<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>IdeaScore blog</title>
    <link>https://ideascore.online/blog/</link>
    <description>Notes on running better calls: rubrics, rounds, grounded AI evaluation and programme administration.</description>
    <language>en-IN</language>
    <lastBuildDate>Fri, 18 Sep 2026 03:30:00 GMT</lastBuildDate>
    <atom:link href="https://ideascore.online/blog/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>How to design a rubric committees trust</title>
      <link>https://ideascore.online/blog/how-to-design-a-rubric-committees-trust/</link>
      <guid isPermaLink="true">https://ideascore.online/blog/how-to-design-a-rubric-committees-trust/</guid>
      <pubDate>Fri, 18 Sep 2026 03:30:00 GMT</pubDate>
      <description>Criteria, descriptors, weights and anchoring, and why a 0 to 10 scale with written descriptors produces a ranking a committee can defend.</description>
      <category>Rubrics</category>
      <category>Evaluation</category>
    </item>
    <item>
      <title>Running a multi-round seed-fund call</title>
      <link>https://ideascore.online/blog/running-a-multi-round-seed-fund-call/</link>
      <guid isPermaLink="true">https://ideascore.online/blog/running-a-multi-round-seed-fund-call/</guid>
      <pubDate>Fri, 11 Sep 2026 03:30:00 GMT</pubDate>
      <description>A timeline, three rounds, promotion rules written before the scores exist, and the workload arithmetic that decides how many evaluators you need.</description>
      <category>Rounds</category>
      <category>Programme design</category>
    </item>
    <item>
      <title>What grounded AI evaluation means</title>
      <link>https://ideascore.online/blog/what-grounded-ai-evaluation-means/</link>
      <guid isPermaLink="true">https://ideascore.online/blog/what-grounded-ai-evaluation-means/</guid>
      <pubDate>Fri, 04 Sep 2026 03:30:00 GMT</pubDate>
      <description>The difference between an assistant that writes a plausible paragraph and one that researches and cites, and where an AI co-evaluator belongs in scoring.</description>
      <category>IdeaScore AI</category>
      <category>Evaluation</category>
    </item>
    <item>
      <title>Resubmissions, duplicates and lineage</title>
      <link>https://ideascore.online/blog/resubmissions-duplicates-and-lineage/</link>
      <guid isPermaLink="true">https://ideascore.online/blog/resubmissions-duplicates-and-lineage/</guid>
      <pubDate>Fri, 28 Aug 2026 03:30:00 GMT</pubDate>
      <description>Teams resubmit for good reasons and bad ones. How to tell the difference fairly, keep a lineage chain, and give evaluators one version to score.</description>
      <category>Review</category>
      <category>Integrity</category>
    </item>
  </channel>
</rss>
