<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: fanyangxyz33</title><link>https://news.ycombinator.com/user?id=fanyangxyz33</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 02 Jul 2026 09:24:01 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=fanyangxyz33" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by fanyangxyz33 in "Devin Fusion: Frontier Performance at 35% Lower Cost"]]></title><description><![CDATA[
<p>I'm very curious about how they trained the lightweight classifier to decide model switching. Is it supervised? Did they use LLM as a teacher? It also seems that the featurization isn't trivial. Like how you build a simple but still meaningful representation of the task(s).</p>
]]></description><pubDate>Tue, 30 Jun 2026 22:25:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=48740062</link><dc:creator>fanyangxyz33</dc:creator><comments>https://news.ycombinator.com/item?id=48740062</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48740062</guid></item><item><title><![CDATA[Devin Fusion: Frontier Performance at 35% Lower Cost]]></title><description><![CDATA[
<p>Article URL: <a href="https://cognition.com/blog/devin-fusion">https://cognition.com/blog/devin-fusion</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48739985">https://news.ycombinator.com/item?id=48739985</a></p>
<p>Points: 2</p>
<p># Comments: 1</p>
]]></description><pubDate>Tue, 30 Jun 2026 22:14:51 +0000</pubDate><link>https://cognition.com/blog/devin-fusion</link><dc:creator>fanyangxyz33</dc:creator><comments>https://news.ycombinator.com/item?id=48739985</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48739985</guid></item><item><title><![CDATA[New comment by fanyangxyz33 in "The Codex App"]]></title><description><![CDATA[
<p>I really look forward to using this. I tried Codex first time yesterday and it was able to complete a task (i.e. drawing Penrose tilings) that Claude Code previously failed at. Also a little overwhelmed by all the new features that this app brings. I feel that I'm behind all the fancy new tools.</p>
]]></description><pubDate>Mon, 02 Feb 2026 18:29:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=46859405</link><dc:creator>fanyangxyz33</dc:creator><comments>https://news.ycombinator.com/item?id=46859405</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46859405</guid></item><item><title><![CDATA[Penrose Tiling Visualization]]></title><description><![CDATA[
<p>Article URL: <a href="https://fanyangxyz.github.io/penrose-ii/">https://fanyangxyz.github.io/penrose-ii/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=45622199">https://news.ycombinator.com/item?id=45622199</a></p>
<p>Points: 7</p>
<p># Comments: 0</p>
]]></description><pubDate>Fri, 17 Oct 2025 21:17:20 +0000</pubDate><link>https://fanyangxyz.github.io/penrose-ii/</link><dc:creator>fanyangxyz33</dc:creator><comments>https://news.ycombinator.com/item?id=45622199</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45622199</guid></item><item><title><![CDATA[Pokémon Color Palette Swapping]]></title><description><![CDATA[
<p>Article URL: <a href="https://fanyangxyz.github.io/2025/10/04/recoloring/">https://fanyangxyz.github.io/2025/10/04/recoloring/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=45553845">https://news.ycombinator.com/item?id=45553845</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Sat, 11 Oct 2025 23:55:33 +0000</pubDate><link>https://fanyangxyz.github.io/2025/10/04/recoloring/</link><dc:creator>fanyangxyz33</dc:creator><comments>https://news.ycombinator.com/item?id=45553845</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45553845</guid></item></channel></rss>