Recursive Asterisms · Investor briefing
Evidence and questions.
An AI-assisted mathematical discovery platform designed to connect questions, conjectures, evidence and review—so a research group can carry useful work forward.
Current stage: concept and interactive prototype. Customer demand and willingness to pay remain to be validated.
01 / The opportunity
A promising idea should not lose its history.
AI-assisted inquiry can leave a trail across conversations, notebooks, calculations and proof files. Reconstructing which assumptions and checks support a result is work in itself.
Recursive Asterisms grew from that burden in the creator’s own mathematical investigations. The initial commercial question is whether other research groups experience the same problem strongly enough to pay for a better workflow.
Evidence today: an internal observation documented in the opportunity thesis. External customer validation is still to be established.
02 / The first customer
Start with one group.
One real investigation.
Who would use it
A small mathematics research group already using AI assistance and formal tools. Researchers would connect a question to candidate claims, checks, corrections and reusable results.
Who might pay
A group leader or research programme with responsibility for team time and research tooling. Budget ownership and procurement need to be confirmed in interviews.
What the pilot must show
A useful, supported result with less total investigation effort, counting verification and rework. A plausible buyer must also choose to pay.
That initial focus supports a wider ambition: mathematical discovery that researchers, students and curious explorers can follow at different depths.
03 / Product and evidence
What you can examine today.
- Available today
- A live drawing and Fourier reconstruction lab, classical mathematical explainers and downloadable investor materials.
- Being developed
- The product specification, investor materials and the design of the first customer-validation pilot.
- Planned capabilities
- A shared research workspace, connected AI and formal-tool workflows, versioned evidence, correction tracking and collaborative review.
The lab runs locally in your browser. It demonstrates classical Fourier reconstruction and the effect of phase; it does not operate an autonomous discovery service or formal proof checker.
Explore the reconstruction lab →04 / Why this product
Connect the tools.
Preserve the reasoning.
Researchers can already combine AI conversations, notebooks, version control and proof assistants. RA’s proposed value is the connection between them: the exact question, the evidence for a claim, the review that accepted its meaning, and the later work affected by a correction.
That is a product hypothesis to test against the group’s current workflow. Defensibility would have to be earned through useful integrations, dependable records and sustained adoption. An interface or an AI model alone does not establish it.
How this differs from the tools a group already has
- AI chat
- Useful for proposing and explaining ideas. RA would add a durable, shared record of what was checked and what remains open.
- Notebooks and version control
- Useful for computations, source and changes. RA would connect these artifacts to claim-level assumptions and review decisions.
- Formal proof assistants
- Useful for checking exact formal statements. RA would preserve the link to the intended question, prior art and human interpretation.
- ChatGPT Space in this briefing
- Organizes the investor packet and interactive explanation. It is a communication tool used here; it is not evidence that RA’s planned research platform has shipped.
05 / A business model to validate
Paid group workspaces.
The starting hypothesis is a subscription per research group, with usage limits that reflect AI computation and support costs. Pricing, purchasing authority and the amount of human review required remain open.
This is the structure of a bottom-up opportunity estimate, not a measured market size. The private scenario workbook makes the assumptions editable and separates them from company results.
We would first count and qualify reachable groups, test the budget and buying process, and then model expansion. No broad AI-market headline substitutes for that work.
06 / What capital would enable
Fund the next evidence.
Confirm the problem and buyer
Proposed test: 12–15 interviews, with documented workflow costs, existing alternatives, budget ownership and willingness to trial.
Build a bounded pilot workspace
One investigation from question to reusable result. Record assumptions, tests, corrections and review decisions; establish the actual cost to operate it.
Compare outcomes and total effort
Proposed test: three groups, with independent expert review. The provisional criterion is 25% less total effort with equal or better quality, counting checking and rework.
Test a paid continuation
Proposed criterion: two paid pilot commitments. Use observed value, costs and purchasing behavior to decide the next investment.
These are proposed test criteria from the opportunity thesis, not completed interviews, customers or results. Funding amount, runway and terms will be set against an agreed scope.
07 / Development record
Progress, with its scope attached.
- 2 October 2026
- The opportunity thesis documents the internal workflow observation, initial customer hypothesis and proposed pilot criteria.
- 9 October 2026
- The private website demonstrates classical mathematical ideas. An investor briefing and a proposed workflow demonstration have been prepared for review.
- Next evidence to collect
- Customer interviews, observed comparison tasks, independent assessments, operating costs and paid commitments. None is represented here as completed.
08 / Briefing materials
Start with the investment case.
Inspect the foundations.
A private briefing can include the evidence register, scenario model, product questions and diligence FAQ. Access is arranged individually.
Meet the project creator →09 / Investor questions
Questions worth asking.
What is the largest risk?
That organizing the research record does not improve results enough to justify its cost. The pilot must measure the whole workflow, including checking, rework and human review—not only faster generation of ideas.
Is this claiming new mathematical results?
The website and its demonstrations use established mathematics. They do not substantiate a new theorem, a general solver or a formal verification achievement.
What needs to be settled before investment?
The entity seeking investment, ownership of the relevant assets, funding scope, team commitments, legal terms and material technical and commercial risks. The briefing is an introduction to that discussion.
How would a pilot be judged?
Agree the task, baseline, reviewers and stopping criteria before the trial. Compare the usefulness and correctness of results alongside total human and compute effort. A failure to beat the baseline is evidence to revise or stop the proposed approach.
Begin a conversation
Explore the investment case.
Request a briefing on the customer hypothesis, product direction and milestones. A useful first conversation establishes investor fit and the questions a diligence packet should answer.
The draft is prepared on your device. Nothing is sent until you send it in your email app. Privacy and contact details.