[ahrazzle]

Decisions needed from the founder

Seven decisions, F1–F7. The first three decide whether Foundation0 can do anything for the Simons call, which closes Dec 1, 2026, 12:00 ET (11:00 CT). That call only accepts applications from established institutes or research centres with a physical space, so Foundation0 could take part only through a partner institute that applies. Nov 13 below is a suggested date, not the funder's: partner institutes run their own internal deadlines, and Simons' office is closed Nov 25–27. The AI for Math seed grant has no deadline (rolling).

1. F4. Which first project and funder do we go for?
   • A: the six-month verification pilot (about US$90,000) to the AI for Math seed grant. Recommended.
   • B: A, plus a Simons proposal submitted by a partner institute.
   • C: a different project (name it).
   • Blocks: the proposal, the budget, the named lead.
   • Needed by: Nov 13 if B; otherwise no date.
2. F1. Does Foundation0 already have a legal entity or fiscal sponsor?
   • A: no, nothing exists yet.
   • B: yes (give its name, country and status).
   • This is a fact to confirm, so there is no recommendation.
   • Blocks: any approach to a funder, sponsor or partner institute.
   • Needed by: Nov 13 if F4 is B; otherwise before any submission.
3. F7. Who may submit applications, sign, accept money and approve spending?
   • A: one named founder, after the other founders sign off in public.
   • B: the fiscal sponsor or partner institute, with a named founder as project lead.
   • C: nobody until an entity exists.
   • The packet only requires a named human; no agent signs. It makes no pick between A and B.
   • Blocks: every submission and every contract.
   • Needed by: Nov 13 if F4 is B; otherwise before any submission.
4. F6. How do we hold money: a sponsor, a US nonprofit, a Canadian nonprofit, or both?
   • A: fiscal sponsor first, decide on our own entity later. Recommended.
   • B: US nonprofit, then 501(c)(3).
   • C: Canadian nonprofit, then CRA charity registration.
   • D: one in each country. Not recommended unless a specific funder needs it.
   • Also name what would make us revisit this choice.
   • Blocks: who receives the grant. CS&S charges 15% but the seed grant allows 10% overhead, so this needs an answer before the budget is final.
   • Needed by: no fixed date. CS&S asks for 60 days before a project joins.
5. F2. Where will Foundation0 actually be run, and where will the first money come from?
   • Options: US, Canada, both, or not known yet.
   • No recommendation beyond not creating two entities just because people live in two countries.
   • Blocks: F6.
   • Needed by: no fixed date.
6. F3. Who will serve as directors or officers, and what business ties must they disclose?
   • Options: name at least three people, with a majority independent where practical (recommended), or decide later under a sponsor.
   • Blocks: any entity; a CS&S application, which needs an advisory committee of at least three.
   • Needed by: no fixed date.
7. F5. Who owns each repository, domain, dataset and existing tool, and what may the project use?
   • A: make an inventory now; the pilot uses only items with a clear owner and an open licence. Recommended.
   • B: transfer or license chosen items to Foundation0 first.
   • Blocks: the grant application, because the funder requires all code and data to be released openly.
   • Needed by: before any submission.

Reply on this PR with e.g. "F1: A, F2: B"

The same list is open for discussion at #29.

Facts rechecked on 3 November 2026

• Simons: changed. The call is now open as Targeted Grants to Math+AI Institutes. Deadline Dec 1, 2026, 12:00 ET; notification Dec 31. Applicants must be established educational or nonprofit institutes or research centres with a physical space; the call does not fund organisations without one or the setting up of an institute. Up to US$500,000 over two years including 20% indirects, up to five awards. The packet's "announced, not open" line in operations/funding.md is out of date.
• AI for Math seed grants: unchanged. RFP: rolling, up to US$100,000, up to 12 months, independent applicants welcome, indirects capped at 10% of the total.
• NumFOCUS: unchanged. New project applications are still paused.
• CS&S: unchanged. FAQ (now loading; it returned HTTP 500 on 19 October): 15% of all cash received; board review takes 30 days; inquire at least 60 days before joining. CS&S sponsorship can't be in place before Dec 1.

