<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: elpakal</title><link>https://news.ycombinator.com/user?id=elpakal</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Fri, 20 Nov 2026 23:36:49 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=elpakal" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by elpakal in "Faster Retrieval for Better, More Efficient RAG"]]></title><description><![CDATA[
<p>Your understanding was applied too strictly and you missed the intent of the joke...</p>
]]></description><pubDate>Wed, 11 Nov 2026 19:16:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=50457285</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=50457285</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50457285</guid></item><item><title><![CDATA[New comment by elpakal in "Faster Retrieval for Better, More Efficient RAG"]]></title><description><![CDATA[
<p>Agreed, but I wonder how that plays out without them entering the frontier model space. I just downloaded the new foundation models cli and ran some inference on my Mac, it wasn't great, but it was neat to see. I asked which model it was using and it said gpt 4o mini.<p>An on-device, 100% private AI assistant seems like it could be so big that it could be a legitimate threat to the frontier model providers they are partnering with right now, imho.</p>
]]></description><pubDate>Wed, 11 Nov 2026 19:12:56 +0000</pubDate><link>https://news.ycombinator.com/item?id=50457237</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=50457237</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50457237</guid></item><item><title><![CDATA[New comment by elpakal in "Claude Whiteboards, Claude Sheets, and Claude Data"]]></title><description><![CDATA[
<p>In some ways it's becoming buy vs spend on tokens</p>
]]></description><pubDate>Wed, 11 Nov 2026 01:49:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=50448882</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=50448882</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50448882</guid></item><item><title><![CDATA[New comment by elpakal in "AI Isn't Replacing SREs. It's Deskilling Them"]]></title><description><![CDATA[
<p>Agreed. I think there's also an argument to be made that it's upskilling non SREs to now be able to do SRE things.</p>
]]></description><pubDate>Fri, 30 Oct 2026 16:33:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=50306335</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=50306335</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50306335</guid></item><item><title><![CDATA[New comment by elpakal in "US tech stocks see sharp falls over AI cost fears"]]></title><description><![CDATA[
<p>Google's AI search is abysmal and is almost always, at least, 30% wrong (outdated, wrong info) ime. I find that I just use the chat interface to save tokens on basic things so I don't want to waste my real AI subscription tokens on basics.<p>In theory I understand their change to make search -> AI search instead, but I have a hard time believing that this was benchmarked in any way and not rushed to desperately remain relevant in the AI race.</p>
]]></description><pubDate>Fri, 30 Oct 2026 16:26:54 +0000</pubDate><link>https://news.ycombinator.com/item?id=50306241</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=50306241</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50306241</guid></item><item><title><![CDATA[New comment by elpakal in "Anthropic S-1"]]></title><description><![CDATA[
<p>What incentive do they have to implement the large token/cost savings model architectures like Jev have introduced? Doesn't the impact Jev/Decisions models have had imply an inefficiency in the business models here?</p>
]]></description><pubDate>Mon, 26 Oct 2026 20:21:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=50250353</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=50250353</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50250353</guid></item><item><title><![CDATA[New comment by elpakal in "Apple HomeView"]]></title><description><![CDATA[
<p>I have an old alexa and a paper calendar we literally write on for our family stuff. And the wall ends up getting written on.<p>So this is two birds for me</p>
]]></description><pubDate>Tue, 13 Oct 2026 17:14:20 +0000</pubDate><link>https://news.ycombinator.com/item?id=50075211</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=50075211</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50075211</guid></item><item><title><![CDATA[New comment by elpakal in "Apple HomeView"]]></title><description><![CDATA[
<p>"I will answer that for you when you put down the knife"</p>
]]></description><pubDate>Tue, 13 Oct 2026 14:50:04 +0000</pubDate><link>https://news.ycombinator.com/item?id=50072643</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=50072643</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50072643</guid></item><item><title><![CDATA[New comment by elpakal in "Decisions API is in public beta"]]></title><description><![CDATA[
<p>I would love to it just needs to stay within our data isolation boundaries (bedrock)</p>
]]></description><pubDate>Wed, 07 Oct 2026 15:11:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=49993937</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49993937</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49993937</guid></item><item><title><![CDATA[New comment by elpakal in "Decisions API is in public beta"]]></title><description><![CDATA[
<p>In bedrock it’s the other way around</p>
]]></description><pubDate>Wed, 07 Oct 2026 15:11:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=49993931</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49993931</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49993931</guid></item><item><title><![CDATA[New comment by elpakal in "Decisions API is in public beta"]]></title><description><![CDATA[
<p>Have there been any signals from Anthropic about matching this? We use AWS bedrock and just switched to Anthropic from OpenAI because of the ZDR guarantee. Would be great to not have to entertain switching back.</p>
]]></description><pubDate>Wed, 07 Oct 2026 02:51:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=49987483</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49987483</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49987483</guid></item><item><title><![CDATA[Let AI agents search across workspace with the Universal Search MCP Server]]></title><description><![CDATA[
