Public summary · PH-SP-001 · Version 0.1 · Effective July 15, 2026
Parallel Human is testing whether an AI-native operator can form, test, and steward a portfolio of ventures, concentrate after evidence appears, and reproduce useful parts of that capability with other founders.
Primary question
Can an AI-fluent founder responsibly create and operate a mixed portfolio long enough to discover and concentrate on stronger opportunities without unacceptable losses in quality, judgment, health, customer trust, or commercial progress?
Can other selected founders reproduce meaningful parts of the pattern?
Can initial capital enter later or in smaller milestone-based amounts in suitable categories?
Can shared infrastructure support several portfolio founders without unsafe concentration?
Registered hypotheses
Suitable ventures can reach equivalent usable-product criteria with materially less cash and specialist labour.
Weakens when comparable builds lose the savings after direction, review, correction, tools, and support are included.Human-dependency load can govern a founder carrying more than one venture better than raw company count.
Weakens when multiple active ventures repeatedly cause missed obligations, unsafe decisions, unresolved failures, or unacceptable strain.Bounded opportunity testing can identify stronger ventures without equal ongoing effort across all ventures.
Weakens when PH preserves weak ventures, cannot decide in time, or portfolio overhead blocks focused progress.In suitable categories, capital can be deployed later against clearer milestones because product formation consumes less cash.
Weakens when compliance, distribution, data, hardware, relationships, or full economic costs restore conventional early capital needs.Selected founders other than Russell can reproduce meaningful parts of the operating pattern.
Weakens when fewer than two independent founders complete the study after selection, support, and infrastructure defects are considered.A viable but deprioritized venture can retain value with a better-matched operator under governed terms.
Weakens when matching fails, context cannot transfer, obligations make transfer unsafe, or transfer cost exceeds likely value.Study phases
May 5 to July 14, 2026. Dated operating records and labelled estimates establish the opening baseline.
Begins July 15, subject to a protected observation store. Founder-week and venture-week evidence replaces retrospective recall.
Two to three founders only after identity, access isolation, consent, agreements, founder workspace, decision history, incident response, and internal dogfooding pass the readiness gate.
Evidence states
Current public boundary
Version 0.1 does not authorize external founder onboarding, applicant data in KoyaOS, fundraising, investment solicitation, superior-return claims, universal founder-capacity claims, or publication of sensitive observations from an unprotected store.
Public reports must name the relevant hypothesis, protocol version, observation window, evidence state, limitations, counterevidence, and correction history.
Follow the evidence