<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/">
    <channel>
        <title>Clint Johnson — LLM</title>
        <link>https://www.clint-johnson.com/articles/tag/llm</link>
        <description>Articles tagged LLM by Clint Johnson.</description>
        <lastBuildDate>Tue, 23 Jun 2026 00:30:56 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>https://github.com/jpmonette/feed</generator>
        <image>
            <title>Clint Johnson — LLM</title>
            <url>https://www.clint-johnson.com/og-image.png</url>
            <link>https://www.clint-johnson.com/articles/tag/llm</link>
        </image>
        <copyright>All rights reserved 2026</copyright>
        <item>
            <title><![CDATA[RAG Isn't Dead. You're Just Using It Wrong.]]></title>
            <link>https://www.clint-johnson.com/articles/rag-isnt-dead</link>
            <guid isPermaLink="false">https://www.clint-johnson.com/articles/rag-isnt-dead</guid>
            <pubDate>Thu, 29 Jan 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The 'RAG is dead' narrative is wrong, but it's wrong in an interesting way. What kills LLM context quality in production, and what to do about it.]]></description>
            <content:encoded><![CDATA[The 'RAG is dead' narrative is wrong, but it's wrong in an interesting way. What kills LLM context quality in production, and what to do about it.]]></content:encoded>
            <author>clint@1putthealth.com (Clint Johnson)</author>
            <enclosure url="https://images.clint-johnson.com/articles/rag-isnt-dead/hero.png" length="0" type="image/png"/>
        </item>
        <item>
            <title><![CDATA[Fine-Tuning a 70B Model on a Consumer GPU: The Q-LoRA Practical Guide]]></title>
            <link>https://www.clint-johnson.com/articles/qlora-consumer-gpu-finetuning</link>
            <guid isPermaLink="false">https://www.clint-johnson.com/articles/qlora-consumer-gpu-finetuning</guid>
            <pubDate>Sat, 17 Jan 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Q-LoRA + SFTTrainer + Flash Attention v2 means you can fine-tune a 70B parameter model on 24GB of VRAM. What that looks like end-to-end, what it costs in quality, and when to just use the API instead.]]></description>
            <content:encoded><![CDATA[Q-LoRA + SFTTrainer + Flash Attention v2 means you can fine-tune a 70B parameter model on 24GB of VRAM. What that looks like end-to-end, what it costs in quality, and when to just use the API instead.]]></content:encoded>
            <author>clint@1putthealth.com (Clint Johnson)</author>
            <enclosure url="https://images.clint-johnson.com/articles/qlora-consumer-gpu-finetuning/hero.png" length="0" type="image/png"/>
        </item>
        <item>
            <title><![CDATA[When Recommendations Meet Language: The LLM-RecSys Convergence]]></title>
            <link>https://www.clint-johnson.com/articles/llm-recsys-convergence</link>
            <guid isPermaLink="false">https://www.clint-johnson.com/articles/llm-recsys-convergence</guid>
            <pubDate>Mon, 04 Nov 2024 00:00:00 GMT</pubDate>
            <description><![CDATA[Most AI stacks treat the recommendation engine and the language model as two separate systems that hand off to each other. A new class of hybrid models eliminates that seam. The implications for domain-specific AI are significant.]]></description>
            <content:encoded><![CDATA[Most AI stacks treat the recommendation engine and the language model as two separate systems that hand off to each other. A new class of hybrid models eliminates that seam. The implications for domain-specific AI are significant.]]></content:encoded>
            <author>clint@1putthealth.com (Clint Johnson)</author>
            <enclosure url="https://images.clint-johnson.com/articles/llm-recsys-convergence/hero.png" length="0" type="image/png"/>
        </item>
    </channel>
</rss>