<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: ChadNauseam</title><link>https://news.ycombinator.com/user?id=ChadNauseam</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Mon, 20 Apr 2026 23:36:04 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=ChadNauseam" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by ChadNauseam in "Claude Token Counter, now with model comparisons"]]></title><description><![CDATA[
<p>How would it be a money grab? If the new tokenizer requires more tokens to encode the same information, it costs them more money for inference. The point of charging per token is that the cost is proportional to the number of tokens. That's my understanding anyway</p>
]]></description><pubDate>Mon, 20 Apr 2026 06:30:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=47831022</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47831022</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47831022</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Changes in the system prompt between Claude Opus 4.6 and 4.7"]]></title><description><![CDATA[
<p>That's just not true. Makes me wondered if you've ever bought a bottle of alcohol before lol. There's no label that says it causes cancer. (Maybe in california because of prop 65?) And I expect cars also have no such labelling, not that it would matter, considering they cause cancer in random passers by who have no opportunity to consent to breathing in auto exhaust or read any labels</p>
]]></description><pubDate>Mon, 20 Apr 2026 03:48:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=47830216</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47830216</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47830216</guid></item><item><title><![CDATA[New comment by ChadNauseam in "The seven programming ur-languages (2022)"]]></title><description><![CDATA[
<p>How does this disagree with "the base language must be total and can't be Turing complete"</p>
]]></description><pubDate>Mon, 20 Apr 2026 03:45:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=47830204</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47830204</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47830204</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Changes in the system prompt between Claude Opus 4.6 and 4.7"]]></title><description><![CDATA[
<p>If you get cancer from drinking alcohol, smoking cigarettes or breathing particles emitted by ICE engines in their standard course of operation, you generally can't sue the manufacturer.</p>
]]></description><pubDate>Mon, 20 Apr 2026 00:10:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=47828907</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47828907</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47828907</guid></item><item><title><![CDATA[New comment by ChadNauseam in "A perfectable programming language"]]></title><description><![CDATA[
<p>> So the claim "tend to grow them" ... it is not completely wrong, but it also does not fully capture an independent want to add them. It comes ALWAYS from people who WANT types.<p>Who else would add them, besides people who want them? I'm confused about what you're even claiming here. It sounds like you feel that there's a vocal minority of type enthusiasts who everyone else is just humoring by letting them bolt on their type systems.</p>
]]></description><pubDate>Mon, 13 Apr 2026 06:32:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=47748378</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47748378</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47748378</guid></item><item><title><![CDATA[Give Them Two Choices]]></title><description><![CDATA[
<p>Article URL: <a href="https://chadnauseam.com/coding/tips/give-them-two-choices">https://chadnauseam.com/coding/tips/give-them-two-choices</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47745751">https://news.ycombinator.com/item?id=47745751</a></p>
<p>Points: 1</p>
<p># Comments: 0</p>
]]></description><pubDate>Sun, 12 Apr 2026 23:45:58 +0000</pubDate><link>https://chadnauseam.com/coding/tips/give-them-two-choices</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47745751</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47745751</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Have Codex review Claude's work"]]></title><description><![CDATA[
<p>It's very important in my opinion to have Codex review Claude's work. They make different types of errors and both get tunnel-vision. I prefer to drive with Claude, but I set up this hook to automatically invoke a Codex code review every time Claude tries to commit. I find it dramatically improves the performance of either of the models individually.<p>You can just paste the entire gist into your Claude session to have Claude set it up for you</p>
]]></description><pubDate>Wed, 08 Apr 2026 23:43:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=47697604</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47697604</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47697604</guid></item><item><title><![CDATA[Have Codex review Claude's work]]></title><description><![CDATA[
<p>Article URL: <a href="https://gist.github.com/anchpop/56688d3526e1d4019db7c4fc1bcb2b2c">https://gist.github.com/anchpop/56688d3526e1d4019db7c4fc1bcb2b2c</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47697603">https://news.ycombinator.com/item?id=47697603</a></p>
<p>Points: 4</p>
<p># Comments: 1</p>
]]></description><pubDate>Wed, 08 Apr 2026 23:43:23 +0000</pubDate><link>https://gist.github.com/anchpop/56688d3526e1d4019db7c4fc1bcb2b2c</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47697603</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47697603</guid></item><item><title><![CDATA[New comment by ChadNauseam in "I won't download your app. The web version is a-ok"]]></title><description><![CDATA[
<p>Same. My app is a PWA. Most users won’t install a PWA and won’t repeatedly navigate to a website so it limits the reach. But the advantage is that I can deploy instantly. I love when someone sends a bug report and I can tell them it’s fixed ten minutes later. Pretty great, compared to “it’ll be fixed in there business days” you get with the iOS app store</p>
]]></description><pubDate>Mon, 06 Apr 2026 16:24:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=47662980</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47662980</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47662980</guid></item><item><title><![CDATA[New comment by ChadNauseam in "How many products does Microsoft have named 'Copilot'?"]]></title><description><![CDATA[
<p>> I guess they really just didn't want a product name to start with the name of a competitor's product.<p>Probably, but I doubt linux wants it either. People might think it's some official linux product.</p>
]]></description><pubDate>Sun, 05 Apr 2026 00:21:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=47644909</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47644909</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47644909</guid></item><item><title><![CDATA[New comment by ChadNauseam in "An NSFW filter for Marginalia search"]]></title><description><![CDATA[
<p>Does marginalia_nu not use embedding models as part of search? I guess I assumed it would. If you have embeddings anyway, decision trees on the embedding vector (e.g. catboost) tend to work pretty well. Fine-tuning modernbert works even better but probably won't meet the criteria of "really fast and run well on CPUs". That said, the approach described in the article seems to work well enough and obviously provides extremely cheap inference</p>
]]></description><pubDate>Mon, 30 Mar 2026 19:55:18 +0000</pubDate><link>https://news.ycombinator.com/item?id=47578954</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47578954</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47578954</guid></item><item><title><![CDATA[New comment by ChadNauseam in "I use Excalidraw to manage my diagrams for my blog"]]></title><description><![CDATA[
<p>Why does it cause tool fragmentation? You can change the font to a normal sans-serif font or to a monospaced font.<p>Personally I like the default font. It looks weird to have my crappy doodles next to a normal computer font. The default one is very legible but has a style (and ligatures) that make it feel not too neat</p>
]]></description><pubDate>Mon, 30 Mar 2026 18:51:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=47578198</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47578198</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47578198</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Copilot edited an ad into my PR"]]></title><description><![CDATA[
<p>That doesn't really make sense. So it's an ad for raycast? But raycast said they didn't know about it. To me the explanation makes perfect sense. "You can use this tool with raycast" seems like a very reasonable tip.</p>
]]></description><pubDate>Mon, 30 Mar 2026 17:34:11 +0000</pubDate><link>https://news.ycombinator.com/item?id=47577280</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47577280</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47577280</guid></item><item><title><![CDATA[New comment by ChadNauseam in "If you don't opt out by Apr 24 GitHub will train on your private repos"]]></title><description><![CDATA[
<p>The situation you describe has dynamics that don't apply when your windows laptop is trying to get you to install an update. A woman can't have 100% confidence that saying no won't trigger a man into rage, so just the question being asked at all is already a bit unpleasant. WinRAR trying to get me to buy a license is not as offensive because I know it won't beat me up for saying no.</p>
]]></description><pubDate>Fri, 27 Mar 2026 21:24:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=47548450</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47548450</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47548450</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Cory Doctorow: Interoperability Can Save the Open Web"]]></title><description><![CDATA[
<p>If you buy a google pixel 9 (the last version for which google released device trees), you can do anything you want on your phone. My pixel runs a version of android I built myself</p>