<p>Article URL: <a href="https://developers.google.com/workspace/guides/universal-search-mcp">https://developers.google.com/workspace/guides/universal-search-mcp</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=49794043">https://news.ycombinator.com/item?id=49794043</a></p>
<p>Points: 1</p>
<p># Comments: 0</p>
]]></description><pubDate>Mon, 21 Sep 2026 22:02:10 +0000</pubDate><link>https://developers.google.com/workspace/guides/universal-search-mcp</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49794043</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49794043</guid></item><item><title><![CDATA[New comment by elpakal in "Astra for Law"]]></title><description><![CDATA[
<p>I find the watermarking dynamic to be really interesting in the legal space, as more large model providers provide increasingly powerful legal capabilities, and adoption (presumably) also increases. Attorneys aren’t the same as developers as their work can be traced back to them, and there are personal bar licenses and reputations at stake. I wonder if knowing the likelihood that AI generated something helps or hurts in that respect.<p>I also see a lot of watermark removal services popping up as a result.</p>
]]></description><pubDate>Thu, 17 Sep 2026 23:19:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=49748073</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49748073</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49748073</guid></item><item><title><![CDATA[New comment by elpakal in "Astra for Law"]]></title><description><![CDATA[
<p>where are the benchmarks</p>
]]></description><pubDate>Thu, 17 Sep 2026 20:58:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=49746477</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49746477</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49746477</guid></item><item><title><![CDATA[New comment by elpakal in "Astra for Law"]]></title><description><![CDATA[
<p>> By using the legal search index, Astra for Law can search U.S. case law, statutes, regulations, court rules, and administrative decisions across a corpus of more than 230 million URLs, with sources added daily. Our work with Free Law Project, the nonprofit behind CourtListener, brings its case-law collection covering more than 99.9% of published U.S. precedential case law (opens in a new window) into this research experience.<p>Found that to be very interesting</p>
]]></description><pubDate>Thu, 17 Sep 2026 20:55:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=49746449</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49746449</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49746449</guid></item><item><title><![CDATA[New comment by elpakal in "Claude Cowork and chat are now one Claude"]]></title><description><![CDATA[
<p>why tho? genuinely curious</p>
]]></description><pubDate>Wed, 16 Sep 2026 19:23:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=49731676</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49731676</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49731676</guid></item><item><title><![CDATA[New comment by elpakal in "Swift 6.4 Released"]]></title><description><![CDATA[
<p>Man I can't count the number of shell wrappers I've written or reached out to over my career. Looking forward to having an official Swift SubProcess helper.</p>
]]></description><pubDate>Tue, 15 Sep 2026 16:39:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=49715124</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49715124</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49715124</guid></item><item><title><![CDATA[New comment by elpakal in "Ask HN: What are you working on? (September 2026)"]]></title><description><![CDATA[
<p>I'm building SMS alerts for Kalshi market movements, upset alerts, big trades, specific teams and more. Starting a private beta (if interested, send a message via X @OddsLineSMS).</p>
]]></description><pubDate>Mon, 14 Sep 2026 14:32:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=49697545</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49697545</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49697545</guid></item><item><title><![CDATA[New comment by elpakal in "Shopify is moving from React Native back to Swift and Kotlin"]]></title><description><![CDATA[
<p>Very true, especially when there are different turn around times and review policies for the various app stores, things can get out of sync quickly</p>
]]></description><pubDate>Thu, 10 Sep 2026 22:20:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=49650913</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49650913</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49650913</guid></item><item><title><![CDATA[New comment by elpakal in "Claude Fable 5.1 and Claude Mythos 5.1"]]></title><description><![CDATA[
<p>From the changelog:<p>Whole-file rewrites for small changes. When editing text files, the model is more likely to rewrite the entire file than make a targeted edit. The result is usually the same, but the rewrite costs more output tokens and time.<p>So we are to catch that somehow? And then add their recommendation (below) to our prompts?<p><a href="https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5-1#prefer-targeted-edits-over-whole-file-rewrites" rel="nofollow">https://platform.claude.com/docs/en/build-with-claude/prompt...</a><p>If Claude Fable 5.1 rewrites whole files for small changes, append the following instruction to the system prompt or the first user message. Claude Fable 5.1 is more likely than Claude Fable 5 to rewrite an entire text file rather than make a targeted edit. The resulting file is usually the same, but unless the file is short or most of it is changing, a rewrite costs more output tokens and time. The instruction brings Claude Fable 5.1 back in line with Claude Fable 5 for small and medium changes.<p>> The number of tokens used to edit files is best minimized, all else being equal. Therefore, when it will not affect the end result, try to surgically edit a file rather than rewrite the entire thing.</p>
]]></description><pubDate>Tue, 01 Sep 2026 20:51:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=49528033</link><dc:creator>elpakal</dc:creator><comments>https://news.ycombinator.com/item?id=49528033</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49528033</guid></item></channel></rss>