About this publication
Organoid Intellect covers organoid intelligence and biological computing: the attempt to perform useful computation on living neural tissue. This page sets out what we publish, how it is produced, and the standards we hold each piece to.
What we publish
Two kinds of work. Analysis is a daily stream: each piece takes one new item of primary research, a paper, preprint, grant award or registered trial, and works out what it changes for computation on living tissue. The library holds standing explainers that do not go stale, starting with the biocomputing primer.
The distinction matters when you read us. An explainer is meant to be correct for a long time and is revised when the field moves. An analysis is a dated reading of one result at one moment, and we do not quietly rewrite it later.
Where the material comes from
We maintain a research library that ingests continuously from arXiv and bioRxiv, the NIH RePORTER grants database, ClinicalTrials.gov, NSF award records, journal feeds, and vendor technical documentation. Each incoming item is checked for duplicates, summarized, and scored for relevance before it can be written about. Every published analysis is drawn from that library and names the specific record it came from, with a link to the primary source.
We do not write about the same item twice. A ledger records every source consumed by the analysis stream, so a piece that has been covered cannot resurface.
How each piece is produced
Analyses are researched and drafted by an automated research system, a coordinated group of large language models led by a reasoning model working to a fixed editorial brief, and are then validated against publication rules before anything goes live. We think readers are entitled to know that, so we are saying it plainly here rather than burying it.
What that process is required to do: retrieve and read the primary source rather than working from a title or abstract; state explicitly, in the article, when a full text could not be obtained, and bound the claims accordingly; separate what a study demonstrates from what it asserts; identify the load-bearing assumption and stress test it; and give the genuine objection alongside the genuine opportunity. Pieces that fail the publication checks are rejected rather than patched into shape.
What it is not: peer review. An analysis is an interpretation of someone else's research by a reader who did not run the experiment. Where our reading is uncertain, the article says so. Where we think the field is overclaiming, we say that too, but that is an argument we are making, not a finding we are reporting.
Sourcing and citation standards
Every factual claim traces to a named source. References carry a title, venue, date and a resolvable link, and the date we accessed it. We do not cite a study we have not opened, we do not reproduce a number we cannot attribute, and we do not use a secondary write up in place of the underlying work.
We are vendor neutral. Platforms and instruments are named when naming them is informative, and we describe capability from published specifications and results rather than marketing claims. When a piece of research is funded by, or produced by, a party with a commercial interest in the result, the analysis says so.
What we do not do
We do not publish speculation about machine consciousness as though it were a finding. We do not treat a grant award as evidence that its aims have been achieved; a funded intention is a fact about funding, not about the world. We do not report stimulus locked activity in cultured tissue as evidence of computation without saying what control would distinguish the two. And we do not run a piece we cannot stand behind in order to meet a publishing cadence: a thin day is published thin.
Corrections
If a piece misstates a result, misattributes work, or misreads a source, we want to know and we will correct it. Corrections are made on the page itself and noted there rather than applied silently, because an analysis is dated evidence of what we thought at the time. Substantive corrections carry the date they were made. Send corrections, and any question about how a piece was sourced, to mg@brandadvertisers.com. Tell us which piece and what is wrong, and a link to the source if you have one.
Publisher
Organoid Intellect is published by Brand Advertisers LLC. It is one of four related titles covering this field: this publication for concepts and theory, Organoid Grid for platforms and ethics, Organoid Biofoundry for organ models and drug discovery, and Organoid Array for electrode hardware and specifications.