]]></description><pubDate>Thu, 26 Mar 2026 16:51:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=47532808</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47532808</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47532808</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Swift 6.3"]]></title><description><![CDATA[
<p>What are you using for rust-swift interop?</p>
]]></description><pubDate>Thu, 26 Mar 2026 16:28:11 +0000</pubDate><link>https://news.ycombinator.com/item?id=47532523</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47532523</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47532523</guid></item><item><title><![CDATA[The Future of Trash [pdf]]]></title><description><![CDATA[
<p>Article URL: <a href="https://dsny.cityofnewyork.us/wp-content/uploads/reports/future-of-trash-april-2023.pdf">https://dsny.cityofnewyork.us/wp-content/uploads/reports/future-of-trash-april-2023.pdf</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47524891">https://news.ycombinator.com/item?id=47524891</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Wed, 25 Mar 2026 23:47:50 +0000</pubDate><link>https://dsny.cityofnewyork.us/wp-content/uploads/reports/future-of-trash-april-2023.pdf</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47524891</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47524891</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Finding all regex matches has always been O(n²)"]]></title><description><![CDATA[
<p>It does but apparently not to the point that it bothers most people. So I guess I wouldn't worry about it. Keep up the good work on making regex faster!</p>
]]></description><pubDate>Tue, 24 Mar 2026 17:18:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=47506072</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47506072</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47506072</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Finding all regex matches has always been O(n²)"]]></title><description><![CDATA[
<p>Haha. But I guess it proves my assertion that "we can still tell" was wrong. Scary to think about the future though. Soon less RLHF'd models will be available, and it won't be so easy to recognize their tells. I don't think the whole post was written by an LLM in one prompt and I'm sure it was actually a lot of work to write, but yeah, it was definitely enhanced by one.<p>It reminds me of how peer review is supposed to be blind, but certain authors like Simon Peyton Jones have such a distinctive voice that surely reviewers instantly know when a paper is by him. The models have been successfully made to talk in a particular way, and it's very clear once you becomes familiar with it. My guess is that most HN users are not familiar and I just sound like a crazy person to them<p>The author admitted it in another comment anyway <a href="https://news.ycombinator.com/item?id=47494833">https://news.ycombinator.com/item?id=47494833</a></p>
]]></description><pubDate>Tue, 24 Mar 2026 17:15:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=47505998</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47505998</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47505998</guid></item><item><title><![CDATA[New comment by ChadNauseam in "Free Multilingual Dictionaries"]]></title><description><![CDATA[
<p>These are some multilingual dictionaries^[1] I made. They're designed to be useful to language learners. They have some cool features I haven't seen elsewhere.<p>1. you can see the frequency of each <i>meaning</i> of each word. For example, the french word "bois" commonly means "drink" and rarely means "wood": <a href="https://yap.town/d/french-to-english/bois/" rel="nofollow">https://yap.town/d/french-to-english/bois/</a><p>2. Each meaning also comes with a huge number of example sentences, taken from "real" sources (movies mostly)<p>3. It also has definitions for common phrases and constructions. For example, "c'est" is technically two words in french, but you should just learn it as one unit, so it gets its own entry: <a href="https://yap.town/d/french-to-english/c-est/" rel="nofollow">https://yap.town/d/french-to-english/c-est/</a><p>You can also see a list of phrases for any language of interest: <a href="https://yap.town/d/french-to-english/top-1000-phrases/" rel="nofollow">https://yap.town/d/french-to-english/top-1000-phrases/</a> . The phrases were discovered via the "unigram" tokenization algorithm (originally invented for training LLMs!)<p>Something remarkable about the phrases is that you can often understand the meaning of a sentence just from the meaning of each phrase in it. The concatenated phrases sometimes even form grammatical english. And of course, if the same phrase has multiple meanings, those are separated and have their own example sentences, just like the words do.<p>^[1]: And by "I", I mean I wrote code to ask an LLM to generate each dictionary entry.</p>
]]></description><pubDate>Mon, 23 Mar 2026 17:25:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=47492469</link><dc:creator>ChadNauseam</dc:creator><comments>https://news.ycombinator.com/item?id=47492469</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47492469</guid></item></channel></rss>