<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: akeel_mohamed</title><link>https://news.ycombinator.com/user?id=akeel_mohamed</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 18 Nov 2026 03:36:57 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=akeel_mohamed" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by akeel_mohamed in "Launch HN: Momentic (YC W24) – Agents that catch bugs by using your app"]]></title><description><![CDATA[
<p>The benchmark is the part I would push on, and I say that from assessment work. A test built by the people selling the tool, with bugs labelled by the same team, will usually flatter the tool, not because anyone cheated, but because the cases reflect what the builders already think counts as a bug.<p>Planted and labelled bugs are also cleaner than real ones. Real failures are ambiguous, they cross features, and reasonable testers disagree about whether they are bugs at all. That disagreement is not noise to remove. In hiring tests we learned it is often the most useful signal, because it shows whether the scoring holds up when the answer is not obvious.<p>19.3 percent recall at 84 percent precision may still be worth $2.67 a bug. I just would not compare that number to another agent run on a different set, under different rules. Hold the cases, the labelling, and the cost accounting constant, or the comparison is mostly marketing.</p>
]]></description><pubDate>Wed, 04 Nov 2026 20:24:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=50370994</link><dc:creator>akeel_mohamed</dc:creator><comments>https://news.ycombinator.com/item?id=50370994</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50370994</guid></item></channel></rss>