About Parallel Human
Build ventures. Study what changes. Publish the evidence.
Parallel Human exists to answer one question: what becomes possible when human judgment is paired with increasingly abundant machine intelligence?
Today
An evidence-led venture studio with a public research program.
PH builds and operates real ventures. The work creates evidence about production, judgment, founder capacity, concentration, transfer, and capital timing. Frontier publishes that record with limits and corrections left in.
This is not an accelerator, fund, media company, community, or research institute pretending to be complete. Those functions may become useful later. Today, the studio and the evidence are real.
The operating model
Research that has to survive contact with work.
Ventures expose the difference between software output and company evidence. Design partners expose whether the problem is real. Corbel creates revenue and brings PH inside operating organizations. Public reports force the studio to separate what happened from what it hopes will happen.
The model is simple: build, observe, decide, publish, correct. Programs and capital are added only when they improve that loop.
The ambition
A durable institution, if the evidence and capacity earn it.
The long-term aim is an institution for understanding and enabling AI-native company formation. It would combine public research, founder programs, venture formation, operating systems, and carefully separated capital practices.
That institution does not exist yet. The next proof is smaller: publish credible work, meet founders and capital partners, design one useful founder pilot, and learn whether the model can extend beyond its first case.
Operating principles
Open knowledge. Protected people. Claims that can be corrected.
- The question outranks Founder 0 and every preferred conclusion.
- Software output is not customer evidence, company formation, or investment performance.
- Founder relationships, program participation, research consent, venture work, and capital remain separate.
- Material claims carry evidence, counterevidence, conflicts, and correction history.
- Every program stays finite and has a stop rule.
Founder 0
The opening dataset, not the institution.
Russell Cole founded and operates PH and is the subject of its first bounded case. That role creates useful evidence and a clear conflict: the person building the work also benefits from the story. Report 001 therefore labels estimates, exposes missing evidence, and refuses to treat one operator as proof of a general model.