<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: stingraycharles</title><link>https://news.ycombinator.com/user?id=stingraycharles</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sun, 12 Apr 2026 09:16:03 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=stingraycharles" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by stingraycharles in "How We Broke Top AI Agent Benchmarks: And What Comes Next"]]></title><description><![CDATA[
<p>This is already well known, all these AI benchmarks use a different model to judge whether or not the solution was correct.<p>It’s… remarkably poor, and as demonstrated in the paper, easily gamed. Worst yet, these benchmarks teach AIs to be very short-sighted and hyper-focused on completing the task, rather than figuring out the best solution.</p>
]]></description><pubDate>Sun, 12 Apr 2026 05:09:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=47736305</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47736305</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47736305</guid></item><item><title><![CDATA[New comment by stingraycharles in "Apple Silicon and Virtual Machines: Beating the 2 VM Limit (2023)"]]></title><description><![CDATA[
<p>This implies anyone doing this using VMware violates the EULA?</p>
]]></description><pubDate>Sun, 12 Apr 2026 02:38:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=47735706</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47735706</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47735706</guid></item><item><title><![CDATA[New comment by stingraycharles in "Apple Silicon and Virtual Machines: Beating the 2 VM Limit (2023)"]]></title><description><![CDATA[
<p>Nah, Apple has been extremely restrictive about virtual machines in all kinds of ways, e.g. the minimum terms anyone is able to lease out a VM or Mac to someone else is 24h, making cloud-like workloads practically impossible. For some reason, Apple really doesn’t like virtual machines, and it’s much more intentional than just “probably multiple users”.<p>It’s extremely frustrating.</p>
]]></description><pubDate>Sun, 12 Apr 2026 02:36:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=47735697</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47735697</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47735697</guid></item><item><title><![CDATA[New comment by stingraycharles in "OpenClaw’s memory is unreliable, and you don’t know when it will break"]]></title><description><![CDATA[
<p>Thanks for this suggestion, I installed it yesterday after seeing this comment and this surely is a breath of fresh air! It appears that everything is designed reasonably well from the ground up. It’s more limited, but what’s there works well.</p>
]]></description><pubDate>Sun, 12 Apr 2026 01:55:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=47735533</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47735533</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47735533</guid></item><item><title><![CDATA[New comment by stingraycharles in "BlueHammer abuses Windows Defender's update process to gain SYSTEM access"]]></title><description><![CDATA[
<p>This website is one of the most user hostile I’ve encountered in a long time. On mobile it’s miserable.</p>
]]></description><pubDate>Sat, 11 Apr 2026 11:11:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=47729549</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47729549</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47729549</guid></item><item><title><![CDATA[New comment by stingraycharles in "OpenClaw’s memory is unreliable, and you don’t know when it will break"]]></title><description><![CDATA[
<p>Yeah I’m amazed how much constantly breaks between every release, I’m dreading to update it. I’m using it as a kind of personal assistant that keeps track of all my outstanding work. Morning and evening briefing, can get an overview of what I did per week / month, it’s pretty solid for keeping track of all this.<p>But it’s so terribly unstable, it’s as if nobody actually tests things before pushing a release. I don’t need 2 updates per day, I just need one every few weeks that’s stable.</p>
]]></description><pubDate>Sat, 11 Apr 2026 04:48:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=47727496</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47727496</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47727496</guid></item><item><title><![CDATA[New comment by stingraycharles in "Industrial design files for Keychron keyboards and mice"]]></title><description><![CDATA[
<p>Why do you like rhe 96% layout? I prefer the full space, recently upgraded to a Q6 Ultra, I’m really happy I can get a wireless version of most of their models these days.</p>
]]></description><pubDate>Fri, 10 Apr 2026 16:54:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=47720838</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47720838</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47720838</guid></item><item><title><![CDATA[New comment by stingraycharles in "Industrial design files for Keychron keyboards and mice"]]></title><description><![CDATA[
<p>Keychron just open-sourced their design files, which I didn’t expect.<p>I’m a happy user of their keyboards.</p>
]]></description><pubDate>Fri, 10 Apr 2026 16:24:18 +0000</pubDate><link>https://news.ycombinator.com/item?id=47720440</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47720440</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47720440</guid></item><item><title><![CDATA[Industrial design files for Keychron keyboards and mice]]></title><description><![CDATA[
<p>Article URL: <a href="https://github.com/Keychron/Keychron-Keyboards-Hardware-Design">https://github.com/Keychron/Keychron-Keyboards-Hardware-Design</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47720419">https://news.ycombinator.com/item?id=47720419</a></p>
<p>Points: 447</p>
<p># Comments: 140</p>
]]></description><pubDate>Fri, 10 Apr 2026 16:22:26 +0000</pubDate><link>https://github.com/Keychron/Keychron-Keyboards-Hardware-Design</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47720419</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47720419</guid></item><item><title><![CDATA[New comment by stingraycharles in "Instant 1.0, a backend for AI-coded apps"]]></title><description><![CDATA[
