<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>RSI4Science Brief</title><link>https://rsi4science.org/brief/</link><description>A free community platform collecting scientific datasets, benchmarks and checking tools for Recursive Self-Improvement. Explore resources, use original citations and contribute.</description><language>en</language><atom:link href="https://rsi4science.org/rss.xml" rel="self" type="application/rss+xml"/><item><title>MetaSkill-Evolve reuses an edited improvement procedure</title><link>https://rsi4science.org/brief/metaskill-evolve-editable-improvement/</link><guid>https://rsi4science.org/brief/metaskill-evolve-editable-improvement/</guid><description>Evolved meta-skills guide later skill and meta-skill updates. The recursive mechanism is explicit; causal attribution and scientific transfer remain separate questions.</description><pubDate>Sat, 03 Oct 2026 00:00:00 GMT</pubDate></item><item><title>ScienceBuddy: scientific adaptation at the boundary of RSI</title><link>https://rsi4science.org/brief/sciencebuddy-feedback-and-procedures/</link><guid>https://rsi4science.org/brief/sciencebuddy-feedback-and-procedures/</guid><description>Three model–harness cycles carry improvements forward, while a fixed auxiliary model edits the harness. This does not establish a stronger self-improver.</description><pubDate>Sat, 03 Oct 2026 00:00:00 GMT</pubDate></item><item><title>What counts as recursive self-improvement for science?</title><link>https://rsi4science.org/brief/what-counts-as-improvement/</link><guid>https://rsi4science.org/brief/what-counts-as-improvement/</guid><description>An accepted capability must help drive later self-improvement. Showing that loop exists and showing that recursion helps are separate claims.</description><pubDate>Sat, 03 Oct 2026 00:00:00 GMT</pubDate></item></channel></rss>