Work

Before the Wisdom Machine: An Analytical History of the Early Whitepapers

Year: 2026
Type: preprint
Abstract: The histories of technical research programs are commonly reconstructed from their consolidated publications. Earlier documents then appear as incomplete versions of the concepts that became prominent later. This retrospective convenience can obscure failed conjectures, independent lines of inquiry, and changes in the evidentiary demands imposed on a proposal. The present paper develops a source-critical reconstruction of an early collection of artificial-intelligence and computational-architecture whitepapers associated with public GitHub repositories in 2025. The working inventory comprises 59 candidate document records associated with 56 repositories, including four separately listed files in a shared collection. It is a candidate archive rather than a verified census of unique papers. A staged documentary audit has examined 17 currently accessible full texts, 22 repository summaries or READMEs, six file-path or catalog records, and 14 metadata-only records. The study distinguishes explicit intertextual citations from conceptual analogy and tests selected relationships against established literature and later publications. Recurring investigational archetypes include recursive inquiry, instrumental self-extension, context continuity, declarative agency, distributed semantic coordination, and qualified judgment. The findings support several local documentary genealogies, but not a single deterministic origin story. They also reveal revision ambiguity, misidentified references, and a chronology in which books sometimes precede the whitepapers that cite them. The contribution is an auditable method and a qualified analytical catalog of the surviving candidate archive. No claim is made that the proposals were scientifically validated, that all candidates are distinct whitepapers, or that later works necessarily descend from them.
Language: English
Open Access status: green
OpenAlex ID: W7235165034
Cites:
Cited by:
Related to: 10