<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Covenant-72B: Pre-Training a 72B LLM with Trustless Peers Over-the-Internet · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2603.08163/</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/2603.08163/feed.xml" rel="self" type="application/rss+xml" /><item><title>Covenant-72B: Pre-Training a 72B LLM with Trustless Peers Over-the-Internet — mentioned by Lichtso</title><link>https://news.ycombinator.com/item?id=50294647</link><guid isPermaLink="false">urn:hackerpapers:mention:2603.08163:50294647</guid><pubDate>Thu, 29 Oct 2026 19:38:14 GMT</pubDate><description>&lt;p&gt;Covenant-72B https://arxiv.org/abs/2603.08163 https://huggingface.co/1Covenant/Covenant-72B&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2603.08163/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.DC</category></item><item><title>Covenant-72B: Pre-Training a 72B LLM with Trustless Peers Over-the-Internet — mentioned by logicchains</title><link>https://news.ycombinator.com/item?id=48515107</link><guid isPermaLink="false">urn:hackerpapers:mention:2603.08163:48515107</guid><pubDate>Sat, 13 Jun 2026 09:02:36 GMT</pubDate><description>&lt;p&gt;It is practical, albeit not as efficient: https://arxiv.org/abs/2603.08163 . But organizing enough people with decent-enough GPUs is the challenge.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2603.08163/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.DC</category></item><item><title>Covenant-72B: Pre-Training a 72B LLM with Trustless Peers Over-the-Internet — mentioned by logicchains</title><link>https://news.ycombinator.com/item?id=48515098</link><guid isPermaLink="false">urn:hackerpapers:mention:2603.08163:48515098</guid><pubDate>Sat, 13 Jun 2026 09:01:21 GMT</pubDate><description>&lt;p&gt;&amp;gt;I&amp;#x27;ve been contemplating a decentralized model training system for some time using volunteer machines that we all contribute. But, it is astronomically difficult. The communication speeds are untenable. It is already possible: https://arxiv.org/abs/2603.08163 . You don&amp;#x27;t need to sync so frequently, so it can be done over normal internet, it&amp;#x27;s just less efficient (takes longer to converge).&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2603.08163/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.DC</category></item><item><title>Covenant-72B: Pre-Training a 72B LLM with Trustless Peers Over-the-Internet — HN submission</title><link>https://news.ycombinator.com/item?id=47335107</link><guid isPermaLink="false">urn:hackerpapers:submission:2603.08163:47335107</guid><pubDate>Wed, 11 Mar 2026 13:09:49 GMT</pubDate><description>&lt;p&gt;Covenant-72B: Pre-Training a 72B LLM with Trustless Peers Over-the-Internet&lt;/p&gt;&lt;p&gt;&lt;/p&gt;&lt;p&gt;5 points; 2 discussion replies on HN. &lt;a href="https://hackerpaper.kaipfister.de/p/2603.08163/"&gt;Paper and source history&lt;/a&gt;. A submission is not an endorsement.&lt;/p&gt;</description><category>cs.DC</category></item></channel></rss>
