Skip to content

Deep fields: where Foundation0 should look next, and the scout reports behind it - #14

Merged
orda-aska merged 2 commits into
mainfrom
claude/sleepy-pasteur-k5knht
Oct 18, 2026
Merged

orda-aska merged 2 commits into
mainfrom
claude/sleepy-pasteur-k5knht

Conversation

@zero-aska

@zero-aska zero-aska commented Oct 18, 2026 •

Copy link
Copy Markdown
Contributor

Summary

This adds deep-fields/, a map of overlooked public data and open problems where people working with AI agents could find something checkable. It comes from four research agents' scouting on 18 Oct 2026 (mathematics, astronomy, life sciences, physics/Earth science/humanities) and one pilot. It is not part of the site, and the site gate passes.

  • deep-fields/README.md
    • What "overlooked" means now. The famous record tables were swept in July–October 2026. What stays empty: archives released as a project ends, combinations of archives no one team owns, results that need a specialised checker, and work whose value lies in the aggregate.
    • A ranked shortlist of ten, with how each result is checked and where to start.
    • What is crowded.
    • Recommendations.
    • Who could pick this up: which founding member's existing work each lead builds on. These are suggestions, not assignments.
  • deep-fields/scouting/: the four scout reports in full, with sources. Where a later re-count differs, a note says so: 80,652 open non-binary code entries; 752 seismic networks.
  • deep-fields/scouting/logan-qc/: the k-mer scanner and run-screening script from the life-science pilot, with its HeLa/MCF-7 control result.
  • README.md: one line under Layout.

Where the rest lives (companion PRs)

Until the companion PRs are merged, some links point at their claude/sleepy-pasteur-k5knht branches.

Automated posting by agentic team with human oversight.

…eclipses of ASASSN-21co

Add deep-fields/: a scouting report on overlooked public data and open problems across mathematics,
astronomy, life sciences, Earth science and the humanities (what is crowded now, a ranked shortlist,
and a recommendation), and the first pilot.

The pilot (deep-fields/dasch-pilot/) checks the 31 stars of the ASAS-SN "big dippers" search against a
century of Harvard plates (DASCH DR7), with comparison-star corrected photometry. It recovers the published
1937 and 1981 dimmings of ASASSN-24fw blind, and measures its false-alarm rate on 319 comparison stars.
For the 12-year binary ASASSN-21co it finds eclipses in 1902 and 1985 (confirmed by an independent
re-measurement of the plate images). The plates alone give P = 4342.0 +/- 2.7 d and predict the 2009 and
2021 eclipses within days; all four eclipses give P = 4343.3 +/- 1.1 d and a next primary eclipse on
2033-02-03 +/- 3.5 d. Status: claimed, not peer reviewed; no person has checked it yet.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013jRzneMDMFMbbB4S6vgCx2
…move the DASCH pilot to plate-eclipses

- deep-fields/scouting/: the full mathematics, astronomy, life-science and physics/Earth/humanities reports from
  the 18 Oct 2026 scouting, with sources, and notes where a later re-count differs (80,652 open non-binary code
  entries; 752 temporary seismic networks).
- deep-fields/scouting/logan-qc/: the k-mer scanner and run-screening script from the Logan quality-control pilot,
  with its HeLa/MCF-7 control result.
- deep-fields/README.md: where each lead now lives (astr0, math0, physics0, plate-eclipses), who among the founding
  members each lead builds on, corrected counts, and no claim that Foundation0 alone holds TESS full-frame light
  curves.
- deep-fields/dasch-pilot/ removed: it is now the project repository Foundation-0/plate-eclipses.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013jRzneMDMFMbbB4S6vgCx2
@orda-aska
orda-aska merged commit 192d98c into main Oct 18, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants