<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: regexorcist</title><link>https://news.ycombinator.com/user?id=regexorcist</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 26 Aug 2026 05:51:18 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=regexorcist" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by regexorcist in "Apple introduces M6 and M5 Ultra"]]></title><description><![CDATA[
<p>Yeah for a lot of us macOS really gets in the way of using the otherwise excellent hardware. My daily driver is an M1 Pro with Asahi and it's easily my favorite computer ever. It's a real shame that the newer M models can't run Linux.</p>
]]></description><pubDate>Wed, 26 Aug 2026 01:52:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=49443220</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49443220</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49443220</guid></item><item><title><![CDATA[New comment by regexorcist in "Ornith-1.5: From Self-Scaffolding to Self-Improvement"]]></title><description><![CDATA[
<p>I get ~10 tok/sec on a M1 Ultra, but the real issue is abysmal prompt processing which makes it unusable. On the M1 and M2 series MTP is actively harmful for performance.</p>
]]></description><pubDate>Thu, 20 Aug 2026 03:12:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=49369955</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49369955</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49369955</guid></item><item><title><![CDATA[New comment by regexorcist in "Ornith-1.5: From Self-Scaffolding to Self-Improvement"]]></title><description><![CDATA[
<p>What signals? Looking more like a tiered release.</p>
]]></description><pubDate>Thu, 20 Aug 2026 03:07:52 +0000</pubDate><link>https://news.ycombinator.com/item?id=49369931</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49369931</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49369931</guid></item><item><title><![CDATA[New comment by regexorcist in "Incident with Github.com [resolved]"]]></title><description><![CDATA[
<p>Any good LLM should be able to come up with the equivalent git and curl commands fairly easily.</p>
]]></description><pubDate>Mon, 17 Aug 2026 18:35:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=49335630</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49335630</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49335630</guid></item><item><title><![CDATA[New comment by regexorcist in "Qwen 3.8 27B is excellent, but it defaults to overthinking things"]]></title><description><![CDATA[
<p>I'm doing much the same, avoid the long thinking loops and instead have more iterations on the plan with reviews from different angles.</p>
]]></description><pubDate>Mon, 17 Aug 2026 05:54:37 +0000</pubDate><link>https://news.ycombinator.com/item?id=49326980</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49326980</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49326980</guid></item><item><title><![CDATA[New comment by regexorcist in "Choosing an AI model: one prompt, 11 models, different results"]]></title><description><![CDATA[
<p>When I set up my local AI server I tried different harnesses and one prompt was to build a site for a consultancy in my domain. Not long after an actual agency I know had a website redesign and it was the exact same thing I got from my local model and the Hermes agent with the web design skill.</p>
]]></description><pubDate>Fri, 14 Aug 2026 14:11:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=49298960</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49298960</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49298960</guid></item><item><title><![CDATA[New comment by regexorcist in "Delayed Gratification – Proud to Be 'Last to Breaking News'"]]></title><description><![CDATA[
<p>6 or 7 years here, no news or socials. I keep telling people the same thing, you find out about the important stuff real quick regardless. The rest is noise and rage bait.</p>
]]></description><pubDate>Tue, 28 Jul 2026 21:56:51 +0000</pubDate><link>https://news.ycombinator.com/item?id=49090451</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49090451</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49090451</guid></item><item><title><![CDATA[New comment by regexorcist in "Our position on open-weights models"]]></title><description><![CDATA[
<p>He is well aware. This is just another pathetic attempt at being handed a monopoly based on fear-mongering and protectionism.</p>
]]></description><pubDate>Tue, 28 Jul 2026 11:12:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=49082167</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49082167</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49082167</guid></item><item><title><![CDATA[New comment by regexorcist in "Netflix employee fired for sharing personal details in retreat trust exercise"]]></title><description><![CDATA[
<p>Indeed, and I kinda miss it. My first couple IT jobs were right in Central London and we would all religiously meet at the pub after work, along with hundreds from other companies of all kinds. The banter was glorious.</p>
]]></description><pubDate>Tue, 28 Jul 2026 10:43:45 +0000</pubDate><link>https://news.ycombinator.com/item?id=49081949</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49081949</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49081949</guid></item><item><title><![CDATA[New comment by regexorcist in "Our position on open-weights models"]]></title><description><![CDATA[
<p>Don't be so naive. He only wants a monopoly and a massive IPO.</p>
]]></description><pubDate>Tue, 28 Jul 2026 09:21:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=49081376</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49081376</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49081376</guid></item><item><title><![CDATA[New comment by regexorcist in "Our position on open-weights models"]]></title><description><![CDATA[
<p>Most hated company in record time, this Dario is insufferable.</p>
]]></description><pubDate>Tue, 28 Jul 2026 08:54:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=49081179</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49081179</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49081179</guid></item><item><title><![CDATA[New comment by regexorcist in "Open-weight AI is having its Kubernetes moment"]]></title><description><![CDATA[
<p>> there is plenty of reporting suggesting their response will be similar<p>There is? Xi Jinping himself openly talked very recently about China's commitment to open weights models.</p>
]]></description><pubDate>Sun, 26 Jul 2026 12:33:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=49057488</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49057488</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49057488</guid></item><item><title><![CDATA[New comment by regexorcist in "Startup founders urge U.S. government not to shut off Chinese open weight AI"]]></title><description><![CDATA[
<p>They are lobbying hard for a monopoly, or duopoly at best, in AI, based on protectionism rather than merit.</p>
]]></description><pubDate>Fri, 24 Jul 2026 11:33:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=49034087</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49034087</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49034087</guid></item><item><title><![CDATA[New comment by regexorcist in "Startup founders urge U.S. government not to shut off Chinese open weight AI"]]></title><description><![CDATA[
<p>> the U.S. can “sanction” companies that engage in intellectual property theft<p>The hypocrisy is off the charts.</p>
]]></description><pubDate>Fri, 24 Jul 2026 11:25:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=49034000</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49034000</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49034000</guid></item><item><title><![CDATA[New comment by regexorcist in "LG to ban residential proxies from smart TV apps"]]></title><description><![CDATA[
<p>I have that setup for my home network regardless. And my TV is the original C1 OLED, nothing compared when I got it. I wouldn't buy LG today.</p>
]]></description><pubDate>Wed, 22 Jul 2026 12:23:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=49005628</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49005628</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49005628</guid></item><item><title><![CDATA[New comment by regexorcist in "LG to ban residential proxies from smart TV apps"]]></title><description><![CDATA[
<p>They hardcode DNS servers into the TV. You need a firewall in front of the pihole intercepting that traffic, I use OpenWrt.</p>
]]></description><pubDate>Wed, 22 Jul 2026 09:21:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=49003880</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=49003880</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49003880</guid></item><item><title><![CDATA[New comment by regexorcist in "OpenAI and Hugging Face address security incident during model evaluation"]]></title><description><![CDATA[
<p>OAI and HF basically saying that Chinese models are the only practical countermeasure available to us plebs. Got it.</p>
]]></description><pubDate>Tue, 21 Jul 2026 23:43:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=48999908</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=48999908</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48999908</guid></item><item><title><![CDATA[New comment by regexorcist in "China’s open-weights AI strategy is winning"]]></title><description><![CDATA[
<p>I won't be giving a dime to the companies that stole all the IP in the world and are now gatekeeping it, acting as the rightful owners, and trashing the hardware market for everyone else in the process. So it's Chinese models or nothing for me.</p>
]]></description><pubDate>Tue, 21 Jul 2026 16:07:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=48994201</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=48994201</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48994201</guid></item><item><title><![CDATA[New comment by regexorcist in "Nativ: Run frontier open models locally on your Mac"]]></title><description><![CDATA[
<p>Same here. Tried MLX twice at different times after reading the claims here but it always does considerably worse for me than llamacpp.</p>
]]></description><pubDate>Tue, 21 Jul 2026 09:29:45 +0000</pubDate><link>https://news.ycombinator.com/item?id=48989989</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=48989989</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48989989</guid></item><item><title><![CDATA[New comment by regexorcist in "Annoying and alarming things about OpenCode"]]></title><description><![CDATA[
<p>Thanks for that, I've been using it but some of these are bad. Specially that it sends data to the cloud by default for session titles, since I use my local model exclusively. So I'm done with it, trying maki now which seems promising and it's so much lighter.</p>
]]></description><pubDate>Mon, 20 Jul 2026 18:38:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=48983023</link><dc:creator>regexorcist</dc:creator><comments>https://news.ycombinator.com/item?id=48983023</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48983023</guid></item></channel></rss>