<p>This is exactly it, and people trying to invent these new tools and formats (wasn’t there a “LLM optimized JSON alternative the other day) are seriously not understanding how LLMs are trained a tremendous amount on understanding JSON and XML and standard tools and frameworks.</p>
]]></description><pubDate>Fri, 10 Apr 2026 12:40:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=47717238</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47717238</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47717238</guid></item><item><title><![CDATA[New comment by stingraycharles in "We've raised $17M to build what comes after Git"]]></title><description><![CDATA[
<p>Yes, this is the obvious solution. Multiple agents working on multiple features should use feature branches.<p>Can’t believe how this whole AI movement seems to want to reinvent software engineering, poorly.</p>
]]></description><pubDate>Fri, 10 Apr 2026 11:31:18 +0000</pubDate><link>https://news.ycombinator.com/item?id=47716513</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47716513</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47716513</guid></item><item><title><![CDATA[New comment by stingraycharles in "Scientists invented a fake disease. AI told people it was real"]]></title><description><![CDATA[
<p>This is already a thing for a year or so, SEO for AI results to make sure that your products are recommended in ChatGPT.<p><a href="https://citeworksstudio.com/" rel="nofollow">https://citeworksstudio.com/</a> is a decent one.</p>
]]></description><pubDate>Fri, 10 Apr 2026 11:11:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=47716325</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47716325</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47716325</guid></item><item><title><![CDATA[New comment by stingraycharles in "I still prefer MCP over skills"]]></title><description><![CDATA[
<p>Obviously opaque binaries are hardly an improvement over MCP, but providing a few curl + jq oneliners to interact with a REST API works great in my experience. Also means no external scripts, just a single markdown file.</p>
]]></description><pubDate>Fri, 10 Apr 2026 07:23:11 +0000</pubDate><link>https://news.ycombinator.com/item?id=47714728</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47714728</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47714728</guid></item><item><title><![CDATA[New comment by stingraycharles in "I still prefer MCP over skills"]]></title><description><![CDATA[
<p>I often just put direct curl commands in a skill, the agent uses that, and it works perfectly for custom API integrations. Agents are perfectly capable of doing these types of things, and it means the LLM just uses a flexible set of tools to achieve almost anything.</p>
]]></description><pubDate>Fri, 10 Apr 2026 06:51:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=47714461</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47714461</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47714461</guid></item><item><title><![CDATA[New comment by stingraycharles in "We moved Railway's frontend off Next.js. Builds went from 10+ mins to under 2"]]></title><description><![CDATA[
<p>Besides, most time is spent running tests anyway.</p>
]]></description><pubDate>Thu, 09 Apr 2026 03:34:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=47698993</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47698993</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47698993</guid></item><item><title><![CDATA[New comment by stingraycharles in "Understanding Traceroute"]]></title><description><![CDATA[
<p>So these types of routers are typically running at 100% CPU and - likely - not keeping up with packets that are not in the fast path? Ie the operators really only care about the packets in the fast path, all others can be ignored?<p>Or are there other types of packets in the slow path that <i>do</i> get a delivery guarantee by the router?<p>I’m curious how these tradeoffs are made.</p>
]]></description><pubDate>Thu, 09 Apr 2026 03:31:10 +0000</pubDate><link>https://news.ycombinator.com/item?id=47698979</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47698979</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47698979</guid></item><item><title><![CDATA[New comment by stingraycharles in "Muse Spark: Scaling towards personal superintelligence"]]></title><description><![CDATA[
<p>You need money to buy resources and manpower.</p>
]]></description><pubDate>Thu, 09 Apr 2026 03:20:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=47698927</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47698927</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47698927</guid></item><item><title><![CDATA[New comment by stingraycharles in "I've been waiting over a month for Anthropic to respond to my billing issue"]]></title><description><![CDATA[
<p>Yeah this would make a lot of sense to crack, given that customer support must be a huge potential revenue stream for them. Starting by fixing their own support would make sense, given that it’s a relatively limited in scope.</p>
]]></description><pubDate>Thu, 09 Apr 2026 01:47:54 +0000</pubDate><link>https://news.ycombinator.com/item?id=47698425</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47698425</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47698425</guid></item><item><title><![CDATA[New comment by stingraycharles in "Sonnet 4.6 Elevated Rate of Errors"]]></title><description><![CDATA[
<p>If you really think the value add by LLMs is comparable to email and calendar, I don’t understand why you don’t understand my point that you’re not the customer Anthropic cares about.</p>
]]></description><pubDate>Wed, 08 Apr 2026 12:59:01 +0000</pubDate><link>https://news.ycombinator.com/item?id=47689566</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47689566</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47689566</guid></item><item><title><![CDATA[New comment by stingraycharles in "Your File System Is Already A Graph Database"]]></title><description><![CDATA[
<p>Using the same logic, a key/value database is also a graph database?<p>Isn’t the biggest benefit of graph databases the indexing and additional query constructs they support, like shortest path finding and whatnot?</p>
]]></description><pubDate>Wed, 08 Apr 2026 12:54:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=47689527</link><dc:creator>stingraycharles</dc:creator><comments>https://news.ycombinator.com/item?id=47689527</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47689527</guid></item></channel></rss>