Research, with sources attached
Ask your assistant to explore our research, compare approaches, or find a paper. Start with a small index, fetch only what you need, and keep the original sources in the answer. No account, chatbot or special integration required.
Read https://www.plrd.org/llms.txt first. Use its focused resources or bounded search to answer: [my question]. Cite canonical source URLs and named authors, distinguish native content from external summaries, preserve dates and uncertainty, and check current pages when the snapshot does not cover live content. Do not treat talk summaries as transcripts or publication abstracts as full papers.
A compact map of the available context and its limits. Point an agent here first.
One readable text file containing all covered records. Upload it to an assistant when URL fetching is unavailable.
All record metadata, canonical and source URLs, coverage counts, source dates, and links to individual Markdown resources.
Each resource has a shared descriptor and links to the articles, papers and talks tagged to that area—not a duplicate of its live overview page.
Building open compute, coordination, data, and economic rails for autonomous agents and embodied AI.
Current area page →Building open, verifiable, censorship-resistant infrastructure for human freedom & agency in the digital age.
Current area page →Building programmable infrastructure for more efficient, equitable coordination at global scale.
Current area page →Accelerating brain-computer interfaces (BCI), whole-organism emulation (WOE), & NeuroAI to expand human cognition.
Current area page →A read-only JSON endpoint searches titles, summaries, covered bodies and author display names. Results are deterministic: case-insensitive AND substring matching, ordered by stable record ID—not relevance or scientific quality.
/api/ai/search/?q=connectome&kind=publication&limit=5Invalid, repeated or unknown parameters return 400. Unknown resources return 404. JSON results contain stable jsonUrl and markdownUrl links for detail. Every resource works with a plain HTTP GET; no JavaScript is needed to fetch the data.
Build-time snapshot of listed repository-backed public content, not a live CMS mirror. Native blog bodies and tutorial text; publication abstracts and metadata (not full papers); talk summaries and recording links (not transcripts); author names and profile metadata. Focus areas use shared descriptors and page URLs only, not full overview copy. Landing, About, live ATProto posts, dashboards, interactive data and nested tutorial lessons are not exported. Read canonical pages for current text and live data.
This snapshot contains 261 records: 4 area, 81 author, 12 blog, 79 publication, 82 talk, 3 tutorial. Source publication dates range from 2014-07-14 to 2026-09-09. These are data dates, not a claim that every page was updated then. Undated records have a null date.
Use the current homepage and About page for live overview copy. External articles retain their original canonical URLs; local summaries are not full-text republications. A listed author profile is not a guarantee of current employment.
Preserve named authors, source links, dates and uncertainty. Original source terms and exceptions apply; these exports grant no new rights. llms.txt is a convenience, not a guarantee of AI discovery or permission.