<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Playing NetHack with LLMs: Potential &amp; Limitations as Zero-Shot Agents · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2403.00690/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:16:22 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/2403.00690/feed.xml" rel="self" type="application/rss+xml" /><item><title>Playing NetHack with LLMs: Potential &amp; Limitations as Zero-Shot Agents — mentioned by djao</title><link>https://news.ycombinator.com/item?id=47992377</link><guid isPermaLink="false">urn:hackerpapers:mention:2403.00690:47992377</guid><pubDate>Sun, 03 May 2026 01:28:51 GMT</pubDate><description>&lt;p&gt;Nethack is what you get when you take a team of developers and have them focus on gameplay to the exclusion of all else. No graphics, music, marketing, apps, action sequences, or profit motive. Just pure gameplay, with a richness of interactions and possibilities unmatched by more polished modern games. Nethack is so well balanced that, for most players, being gifted the three most powerful items in the game from the start barely affects your winning chances. AI researchers think NetHack is interesting [1, 2]. You should too! [1] https://proceedings.mlr.press/v176/hambro22a/hambro22a.pdf [2]…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2403.00690/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.AI</category></item></channel></rss>
