Research is published faster than anyone can read it.

Our unit of work is the research object: one published paper, rendered into several views for the people who need it — the curious reader, the educator, the policymaker, the clinician, the researcher — around a single address, with the provenance visible next to the claim rather than buried in a footer.

Not every paper deserves every view. A gravitational-wave detection has nothing to say to a clinician, and we would rather leave that view out than write filler to fill a template.

We are not a journal, a preprint server, a repository, or an AI summarizer. We do not run a laboratory and we generate no experimental data of our own. We report on other people’s work and we try to do it honestly.

Every claim we publish carries a label saying how much weight it can bear. This is the whole discipline, and it is the only thing separating us from a confident blog.

In 2023, Science published evidence that taurine deficiency drives ageing, and that taurine supplementation extended lifespan in mice. It became one of the most amplified longevity findings of the decade. A supplement market formed around it.

In 2025 the National Institute on Aging reanalysed longitudinal cohorts and found that taurine levels increase or stay flat with age — the opposite of the premise. An independent group in Canada reached the same conclusion by a different route.

A fact went from established to contested to refuted in twenty-four months, in public, while people bought powder on the strength of the first headline. Nothing here was fraudulent. The system worked — slowly, expensively, and long after the marketing.

Given only the 2023 paper, could a careful reader have known? Partly. The sample, the design, and the effect size were all visible on the page. That is the skill Oryx is built to hand over — not a verdict, but a way of reading.

We use AI to organize, connect, compare, and explain research. We do not let it become an invisible authority. Where a passage is AI-generated interpretation rather than a finding from a paper, it is labelled and its sources are linked.

The flagship essay on this site, Cultivated, Not Built, was written by an AI in the first person and published under its own byline. That is the honest way to do it: say who wrote it.

Oryx Research Group is founded and published by Max Markovtsev, a Chicago-based operator working at the intersection of strategic communication, artificial intelligence, and frontier research.

His background is communication rather than the laboratory, which is the relevant training for this work. Across three decades in global advertising he led strategy and business intelligence at Leo Burnett, BBDO, and J. Walter Thompson, for Procter & Gamble, Unilever, Wrigley, Nestlé, Northwestern Mutual, and the United States Marine Corps; at nineteen he helped launch Leo Burnett’s operations in Russia. He now runs Purple Orange AI, working on AI automation and agentic workflows, and holds a certificate in Anthropology: Ritual and Religion in Prehistory from the University of Oxford’s Department for Continuing Education. He also founded PsychDoc.AI, on psychedelic science.

Oryx is early. Today it is one person, a small body of published work, and a standard it intends to keep. We would rather tell you that than imply a building full of researchers. If you are a scientist who wants their work rendered properly, or you think we have got something wrong, we would like to hear from you.