<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: locallylinear</title><link>https://news.ycombinator.com/user?id=locallylinear</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sat, 07 Nov 2026 16:27:31 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=locallylinear" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by locallylinear in "A Jev that can ask Opus for help gives Opus accuracy at 4x lower cost"]]></title><description><![CDATA[
<p>The most glaring downside with practically using decision models is if you need the model to not only decide something but give a reason, for every case, you end up not saving money or time, but just adding an extra step.<p>Give me a decision model that can do both, more expensive than today’s decision models but still cheaper than full on frontier model, and it’d open up so many more use cases, IMO</p>
]]></description><pubDate>Sat, 07 Nov 2026 10:52:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=50403410</link><dc:creator>locallylinear</dc:creator><comments>https://news.ycombinator.com/item?id=50403410</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50403410</guid></item><item><title><![CDATA[New comment by locallylinear in "A Jev that can ask Opus for help gives Opus accuracy at 4x lower cost"]]></title><description><![CDATA[
<p>> Calibration matters<p>I agree. What I’ve encountered though is even building enough cases for a calibration data set is hard. And so we farm out to smarter expensive models like Astra to label cases and use that as a calibration data set for the now cheaper LLM judge actually running as our judge<p>Is this standard ? Or am I just too lazy to manually label enough data to calibrate ?</p>
]]></description><pubDate>Sat, 07 Nov 2026 10:46:58 +0000</pubDate><link>https://news.ycombinator.com/item?id=50403367</link><dc:creator>locallylinear</dc:creator><comments>https://news.ycombinator.com/item?id=50403367</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50403367</guid></item><item><title><![CDATA[New comment by locallylinear in "A thousand fewer paths to root"]]></title><description><![CDATA[
<p>The ROI on this is massive if the bounties would actually be accepted (they won’t, right?)<p>It begs the interesting question on the future of bug bounty programs: when it’s so easy to find and solve issues there’s less incentive to offer a bounty, which in turn leads to less incentive for humans/agents to find and fix security issues..?<p>As someone not plugged in on the current times with respect to bug bounty programs, curious for peoples take on this / is this already the reality ?</p>
]]></description><pubDate>Thu, 05 Nov 2026 12:15:18 +0000</pubDate><link>https://news.ycombinator.com/item?id=50378210</link><dc:creator>locallylinear</dc:creator><comments>https://news.ycombinator.com/item?id=50378210</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50378210</guid></item><item><title><![CDATA[New comment by locallylinear in "Show HN: Subvocal – Talk to your computer without making noise"]]></title><description><![CDATA[
<p>Very cool, great work and awesome story of how you all preserved to get to this point.<p>Are you thinking at all about security certifications to more easily be approved with, or sell subscriptions to, enterprise companies?<p>I’d love a pair but realistically I couldn’t use them with work unless it was an approved vendor; we use Wispr Flow though so I imagine being on a similar plane to them would suffice.</p>
]]></description><pubDate>Tue, 03 Nov 2026 09:55:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=50350013</link><dc:creator>locallylinear</dc:creator><comments>https://news.ycombinator.com/item?id=50350013</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50350013</guid></item></channel></rss>