Kasuri

Sponsorship and donations

Fund the experiment that could end the programme.

We are happy to discuss sponsorship and donations. What follows is what the money is for, and — more importantly — what it does not buy.

This is a patient-money programme with a published set of criteria capable of ending it. That shapes what sponsorship can and cannot be, and it is worth stating plainly before any conversation rather than during one.

What the money is for

The single most useful thing anyone could fund is the pre-registered benchmark. It is the instrument that tests whether the central claim survives contact with evidence, and it is the one thing standing between a well-argued position and a measured one.

Benchmark
The minimal kill-capable evaluation Roughly $25,000–27,000. Sealed domains, criteria frozen before any run, compared against the incumbent stack at its best. It is designed to be capable of returning a result that ends the programme. Currently unfunded.
Reference implementation
Interpreter, then a compiler slice The trusted computing base is sized at roughly 15,000–25,000 lines across five enumerable items. Estimated at 2–4 person-years.
Independent verification
A second checker, deliberately diverse Different model family, different implementation language, prompted from the specification only. This is the commitment most likely to be dropped under schedule pressure, which is exactly why funding it separately is useful.
Running cost
A floor of roughly $150,000–250,000 a year Patient money only. The staged gates in Chapter 19 exist because this work cannot be rushed into being correct.

What sponsorship does not buy

These are not courtesies. The programme's entire argument rests on published, falsifiable evidence, and evidence that money can move is not evidence.

  • Not the result. Kill criteria are numeric and frozen at benchmark registration, before any test runs, and are never adjusted afterwards. A sponsor cannot change them, and neither can we.
  • Not the publication decision. Results are published either way, including results that end the programme. There is no arrangement under which an unfavourable finding stays private.
  • Not the comparison. The incumbent arm is the strongest available configuration with current tooling, never a strawman, and it stays that way regardless of who is paying.
  • Not exclusivity. Findings are not licensed, embargoed for one party, or released early to sponsors.
  • Not endorsement in either direction. We will not describe a sponsor's products as secure, and sponsorship does not imply a sponsor endorses ours.
Disclosure is a condition, not an option

Any funder of benchmark work is disclosed alongside the results, with the full logs, under the conflict-of-interest mechanism the evaluation was designed with. An undisclosed sponsor would make the numbers worthless — to you as much as to us.

If a sponsorship cannot be disclosed, it is one we should not take, and saying so now is cheaper than discovering it later.

What sponsors do get

  • Influence on scope, not on outcome. Which workloads the benchmark models, and which sector obligations the specification must satisfy, are genuinely open questions where an operator's input is worth more than our guess.
  • Technical due diligence access. The research corpus, the decision records, the specification and its gap register — artifacts rather than assertions. Chapter 18 describes what exists.
  • Early sight of the questions, not of the answers. You will know what is being measured and why before it is measured.

Where this actually stands

There are no sponsors and no donations to date. There is no product to buy, no certification, no deployment and no customer. Nothing on this page implies otherwise, and Chapter 20 sets out the current stage in full.

Donations are welcome and are treated identically to sponsorship for disclosure purposes. If you would rather fund something specific than contribute generally, the benchmark is the answer — and it is the item most likely to produce a result you can check.

Get in touch

Tell us what you are considering and any constraints that come with it. If disclosure is a problem, say so first — it saves both of us a conversation.

Goes to the same mailbox as general enquiries, flagged as a sponsorship enquiry. No analytics, no third-party scripts, no client-side JavaScript — this form is an ordinary HTML POST. Your message is emailed to us and not stored anywhere else.