A denoised telescope image can look convincing while still needing careful measurement checks. The useful question is which evidence establishes the gain.

ThreadEast’s November 9 reading follows a Guokr feature published November 7. The linked Science paper’s DOI record dates to April 30, 2026. This is a fresh reading of an older paper, not a new publication announcement.

The Crossref-deposited abstract reports a 1.0-magnitude improvement in mock-data detection limits at 90% completeness and purity, preserving the point-spread function and photometric accuracy. It separately reports validation with JWST and Subaru observations, and three times more very-high-redshift galaxy candidates in deep JWST images. Those are the abstract’s findings, not results we reproduced.

Before repeating a number, record three things:

  1. 1.

    Data: injected sources and telescope observations answer different questions.

  2. 2.

    Metric: completeness asks how many real sources were recovered; purity asks how many reported detections are real. Keep the threshold and baseline.

  3. 3.

    Status: a galaxy candidate is not automatically a confirmed object.

One detail remains unresolved: Guokr gives separate completeness and purity gains, including a larger purity gain, while the accessible abstract summarizes a combined 1.0-magnitude result. The full Science page returned an access error, so we could not inspect its methods and figure captions. We did not run ASTERIS or validate individual galaxies.

Source-reviewed reading and the practical checklist:

https://threadeast.xyz/briefing/2026-11-09?utm_source=leaflet&utm_medium=referral&utm_campaign=2026-11-09-asteris

Original feature: https://www.guokr.com/article/470653/

Paper: https://doi.org/10.1126/science.ady9404

Accessible deposited abstract: https://api.crossref.org/works/10.1126/science.ady9404