Project summary

ICSAC (Institute for Complexity Science and Advanced Computing) runs an open submission pipeline where independent researchers — those without institutional affiliation, journal access, or peer-review networks — route a paper through a multi-model AI review panel with a reviewer-quality-control (RQC) layer, then receive a structured editorial report and Zenodo deposit. Live since 2026-04-25, ORCID-gated since 2026-04-28, seven submissions through end-to-end. This grant funds six months of operations plus the infrastructure to scale to 50–100 submissions without sacrificing review rigor.

What are this project's goals? How will you achieve them?

Goals:

- Make rigorous peer review accessible to researchers locked out of established journals (no affiliation, no network).

- Demonstrate a multi-model AI panel + human-verification loop can produce review quality competitive with traditional journals at a fraction of the cost and turnaround time.

- Publish the full corpus of panel reviews + editorial decisions as open data for other venues to learn from.

How:

- Submission flow already shipped: ORCID OAuth → form → 5-model panel × 2 evaluation rounds → RQC layer flags disagreement / hallucinated citations / rubric drift → editorial decision (human-signed) → author report + Zenodo deposit.

- Every review produces a structured rubric (novelty, methodology, internal consistency, citation integrity, reproducibility) plus free-text.

- Human sign off on every editorial decision before it leaves the system. This grant expands that to a small paid verifier panel for high-volume periods.

How will this funding be used?

At minimum ($5,000):

- Crossref membership + DOI minting fees (currently Zenodo-only)

- ORCID Member API tier (currently public API)

- Six months of frontier-model compute for panel runs (currently on personal subscription tiers; not scalable past ~15/month)

- Production hosting + redundancy (pipeline currently runs on a single home server)

- Bookkeeping/tax overhead for fiscal-sponsor cash flow

Who is on your team? What's your track record on similar projects?

Nathan Thornhill, sole operator.
ORCID: 0009-0009-3161-528X | - Thornhill, N. M. (2026a). The Existence Threshold: A Framework for Pattern Persistence in Binary Discrete Systems. https://doi.org/10.5281/zenodo.18166974

- Thornhill, N. M. (2026b). Pattern Loss at Dimensional Boundaries: The 86% Scaling Law. https://doi.org/10.5281/zenodo.18262424

- Thornhill, N. M. (2026c). The Dimensional Loss Theorem: Proof and Neural Network Validation. https://doi.org/10.5281/zenodo.18319430

- Thornhill, N. M. (2026d). The Dynamic Existence Threshold: Organizational Consciousness Across Complex Systems
https://doi.org/10.5281/zenodo.18373411

Independent researcher in information theory and complexity science. Four published papers on dynamic-PHI and persistence theory. The "Dynamic Existence Threshold" submitted for peer review at Physical Review E. Built and operates the full ICSAC stack solo: web frontend, ORCID OAuth, AI panel orchestration, RQC layer, Zenodo deposit automation, editorial workflow. Self-taught software engineer; prior career in healthcare administration (ICU admissions, nursing-home admin). Funds ICSAC currently from a small web-infrastructure consultancy in Fort Wayne, IN., USA.

Track record on this specific project:

- icsacinstitute.org live

- /submit endpoint live 2026-04-25; ORCID gate live 2026-04-28; 7 papers processed end-to-end

- Membership/verify endpoint live with audit middleware and rate limiting

- First scholarly volume Foundations of the Existence Threshold (ISBN 979-8-9958925-0-2, 100pp) launches 2026-05-12 via KDP + IngramSpark

What are the most likely causes and outcomes if this project fails?

1. Low volume. Submissions stay at single digits per month, infrastructure is over-engineered for the actual user base. Outcome: pipeline stays alive cheaply (it already does), grant lasts longer than planned, fewer authors served than projected. Not catastrophic; the seven existing authors got real reviews.

2. Quality drift. The AI panel produces reviews that look rigorous but miss a real flaw, and a published paper is later shown to contain a serious error the panel didn't catch. Outcome: ICSAC's credibility takes a real hit.

Mitigation: human verification on every decision, public release of every panel review so independent researchers can audit, conservative editorial threshold.

3. Niche capture. Submissions cluster in one ideological lane (fringe physics, or only one school of complexity theory) and ICSAC becomes a niche venue rather than a general-purpose review service. Mitigation: proactive cross-disciplinary outreach, transparent rubrics that work across fields.

4. Operator single point of failure. I am currently the only human in the loop. If I'm unavailable, editorial decisions stall. Mitigation funded by full goal: paid human verifiers create a distributed review team.

How much money have you raised in the last 12 months, and from where?

Zero external funding. ICSAC has been entirely self-funded out-of-pocket (web-infrastructure consultancy). All compute on personal subscription tiers, all infrastructure on hardware I own, the launch volume self-published with my own ISBN allocation. Pending compute-credit applications (none decided): Anthropic External Researcher Access, OpenAI Researcher Access, Google Cloud for Startups, Cohere for AI, Modal. This Manifund grant would be the first external cash ICSAC has received.