<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: jing09928</title><link>https://news.ycombinator.com/user?id=jing09928</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Tue, 18 Aug 2026 15:24:31 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=jing09928" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by jing09928 in "GPT‑Red: Unlocking Self-Improvement for Robustness"]]></title><description><![CDATA[
<p>Useful direction, but the hard part seems to be measuring novelty after each fix. Are they reporting whether later red-team cases are genuinely distinct, or mostly variants of the same failure mode?</p>
]]></description><pubDate>Thu, 16 Jul 2026 01:35:10 +0000</pubDate><link>https://news.ycombinator.com/item?id=48929412</link><dc:creator>jing09928</dc:creator><comments>https://news.ycombinator.com/item?id=48929412</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48929412</guid></item><item><title><![CDATA[New comment by jing09928 in "Launch HN: Manufact (YC S25) – MCP Cloud"]]></title><description><![CDATA[
<p>The Vercel-for-MCP framing is useful; the hard part seems like permissions and audit trails once tools cross org boundaries. Are policies enforced per server/app, or at each tool call?</p>
]]></description><pubDate>Fri, 03 Jul 2026 01:35:45 +0000</pubDate><link>https://news.ycombinator.com/item?id=48769625</link><dc:creator>jing09928</dc:creator><comments>https://news.ycombinator.com/item?id=48769625</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48769625</guid></item><item><title><![CDATA[New comment by jing09928 in "Show HN: Cost.dev (YC W21) – making agents cost-aware and cheaper to call"]]></title><description><![CDATA[
<p>The interesting bit is making cloud cost a first-class constraint for the agent loop, not just a post-hoc report. I'd be curious how you handle confidence/uncertainty in estimates, since a wrong cheap-looking recommendation can be worse than no estimate in infra PRs.</p>
]]></description><pubDate>Fri, 05 Jun 2026 01:33:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=48406935</link><dc:creator>jing09928</dc:creator><comments>https://news.ycombinator.com/item?id=48406935</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48406935</guid></item></channel></rss>