About

The Human Data Collective (HDC) is a draft proposal for a decentralized network of human-generated content created, curated, and labelled specifically for the purpose of training large language models, paired with an open inference standard, AIISP-1, that routes verifiable per-request payments into environmental remediation and direct creator and reviewer compensation.

This site

This site exists for one purpose: to host the whitepaper in a form where any reader can leave a comment on any sentence, the way you would mark up a Google Doc. Comments are public, threaded, and anchored to the exact text you selected.

Suggesting changes to the paper or the standard

The paper sources, the AIISP-1 specification text, the reference implementation, and the open tracking issue for this draft all live at github.com/97115104/aiisp-spec ↗ Substantive edits to prose, citations, schemas, or the spec itself should be proposed in the GitHub repos as pull requests or issues. Comments left on this site are read by the author and folded into subsequent drafts; PRs against the spec are versioned on the public record.

Get in touch

Editorial review, partnership, pilot funding, worker-organization or community co-authorship, and similar correspondence: [email protected].

Author

Austin Harshberger, Happy Stack Calculus. Personal links: links.97115104.com ↗

Licences and attestation

Paper © 2026 Happy Stack Calculus LLC, released under CC BY-NC-ND 4.0; share and attribute, no derivatives or commercial use without permission.
Spec CC BY 4.0 for the text and MIT for the reference code, at the spec repository.
This site MIT.
Attestation View the attestation record for this paper here ↗