<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Internet-augmented language models through few-shot prompting for open-domain question answering · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2203.05115/</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/2203.05115/feed.xml" rel="self" type="application/rss+xml" /><item><title>Internet-augmented language models through few-shot prompting for open-domain question answering — HN submission</title><link>https://news.ycombinator.com/item?id=49333187</link><guid isPermaLink="false">urn:hackerpapers:submission:2203.05115:49333187</guid><pubDate>Mon, 17 Aug 2026 16:03:11 GMT</pubDate><description>&lt;p&gt;Show HN: OpenWebSearch – A router for web search indexes&lt;/p&gt;&lt;p&gt;Hi HN, I&amp;#x27;m one of the people behind OpenWebSearch (https://openwebsearch.ai). It&amp;#x27;s a router for web search indexes. You POST to one endpoint with a `provider` field, and it normalizes both the request and the response for web search indexes like Parallel, Brave, Exa and more Why we built it: we run a model company (Interfaze) and a lot of our models are smaller in size and we&amp;#x27;re experimenting if given web search can a smaller 9b or 70b model perform the same as 300b or 600b model and we found that it does extremely better when given web search similar to this paper…&lt;/p&gt;&lt;p&gt;3 points; 3 discussion replies on HN. &lt;a href="https://hackerpaper.kaipfister.de/p/2203.05115/"&gt;Paper and source history&lt;/a&gt;. A submission is not an endorsement.&lt;/p&gt;</description><category>cs.CL</category></item></channel></rss>
