
  <rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
      <title>Ashim Sharma</title>
      <link>https://ashimsharma10.github.io/blog</link>
      <description>Software engineer sharing projects, notes, and guides on ML infrastructure.</description>
      <language>en-us</language>
      <managingEditor>sharmaashim00@gmail.com (Ashim Sharma)</managingEditor>
      <webMaster>sharmaashim00@gmail.com (Ashim Sharma)</webMaster>
      <lastBuildDate>Mon, 07 Sep 2026 00:00:00 GMT</lastBuildDate>
      <atom:link href="https://ashimsharma10.github.io/tags/physical-ai/feed.xml" rel="self" type="application/rss+xml"/>
      
  <item>
    <guid>https://ashimsharma10.github.io/blog/world-models</guid>
    <title>World Models and the Path to Physical AI</title>
    <link>https://ashimsharma10.github.io/blog/world-models</link>
    <description>A world model is a learned simulator. Give it a state and an action and it predicts what happens next. This write-up follows the idea from Ha and Schmidhuber&#39;s V+M+C agent through Dreamer, MuZero and TD-MPC, to Genie, Cosmos and V-JEPA, the drift that comes with imagining too far ahead, LeCun&#39;s case for predicting representations instead of pixels, and why robots need all of this.</description>
    <pubDate>Mon, 07 Sep 2026 00:00:00 GMT</pubDate>
    <author>sharmaashim00@gmail.com (Ashim Sharma)</author>
    <category>world-models</category><category>reinforcement-learning</category><category>robotics</category><category>physical-ai</category><category>jepa</category><category>model-based-rl</category>
  </item>

    </channel>
  </rss>
