<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: barrell</title><link>https://news.ycombinator.com/user?id=barrell</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 19 Nov 2026 21:15:43 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=barrell" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by barrell in "The AI Superintelligence Debate Has Serious Flaws"]]></title><description><![CDATA[
<p>I would go so far as they are optimizing for such things.<p>I’ve told the story many times, but when gpt voice came out, I used it every day. I found it great to be able to talk with the internet as a way of research.<p>Then after a few months, it started clearing its throat. Then taking breaths. I stopped using it over the course of a week.<p>I don’t see any reason why a voice agent should sound convincingly human, clear it throat, take breaths, say hmm, giggle like a schoolgirl, sigh etc for any non-nepharious reasons. For me, it actively decreases the non-nefarious uses of the product, and increases more-or-less only the nefarious uses.<p>Then frontier labs go on to talk about how they’re worried their product would be used nefariously.</p>
]]></description><pubDate>Thu, 19 Nov 2026 14:04:20 +0000</pubDate><link>https://news.ycombinator.com/item?id=50560269</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50560269</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50560269</guid></item><item><title><![CDATA[New comment by barrell in "AI is killing All About Berlin"]]></title><description><![CDATA[
<p>I have a similar feeling. Personally I combat it by running my own company. While I use a bunch of LLMs, its an “AI-last company”. I code by hand, design by hand, write by hand, and edit by hand.<p>There are two things I tell myself to keep my head:<p>1. There are really two possible outcomes. Either “AGI/SI/RSI” is achieved, or it’s not. The former will change everything so drastically, there is literally no way to prepare for it or imagine the reprocusions. Nothing about that path is in my control, so I don’t spend any time thinking about it or worrying it. Doing so would be the equivalent of us practicing hiding under our desks for when the nuclear alarms go off. It helps that I don’t find it likely, but that “1% chance” lingered in the back of my brain until I embraced this mentality.<p>2. The amount of investor capital and debt involved in these businesses give me very little hope for their ability to innovate. The squeeze/enshittification will come faster than with any other technology. Those business that can outlast it will come back in demand. Private drivers were almost wiped out by uber, now I know several private drivers making a killing because people are tired of Ubers. AirBNBs were posed to take out hotels or BnBs or alternative lodging; nearly everyone I know now has sworn off airbnbs, and there are tons of cool alternative options.<p>It sucks that these companies can steal and get away with it. They’re destroying large swathes of society and ruining millions of people’s lives to make a few rich people even richer.<p>But eventually the check will come due and the music will end. Forest fires burn away hundred year old trees, but in doing so allow a new generation of seeds to take root. I refuse to believe the money inferno will burn up all human desire for uniqueness, creativity, and artisanry.<p>Keep doing what you’re doing, follow your passions, hone a craft, use AI as little as possible, refuse AI when you don’t want it, don’t get doomtrolled, and wait for the fire to burn out.</p>
]]></description><pubDate>Wed, 18 Nov 2026 14:20:34 +0000</pubDate><link>https://news.ycombinator.com/item?id=50546644</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50546644</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50546644</guid></item><item><title><![CDATA[New comment by barrell in "Apple has new Mac app coming to make running local AI models easy"]]></title><description><![CDATA[
<p>I would be extremely surprised if declining App Store usage due to LLMs ever effects Apple’s bottom line. Pretty sure the increase in developer licenses could easily offset any loss in profits, and might actually result in more revenue.<p>And it’s not like Apple is fixed. If this really did become a threat, Apple could surely integrate into Siri some ability to generate apps from your iPhone and put it behind their iCloud subscription.<p>They’ve had XCode for years, managed their entire UI library basically since conception, and have their own language for their apps. Fine tune DeepSeek for a while on swift iOS apps and I’m sure you could generate apps rivaling frontier models.<p>I really don’t see any existential threats for Apple here.</p>
]]></description><pubDate>Wed, 11 Nov 2026 15:07:39 +0000</pubDate><link>https://news.ycombinator.com/item?id=50454334</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50454334</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50454334</guid></item><item><title><![CDATA[New comment by barrell in "Apple has new Mac app coming to make running local AI models easy"]]></title><description><![CDATA[