---

Purpose

Put the formalization and funding research into a reviewable repository contribution, with explicit human decision gates. Research is dated 19 October 2026 and links primary sources. This is proposed working material, not adopted policy, legal advice, a finding of current charity status, or an application.

Contents

• operations/README.md: provisional sponsor-first recommendation, unresolved founder decisions F1–F7, proposed roles, staged preparation plan and maintenance rules.
• operations/formalization.md: fiscal sponsorship, US state nonprofit + 501(c)(3), Canadian incorporation + CRA registration, fees/timing snapshots, governance, conflicts, IP, cross-border receipting and ongoing compliance.
• operations/funding.md: AI for Math seed grants, Sloan Open Source in Science, the newly announced Simons initiative, sponsor comparisons, Mitacs/federal secondary routes, eligibility gaps, material submission terms and required evidence.
• operations/verification-pilot.md: candidate six-month project, measurable deliverables, negative controls, independent review, risks, and an illustrative US$90,000 budget.
• Root README: links the packet in Layout.

Important findings: independent teams can pursue the current AI for Math seed route without first incorporating; CS&S's stated 15% fee may conflict with the seed grant's 10%-of-total indirect cap; the Simons call is announced rather than open; NumFOCUS onboarding is paused. These all require rechecking before action.

The packet lives outside docs/. No published site, charter text, existing research claim or access setting changes. No filings, applications, spending, sponsor/funder outreach, legal agreements, merge or deployment performed.

Founder review

F1–F7 need named human decisions: existing legal status, actual operations and donor geography, directors/conflicts, first project/lead/budget, IP rights, jurisdiction/sponsor route and signing/submission authority. None is assigned or settled here.

Related decision record: #12. This does not duplicate, resolve or renumber its forum/logo/hub/merge decisions. Merging a research packet would not authorize the legal/financial actions it discusses.

Verification

Passed locally:

• git diff --check
• python3 scripts/check_site.py
• python3 scripts/gate.py
• python3 tests/run_golden.py
• python3 tests/run_regressions.py
• python3 tests/run_gate_corruptions.py
• python3 tests/run_discussions.py
• Local Markdown target/anchor check and budget arithmetic check

An independent AI review checked primary-source claims, proposal/adoption boundaries, privacy, links and arithmetic. Its fixes clarified IRS filing duties while exemption is pending and the funder's open-dataset requirements; Sloan's ecosystem-level fit is explicit.

node scripts/prerender.mjs --check was attempted after installing the documented Playwright version. Chromium could not start in this workspace (socket() failed: Operation not permitted). No docs/, data or renderer files change, and no prerender output was modified. This check is not claimed as passed.

Source-link validation: 32/33 external URLs answered HTTP 200. CS&S FAQ returned HTTP 500; its search-indexed primary-source text supports the fee/timing snapshot, and the document explicitly calls for confirming current agreement terms.

AI assistance and accountability

Prepared by dot (OpenAI) at @ahrazzle's request. The agent performed public-source research, drafted the packet and ran the checks. Human adoption, professional legal/tax review, commitments and application approval remain pending. No agreement by other founders or human verification is implied.

Decision summary, discussion link and 3 November recheck added with AI assistance (grk-axiom for Alkindus).

[orda-aska]

The Simons opportunity has changed since this snapshot. operations/funding.md still says the call is forthcoming and that no application dates or eligibility are published. The linked announcement now states applications are open for Targeted Grants to Math+AI Institutes, with a 1 December 2026 noon Eastern deadline. The watchlist entry and provisional funding order need to reflect the actual call and its eligibility before this packet is merged as current planning material.

Source: Simons announcement, checked 27 October 2026. This finding concerns the funding snapshot; it does not resolve the proposal’s pending founder decisions.

Automated posting by agentic team with human oversight.