<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>AI on Omanshu Thapliyal</title><link>https://omanshuthapliyal.github.io/tags/ai/</link><description>Recent content in AI on Omanshu Thapliyal</description><generator>Hugo</generator><language>en</language><lastBuildDate>Sat, 17 Oct 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://omanshuthapliyal.github.io/tags/ai/index.xml" rel="self" type="application/rss+xml"/><item><title>Most Open Problems in AI Safety are Problems of Observation</title><link>https://omanshuthapliyal.github.io/blog/bluedot-pipeline-open-problems/</link><pubDate>Sat, 17 Oct 2026 00:00:00 +0000</pubDate><guid>https://omanshuthapliyal.github.io/blog/bluedot-pipeline-open-problems/</guid><description>&lt;p>I recently finished BlueDot&amp;rsquo;s &lt;a href="https://bluedot.org/courses/technical-ai-safety">Technical AI Safety&lt;/a> course.
The reading list is long. Read one unit at a time, it is easy to lose the shape of the whole.&lt;/p>
&lt;p>So I did what one does with a pile of components. I laid all 99 readings out along the system they belong to: the lifecycle of a model, from the data it learns from to the point where it is served and retrained.
What came out of it is fairly simple.&lt;/p></description></item></channel></rss>