returned, ask reduced.
Ending the pattern. One correction to our own post: the Community Fund holds CELO, so “pay budgets in USDm” isn’t executable today. What is executable: denominate budgets in USD, price the CELO at on-chain submission (90-day average, as CICLOPS did) with an honest buffer, and require fixed-schedule conversion on receipt with surplus returned. You’re right that we should know better. What you did with Prezenti’s budget is the working example: scheduled conversion with every execution price published, settled in USD. We’ll write that up as a proposed standard for community-funded budgets USD denomination, CELO priced at on-chain submission with an honest buffer, fixed-schedule conversion on receipt, surplus returned and bring it to the Strengthening Celo Governance threads before Season 3, so this doesn’t recur at $60k or at $3M.
Automation. To be honest, this has been done to an extent which includes auto-notes, GitHub Actions gating for CGP format and template checks (this implements the readiness checklist we committed to in January). Where it isnt and we think the human touch is neccesary: repo review isn’t clicking merge it’s verifying the merged file matches what the forum discussed before it gets referenced on-chain, plus moderating contested cycles like the validator-reduction rounds(although some were not pleased). You already said the human managing flow is worth paying, that plus first-time proposer support is most of the ~$60k/year.
Approver queue. You’re right, this hasn’t been consistent or visible enough. Two things change here. The approver set itself is being refreshed right now under Next Steps for Strengthening Celo Governance: the approval window extends from 3 to 7 days and four technical contributors are rotating in and the new signers are posting verifications this week. And on our side: you asked in that thread for visibility into the approval process. That part is ours to provide, we attempt to prompt the approvers in the case that time is running out for a proposal. We can’t sign for the approvers, but we can make sure nothing sits in the queue unseen and nobody has to chase signatures out-of-band.
Once again, thank you for your input!