<p>Maybe I’m not the largest consumer of apps, but the only time I use an app is if I want to interface with a service or with other people. Apps are just the man in the middle.<p>Do I want to set up my scale to sync its data to my vibe coded app? Not really. Do I want to set up the dozens of visualizations and hundreds of different actions and thousands of different items for calorie tracking, where even a 5% error is extremely significant? No way. Am I going to convince any of my friends to download my vibe coded chat application? Absolutely not. Do I want to store my photos in my own vibe coded solution and trust it to run in 20 years? Not a chance.<p>While a segment of apps will surely become obsolete, I don’t think the majority of users time will be spent on their own vibe coded replacements. Even as a developer I have no interest in turning all my conveniences into maintenance burdens, with risk of catastrophic failure</p>
]]></description><pubDate>Wed, 11 Nov 2026 14:47:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=50454109</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50454109</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50454109</guid></item><item><title><![CDATA[New comment by barrell in "Person of Interest"]]></title><description><![CDATA[
<p>Yeah I can’t remember when it got good but I do remember it just kept getting better and better.<p>I still think about it so often. I haven’t rewatched it because I still remember so much if it given how often I’ve thought about it - I’m waiting to forget forget.</p>
]]></description><pubDate>Sun, 08 Nov 2026 15:25:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=50414066</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50414066</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50414066</guid></item><item><title><![CDATA[New comment by barrell in "Person of Interest"]]></title><description><![CDATA[
<p>I first saw an episode a few seasons in, I can’t remember which. It came on in the background while at a friends house and it kept distracting me. I didn’t catch the name of the show and spent years looking for it.<p>I actually stumbled upon Person of Interest a couple times, but the first couple episodes were so unlike what I saw I figured it was a different series.<p>Eventually I realized it was the same show, and was able to watch the first couple seasons with baited breath trying to figure out how it became the show in the episode that I glimpsed before… and if it even was as good as I thought it was years prior.<p>So maybe watch a random episode in the fourth season of you don’t like season one? XD<p>edit: it was even better than I remember. It became one of my favorite series, probably the one I still think about the most to this day</p>
]]></description><pubDate>Sun, 08 Nov 2026 14:40:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=50413752</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50413752</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50413752</guid></item><item><title><![CDATA[New comment by barrell in "Apple’s grey list: Words and phrases your iPhone won’t help you spell"]]></title><description><![CDATA[
<p>Tried this myself. Got 10/10 on propaganda, 0/10 on left wing propaganda, 10/10 on right wring propaganda.<p>I think that’s a bigger issue regardless of your political stance.</p>
]]></description><pubDate>Sun, 08 Nov 2026 09:33:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=50411708</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50411708</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50411708</guid></item><item><title><![CDATA[New comment by barrell in "The Nordic Token Tax"]]></title><description><![CDATA[
<p>I’d be curious to see this ran through cohere’s tokenizer - they seem to be the largest lab with a focus on multilingualism.</p>
]]></description><pubDate>Tue, 03 Nov 2026 18:17:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=50355867</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50355867</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50355867</guid></item><item><title><![CDATA[New comment by barrell in "We're releasing the weights of Mistral Large 4"]]></title><description><![CDATA[
<p>I’m not sure it speaks so poorly on the company. I’m not sure whether they’re using GLM 5.3 or 5.3 flash. The former is twice the price (probably better results?) and the latter is a quarter the price.<p>If they don’t end up switching models by the end of the year, I think that says a lot about the models. If they’re just waiting for an appropriately large or small model, or waiting to get large 4 put through its paces before switching, I think that’s a pretty ego-less way to go about it, and could speak pretty highly in a way.<p>Mistral large 3/medium 3.5/small 4 have all been pretty dated for a while so if they just needed a stopgap, I could understand. Now that large 4 is out though, we will see if that was the case.</p>
