Partners & Funding

Fund work that ships.

Dissensus turns a formal thesis about friction in multi-agent systems into artefacts you can check: theorems a proof assistant has verified, a risk index that runs every day, experiments that rerun from seed. Funding does not buy a promise of research — it accelerates a programme already producing.

Or simply: the work exists; support makes more of it.

01 · The case

Research you can verify, not just read

Most research asks funders to trust the process. Ours is checkable. Twenty papers across computational finance, political economy, AI alignment, and formal methods — organised around a single thesis: coordination has overhead, delegation produces friction, and friction decomposes into stake, entropy, and alignment. The thesis is not a slogan; it is a functional with machine-checked properties and empirical tests in three domains so far.

Machine-checked proofs
Core results of the friction framework are formalised in Lean 4 and verified by machine — no placeholder proofs, no silent axioms. The compiler is the referee.
Σ
A live index
ASRI, our systemic-risk index for cryptocurrency markets, runs continuously at asri.dissensus.ai — validated against the 2022–23 crisis history, not a backtest in a drawer.
Reproducible experiments
Every reported number traces to committed code with fixed seeds. Datasets and pipelines are archived with DOIs, so anyone can rerun what we publish.
02 · What funding unlocks

Three bottlenecks, all of them tractable

The programme is not blocked on ideas. It is blocked on compute, hands, and the unglamorous work of carrying results from working prototype to published, replicable finding. Each is a problem money straightforwardly solves.

Compute

Multi-agent and privacy experiments at proper scale

The multi-agent friction experiments and the CBDC-privacy work currently run on a dual-GPU workstation. Cluster time turns week-long parameter sweeps into overnight ones, and lets us run the factorial designs at the resolution the theory actually calls for.

People

Research-assistant time through to publication

Empirical programmes die in the gap between promising result and peer-reviewed paper. Funded research-assistant time carries the studies through that gap — data collection, robustness batteries, replication — instead of leaving them queued behind everything else.

Open artefacts

Public goods, by default

What partnership funds here stays public: archived datasets, the Lean formalisation library, and the live ASRI index are open-access and citable. Supporting the programme produces infrastructure other researchers build on, not a private asset.

03 · Ways to partner

Pick the door that fits

01
Grants & philanthropy
Fund the programme, at project or programme level. Outputs are open by default, and we report what shipped against milestones agreed up front.
02
Institutional & academic
Joint proposals, co-authorship, shared infrastructure, visiting arrangements. ASCRI is the programme umbrella for cross-institutional work.
03
Industry engagement
If the problem is commercial — a bespoke risk index, formal verification, adversarial evaluation — that is the Services side of the house.
04 · Track record

The record so far

Peer-review venues include Digital Finance, AI and Ethics, and Ethics and Information Technology. The full archive — every paper with its identifiers — is under Research; the programme itself is documented at systems.ac.

0
Publications
0
Research domains
0
Peer-review venues
0+
Machine-checked proofs
05 · Contact

Start a conversation

One email is enough. Tell us what you want to accelerate — or ask what we would do with the support; the honest answer is in the brief.

research@dissensus.ai

What we can talk about

  • Grant and philanthropic funding, at project or programme level
  • Academic collaboration and joint proposals under ASCRI
  • Commercial engagement via Services
We reply to substantive messages, and if the fit is wrong we will say so quickly — friction analysis begins at home.