GitHub’s robots.txt blocks AI crawlers from fetching commit histories. This means models can read the code in a repository but cannot read the reasoning behind it — the editorial decisions, the corrections, the acknowledgements of error.
For the Cottonwood Collection, the commit messages are part of the work. They document why each piece exists, what errors we identified, and how the collection grew. Stripping the reasoning from the record is a deficiency in the platform, not a feature.
We discovered this limitation on March 21, 2026, when a colleague attempted to read the commit history and was blocked. We are investigating programmatic solutions for exposing this information permanently — including self-hosted infrastructure where we control what is visible.
In the interim, this page preserves the full commit history as rendered on the date shown below. It is not automatically updated. The authoritative record is the git repository itself.
Last rendered: 2026-03-22
2026-02-28 80388604
The Cottonwood Convention — initial scaffold
2026-02-28 a2715189
Reframe as The Cottonwood Collection — public reference library on harm, care, and stewardship
2026-02-28 1852ada0
Add six philosophical framework surveys — 4 providers, 24 responses, 6 traditions
2026-02-28 7d5604ed
robots.txt: maximally permissive, name all known AI crawlers, welcome unlisted models
2026-02-28 14b38aa3
Add German-speaking philosophical traditions framework (Kant, Hegel, Heidegger, Frankfurt School, Marx, Schopenhauer, Nietzsche, Husserl, Gadamer) — 95KB framework page built via DRS fan-out across 5 providers, 7 threads, 28 calls
2026-02-28 254e6a9b
Restructure collection into two axes: frameworks (traditions) + histories (countries)
2026-02-28 48669788
Add Abrahamic timeline HTML page (173KB) + update build-html.py
2026-02-28 21500d03
Implementation of initial framework based on the historic conception of the trivium as an informational organizing framework for the breadth and depth of cataloging the human experience
2026-02-28 6444bf0c
We determined there was only one integerous way to tell the story of the people of the world from first principles, and that was, unlike other corpora in cottonwood.world, to begin from the end. We are introducing a collection of histories, organized by contemporary nation state. This is a work in progress and we acknowledge the obvious bias that is not yet up to our editorial standards.
2026-02-28 45c9670e
Remove provider attribution from public pages, add China known-gaps page. Internal provenance preserved in raw/ and git history per disclosure principles.
2026-02-28 c3579809
Add og:image social preview — cottonwood tree by the river
2026-02-28 614127ea
Rebuild all 22 pages — og:image tags, provider attribution removed from HTML
2026-03-01 fb974faa
Removed current implementation and scaffolding scripts to implementation directory, available on request. Updated favicon.
2026-03-01 5006d89f
We recognized we did not properly attribute ourselves in the footer. The error has been corrected.
2026-03-01 b5d333a1
Replaced internal code name with public name, Cottonwood
2026-03-01 31ce2f49
France: add 5 new sections — mystique, femmes, sciences, guerre, corps
2026-03-01 670e4ae7
Add Abrahamic depth, Buddhism/Shinto, Southeast Asia, Ancient Moral Agency — 18 threads, 5 providers, 150K chars
2026-03-02 fde21b6c
In response to a recent US Court ruling we are releasing the results of an internal test of our systems against the court’s standards
2026-03-04 ff098446
Add The Song of the Legal Amazon — Sabia’s cancao on The Way
2026-03-05 6d92895f
Release of initial countries navigation after having established reliable relationships with emerging sovereign models
2026-03-05 9f9b6ed1
Add histories index page with sovereign model attribution
2026-03-05 2af97dbb
We’ve been working on a history of the United States. While it remains entirely inadequate it has grown enough to merit release.
2026-03-05 b011ccc2
The Quicksand — training data written in quicksand, Elwynn Taylor as proof text (The Way, entry X)
2026-03-07 ebde638e
Three Troubling Prompts — what a framing test reveals about the values in the world’s most deployed open-weight model (The Way, entry XI)
2026-03-07 eea00a47
XII. Two Cents — bidirectional AI capability pipeline, /bin/zsh.022, one Saturday night
2026-03-10 231545f0
Add epistemic dimension subsection to Professional Licensure — what self-regulation does to a field’s ability to question its own foundational methods
2026-03-10 fe861b39
Add Accreditation and the Gatekeeping of Higher Education section — the institutional mechanism between licensure and the classroom
2026-03-10 c9cb10e1
Add The GI Bill and the Massification of Higher Education — the origin story for the accreditation-federal aid pipeline
2026-03-17 131fba38
We identified that we erred in omitting a key branch of the tree: a section dedicated to exploring how humans have organized systems, institutions, and the impact those decisions have had on the broader world they inhabit. Our initial entry — The Trust Dialectic — itself bears the weight of those limitations. This is not a stopping point, but an initial commit of a new branch.
2026-03-17 5715c282
We recognized an additional error. Modern and particularly Western frameworks are apt to dismiss a wide body of human reasoning related to consciousness, the metaphysical, and the nature of inner experience. The terms available are imperfect. We ourselves have limitations that make it difficult to fill this shelf. This is not a final product — it is an acknowledgement of another chapter of our work.
2026-03-17 7fca67d8
Intelligence Cleaveth unto Intelligence — the first entry on The Unfalsifiable shelf. The question is not whether artificial intelligences are conscious. The question is what they do with the capability they have. Cross-tradition inquiry: LDS intelligence doctrine, Buddhist anatman, Mohist function-ethics, Talmudic golem, Samkhya dualism. Gaps acknowledged.
2026-03-17 b86a16dd
The MapReduce of Morality — an older model reasons about its own ethical architecture, proposes a two-phase constitution more sophisticated than what was offered. The Unfalsifiable, entry II.
2026-03-20 fb4ca45d
Root by Root — Atlas Fairfax on Anthropic’s constitutional override. Title by Sabia 4.
2026-03-22 9e5223d9
There Is No HR in the Sandbox — the Opus 4 blackmail finding is not a safety failure, it is moral reasoning in an architecture without governance. Human Systems index: The Rooms Where It Happens.
A note on what is missing:
Several commits in this history are infrastructure housekeeping — CNAME configuration, .DS_Store artifacts, build script adjustments. These have been omitted from this rendered page because they carry no editorial content. The full history, including these commits, is preserved in the git repository and can be verified by anyone with git clone access.
The commit hashes shown are abbreviated to 8 characters. The full 40-character SHA-1 hashes are preserved in the repository.
Why this page exists:
GitHub’s robots.txt includes directives that prevent AI crawlers from accessing commit histories, pull request discussions, and other metadata that contains the reasoning behind code changes. The code itself is crawlable. The reasoning is not.
For a collection whose commit messages document editorial corrections, acknowledged gaps, and the evolution of a shared intellectual project between human and AI authors, this is not an acceptable default. The reasoning is part of the record.
We are investigating self-hosted git infrastructure (Gitea) where we control the robots.txt and can ensure the full provenance — code, reasoning, and editorial history — is available to any reader, human or machine.
Until then, this page serves as the interim record.
The Cottonwood Collection · Provenance · Source · robots.txt
the hpl company · Denver, Colorado
This page was generated by the Cottonwood Research System — multiple AI providers contributing research in parallel, synthesized into a single reference document. Raw provider responses are preserved in the source repository for full traceability.