]]></description><pubDate>Fri, 30 Oct 2026 12:14:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=50302874</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50302874</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50302874</guid></item><item><title><![CDATA[New comment by barrell in "We're releasing the weights of Mistral Large 4"]]></title><description><![CDATA[
<p>I didn’t know mistral started serving deepseek-flash 4.1! This is awesome. A little more expensive than most providers though. $0.03/$0.007/$1.20 on mistral vs deepseeks own peak pricing of $0.03/$0.006/$1.20 (vs deepseeks off peak pricing of $0.015/$0.003/$0.60).<p>I prefer using a European provider though so this is actually cheaper than I can find elsewhere.</p>
]]></description><pubDate>Fri, 30 Oct 2026 08:01:51 +0000</pubDate><link>https://news.ycombinator.com/item?id=50300987</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50300987</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50300987</guid></item><item><title><![CDATA[New comment by barrell in "We're releasing the weights of Mistral Large 4"]]></title><description><![CDATA[
<p>I also found this odd. After mistral announced v4 preview, I just assumed it was what would be used in Le Chat. I gave it a try and was impressed.<p>Turns out they’re using GLM in le chat. Quite concerning if I’m being honest, and this is coming from a huge mistral fanboy.<p>I wish we could toggle mistral large 4 in le chat so I could get a feel for it.</p>
]]></description><pubDate>Fri, 30 Oct 2026 07:56:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=50300958</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50300958</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50300958</guid></item><item><title><![CDATA[New comment by barrell in "We're releasing the weights of Mistral Large 4"]]></title><description><![CDATA[
<p>Not sure it’s particularly odd. Several models of late have launched with an introductory promotional pricing, only to later lock in the promotional pricing.<p>Too many models now to remember specifics - I believe deepseek v4 did this and one of the gpt models (5? 5.1? 6?) did this if I recall correctly. I’m also fairly confident Google has had a few promotional periods on launch of the model, I seem to remember them starting the trend.</p>
]]></description><pubDate>Fri, 30 Oct 2026 07:54:07 +0000</pubDate><link>https://news.ycombinator.com/item?id=50300952</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50300952</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50300952</guid></item><item><title><![CDATA[New comment by barrell in "Voxtral 3"]]></title><description><![CDATA[
<p>I’ve been using mistral-large-3 a tremendous amount. Price/speed/quality ratio was the best I could find. Been using the mistral-large-4 now since it came out for a few things, and while it’s not price competitive for the quality, it is really good. I’ve replaced a lot of my browser-based-chat-ui-llm-usage with mistral since it launched.</p>
]]></description><pubDate>Tue, 20 Oct 2026 21:33:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=50172990</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50172990</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50172990</guid></item><item><title><![CDATA[New comment by barrell in "DeepSeek-v4.1-Flash-1017"]]></title><description><![CDATA[
<p>To me this implies they continued the training run for 4.1 and it improved. A 4.2 would imply they changed something about how it’s trained/how inference works, and a 5.0 would be a major update (who knows what would have changed).<p>This is just my headcanon though.</p>
]]></description><pubDate>Sat, 17 Oct 2026 14:42:55 +0000</pubDate><link>https://news.ycombinator.com/item?id=50131618</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=50131618</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50131618</guid></item><item><title><![CDATA[New comment by barrell in "Sharing AI progress in mathematics"]]></title><description><![CDATA[
<p>The letter counting issue <i>is</i> due to tokenization. And most models still get this wrong often enough, even with reasoning. Probably less so on strawberry given how prevalent it is, and less so than without reasoning, but this not a historical issue. It’s becoming less of one though.</p>
]]></description><pubDate>Wed, 07 Oct 2026 11:03:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=49991032</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=49991032</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49991032</guid></item><item><title><![CDATA[New comment by barrell in "Mistral Large 4"]]></title><description><![CDATA[
<p>That is the promotional price, after which it doubles :(</p>
