<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>CodeClash: Benchmarking Goal-Oriented Software Engineering · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2511.00839/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:27:26 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/2511.00839/feed.xml" rel="self" type="application/rss+xml" /><item><title>CodeClash: Benchmarking Goal-Oriented Software Engineering — HN submission</title><link>https://news.ycombinator.com/item?id=45824582</link><guid isPermaLink="false">urn:hackerpapers:submission:2511.00839:45824582</guid><pubDate>Wed, 05 Nov 2025 16:13:16 GMT</pubDate><description>&lt;p&gt;Show HN: New eval from SWE-bench team evalutes LMs based on goals not tickets&lt;/p&gt;&lt;p&gt;Current evals test LMs on tasks: &amp;quot;fix this bug,&amp;quot; &amp;quot;write a test&amp;quot; But we code to achieve goals: maximize revenue, cut costs, win users Meet CodeClash: LMs compete via their codebases across multi-round tournaments to achieve high-level goals. Because real software dev isn’t about following instructions. It’s about achieving outcomes. Here&amp;#x27;s how it works: Two LMs enter a tournament. Each maintains its own codebase. Every round: 1. Edit Phase: LMs modify their codebases however they like 2. Competition phase: Codebases battle in an arena. 3. Repeat The LM that wins the majority of rounds is…&lt;/p&gt;&lt;p&gt;5 points; 1 discussion reply on HN. &lt;a href="https://hackerpaper.kaipfister.de/p/2511.00839/"&gt;Paper and source history&lt;/a&gt;. A submission is not an endorsement.&lt;/p&gt;</description><category>cs.SE</category></item></channel></rss>
