<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Self-Improving on Hanchen's Space Bar</title><link>https://hanchenli.github.io/blog/tags/self-improving/</link><description>Recent content in Self-Improving on Hanchen's Space Bar</description><generator>Hugo -- 0.148.0</generator><language>en-us</language><lastBuildDate>Thu, 30 Apr 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://hanchenli.github.io/blog/tags/self-improving/index.xml" rel="self" type="application/rss+xml"/><item><title>Hanchen's ICLR 2026 Reflections on Self-Improving AI</title><link>https://hanchenli.github.io/blog/posts/continual_learning/</link><pubDate>Thu, 30 Apr 2026 00:00:00 +0000</pubDate><guid>https://hanchenli.github.io/blog/posts/continual_learning/</guid><description>This blog reflects on my experience at ICLR 2026, focusing on self-improving AI. It discusses the reclassification of continual learning based on assumptions about future tasks and introduces the three pillars of evolving agents: models, harness, and environments.</description></item></channel></rss>