]]></description><pubDate>Tue, 06 Oct 2026 14:52:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=49979391</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=49979391</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49979391</guid></item><item><title><![CDATA[New comment by barrell in "Dots: Always-on agents"]]></title><description><![CDATA[
<p>I have been waiting for the next generation of mistral models for so long now. They were supposed to come out this summer… I hope the delay is because they’re making something better, not just because they’ve fallen too far behind.<p>Mistral models were the best for my use case when they came out, but they’re mostly almost a year old now. Hard to keep justifying using them, especially with all the price cuts this year on other models.</p>
]]></description><pubDate>Wed, 30 Sep 2026 06:07:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=49905007</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=49905007</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49905007</guid></item><item><title><![CDATA[New comment by barrell in "When did Google get so weird?"]]></title><description><![CDATA[
<p>> are you ready to just walk away<p>Many comments, but I'm surprised no one has mentioned the ease and quickness LLMs have to advise you to leave relationships. Doesn't matter the context, how little you tell them, or how close the person is, an LLM is always ready to drop them at the tip of a hat.<p>It's weird enough chatbots try to be your friend. Doubly weird that the search engine tries to be your friend. The downright dystopian part is how often they tell you to leave your real friends.</p>
]]></description><pubDate>Mon, 28 Sep 2026 04:58:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=49873836</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=49873836</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49873836</guid></item><item><title><![CDATA[New comment by barrell in "What About Rails?"]]></title><description><![CDATA[
<p>I’m building a language learning application [1]. It uses spaced repetition, built upon FSRS, an open source algorithm.<p>I’ve spent the last three years tweaking away at this algorithm. There are so many little details that have had to come together to make it an enjoyable experience. For example, take the one problem of balancing multiple language: What is the best way to balance multiple languages? How should you switch between them, and how often? What happens if you are over performing in one language and underperforming in another? What happens if you have more reviews in one language than another that isn’t in line with your priorities? How do you deal with competing learning speeds and language priorities? Spiky review burdens? What happens when you change your language priorities?<p>This is probably the most trivial part of the algorithm, and it took years of trial and error and dozens of iterations to get it to feel right. It requires knowing about SRS, the FSRS algorithm, language learning principles, and even then tons of experimentation, trial and error, and hundreds of days of daily usage.<p>I’ve asked LLMs every step of the way what to do. I’ve probably taken <1% of the advice I was given. Just asking an LLM “build an SRS app that balances multiple languages” would produce a result, but one that would definitely be terrible to use, and probably would end up slowing you down, not speeding you up. Explaining to it any of the issues that you would come across when using it will likely suggest to changes that will only make things worse, based on the thousands of terrible ideas I’ve seen confidently produced.<p>The hard part of a product is not the idea, but the experience of using it. The valueable work is normally not in the big feature specs, but the hundreds of little paper cuts that have all been smoothed over. The time saved is not because you don’t have to code it, but is from not having to become an expert in several different fields.<p>[1] <a href="https://phrasing.app" rel="nofollow">https://phrasing.app</a></p>
]]></description><pubDate>Fri, 25 Sep 2026 17:14:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=49847279</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=49847279</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49847279</guid></item><item><title><![CDATA[New comment by barrell in "I am done with this shit"]]></title><description><![CDATA[
<p>I agree artisan can be a bit pretentious, but calling someone an artisan != calling someone pretentious, whereas calling someone a craftsman (to some people) == calling someone blue collar, someone who works in the crafts/went to craft school.</p>
]]></description><pubDate>Thu, 24 Sep 2026 04:35:58 +0000</pubDate><link>https://news.ycombinator.com/item?id=49826298</link><dc:creator>barrell</dc:creator><comments>https://news.ycombinator.com/item?id=49826298</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49826298</guid></item></channel></rss>