http://blog.asleson.org/tags/ai-datasets/ Recent content in AI Datasets on blog.asleson.org Hugo en-us tony dot asleson@gmail.com (Tony Asleson) tony dot asleson@gmail.com (Tony Asleson) &copy; Copyright 1999-2025 Tony Asleson Tue, 03 Mar 2026 00:00:00 +0000 http://blog.asleson.org/2026/03/03/honey-i-shrunk-the-model-maybe-blk-archive-vs-ai-data/ Tue, 03 Mar 2026 00:00:00 +0000tony dot asleson@gmail.com (Tony Asleson) http://blog.asleson.org/2026/03/03/honey-i-shrunk-the-model-maybe-blk-archive-vs-ai-data/ <h2 id="because-it-should-work-isnt-data">Because “It Should Work” Isn’t Data</h2> <p>After reading about the billions spent on AI infrastructure, I kept wondering: how much of that storage is just&hellip; the same bytes over and over? So I decided to find out. As I&rsquo;m involved with <a href="https://github.com/device-mapper-utils/blk-archive">https://github.com/device-mapper-utils/blk-archive</a> I thought it would be good to understand how much storage <code>blk-archive</code> can realistically save when pointed at AI-style datasets.</p> <p>Let me be clear upfront: this isn&rsquo;t a speed test or hardware benchmark. I&rsquo;m only interested in one question: how many bytes go in, and how many come out?</p>