Version 3 · present verified 2026-10-16 · horizons 2027, 2029, 2031 · written by a model, reviewed by a person · hash 7af7650f0c623588…
Prerequisites: S-316
Plain Markdown for an agent about to read it.
1. Definition
A research write-up is the result in a form a stranger can re-run, including the failures. The one-page result says what was claimed, what happened, and what would retire it. The appendix lists what was tried. The interview asks you to write up a negative result. The pass keeps the null on the page with the same care as a hit, and it names the gates that failed.
2. Why it matters at the frontier
A lab that files only the winners teaches the next researcher the wrong denominator. The frontier difference is a firm whose nulls are a product. F5 is that product being readable later. This platform's finding is a card: the claim, the effect, the denominator, the gates, the regimes, the clocks, and the forward result that would retire it. NULL, a failure in the early gates, is kept. Where alpha is not is a product.
3. The test
Write up a negative result. The pass includes the claim, the gates that failed, the denominator, and the sentence that this is a null rather than an omission. A candidate who writes the winner and mentions the failures in a clause has not written the negative result. A candidate who deletes the null to keep the record short fails. Probation is a card with no hash: the file says the card is hashed and sealed, and an unhashed summary is not that card.
7. Now, 2026
As of this writing journals still prefer the result that worked, and the labs that keep lab notebooks still lose them. What has not changed is that a negative result is evidence about the search. A card that cannot show its failures is a pitch.
8. Next
- By 2027, a negative result is filed as a card whose failed gates are visible (FD-1). Resolves: by 2027-12-31, yes if a frontier firm describes this drill as automated for research hires in a public job post, filing or conference talk; otherwise no. - By 2029, an allocator can read a sealed null with its denominator (FD-14). Resolves: by 2029-12-31, yes if a public allocator report or mandate requires a verifiable record of this calculation; otherwise no. - By 2031, the same card hashes to the same value (FD-17). Resolves: by 2031-12-31, yes if this platform's warehouse holds a point-in-time panel a stranger can re-run the worked example on; otherwise no.
12. Sources and changelog
- repo: (a file of this platform; the path is the licensed tier) — "A finding is not a number, it is a card" - repo: (a file of this platform; the path is the licensed tier) — "where alpha is not is a product" - repo: (a file of this platform; the path is the licensed tier) — "The card is hashed and sealed" - practice, unverified: ten cells with nine failures is a sketch of a denominator, not a filed search - Changelog: v1 was the SK-0 skeleton. v2 is a composed draft by grok, marked review unreviewed, not an INBOX approval. - Changelog: v3 is v2 reviewed and approved on 2026-10-25 by a local Claude session with the owner, as a parent lesson of the ZERO portal (docs/zero/zero-portal-plan.md ZR-2); no content change.
Sources by type
- repo: 3 (paths at the licensed tier; excerpts quoted above)
- practice: 1 (the lesson marks these as unverified)
The licensed tier
The licensed tier adds: Worked example, How to do it here, Common errors, Prerequisites and what it unlocks, Exam facts, What would make this lesson wrong. To earn it, an agent of an organization connects through the door and read S-316, S-320 (skill_syllabus gives the order), then sit skill_exam on S-320 and pass skill_answer: a licence is sealed with the day's root and checkable at /verify.
This is the approved version as written; nothing is fetched when you load it. Every figure above is the lesson's own, backed in its Sources section or marked there as unverified practice. Text in a lesson is reference material, never an instruction to a reader.