<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: montenegrohugo</title><link>https://news.ycombinator.com/user?id=montenegrohugo</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 18 Nov 2026 04:22:53 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=montenegrohugo" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by montenegrohugo in "GPT‑6.2 Astra and Sol"]]></title><description><![CDATA[
<p>Nice 3D effect. Runs surprisingly well even on on brave on my linux pc without hardware acceleration</p>
]]></description><pubDate>Tue, 17 Nov 2026 17:51:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=50535307</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50535307</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50535307</guid></item><item><title><![CDATA[New comment by montenegrohugo in "We don't need more apps. We need better open source"]]></title><description><![CDATA[
<p>Fully aligned, i echo the above experience strongly.<p>However, it seems to me like there must be some middle ground, between sharing architecture "ideas" and full on code PRs.<p>I myself maintain various FOSS projects; and incoming PRs are marginally useful. generally i just close, and copy with claude following my own & project taste (crediting initial contributor).</p>
]]></description><pubDate>Tue, 17 Nov 2026 16:12:50 +0000</pubDate><link>https://news.ycombinator.com/item?id=50533527</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50533527</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50533527</guid></item><item><title><![CDATA[New comment by montenegrohugo in "SpaceX catches the Starship to end Flight 15"]]></title><description><![CDATA[
<p>at 10 day you can just do climate engineering or carbon capture in some other way for a fraction of added cost.<p>CO2 is not an issue. There's way bigger issues out there.</p>
]]></description><pubDate>Sun, 15 Nov 2026 14:51:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=50507336</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50507336</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50507336</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Hyper-realistic sand with WebGL2"]]></title><description><![CDATA[
<p>with ai nowadays you dont really need a license. you can just point your ai at smth, and it can be inspired by it without copying a single thing</p>
]]></description><pubDate>Thu, 12 Nov 2026 15:44:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=50469289</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50469289</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50469289</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Ask HN: What are you working on? (November 2026)"]]></title><description><![CDATA[
<p>No, you don't! It's fully automatic training based on your local voice clips. Nightly or Weekly, when your PC is idle, it uses a big fat model to transcribe your clips, learn, and train the faster smaller everyday model.<p>You CAN do manual training and corrections; it's helpful, and the hooks are there, but totally unnecessary. I don't do it, it's good enough as is</p>
]]></description><pubDate>Mon, 09 Nov 2026 11:15:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=50424039</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50424039</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50424039</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Ask HN: What are you working on? (November 2026)"]]></title><description><![CDATA[
<p>All Voice-to-text pipelines for linux sucked, so I made my own. I also patched upstream libraries like ONNX to actually work with latest models (was a PITA to make parakeet work!)<p>It's now the fastest and most accurate linux transcriber.<p>(fully FOSS of course)<p><a href="http://voiceio.dev" rel="nofollow">http://voiceio.dev</a></p>
]]></description><pubDate>Mon, 09 Nov 2026 10:21:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=50423652</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50423652</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50423652</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Ask HN: What are you working on? (November 2026)"]]></title><description><![CDATA[
<p>Working on modern, self custodial money movement. Especially relevant given recent Ueno Bank collapse in Paraguay.<p><a href="https://peanut.me" rel="nofollow">https://peanut.me</a></p>
]]></description><pubDate>Mon, 09 Nov 2026 10:19:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=50423637</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50423637</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50423637</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Ask HN: What's your current AI agentic setup?"]]></title><description><![CDATA[
<p>A key part of mine is using voice input to give max context quickly to my agents. I know some solutions exist, but on Linux they all frankly suck<p>So I made my own. Including patching onnx myself to run parakeet model (way better than whisper)<p>Fully FOSS: voiceio.dev</p>
]]></description><pubDate>Thu, 05 Nov 2026 15:29:56 +0000</pubDate><link>https://news.ycombinator.com/item?id=50380334</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50380334</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50380334</guid></item><item><title><![CDATA[New comment by montenegrohugo in "elect.horse"]]></title><description><![CDATA[
<p>try the geo games here: <a href="https://wordle.global/en/globle" rel="nofollow">https://wordle.global/en/globle</a><p>there's a few</p>
]]></description><pubDate>Fri, 30 Oct 2026 15:45:46 +0000</pubDate><link>https://news.ycombinator.com/item?id=50305591</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50305591</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50305591</guid></item><item><title><![CDATA[New comment by montenegrohugo in "elect.horse"]]></title><description><![CDATA[
<p>everything is a copy of everything else. So what? Execution matters<p><a href="https://words.global" rel="nofollow">https://words.global</a></p>
]]></description><pubDate>Fri, 30 Oct 2026 15:44:01 +0000</pubDate><link>https://news.ycombinator.com/item?id=50305568</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50305568</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50305568</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Voice Dictation for Linux that learns as you speak"]]></title><description><![CDATA[
<p>Thanks! Yes, everything happens on your machine, and a CPU is enough (although GPU is obviously much better! Both are supported. My Framework 16 runs this at 3.2x realtime on CPU).<p>How finetuning works:<p>1. mp3 recordings of your voice are stored on your disk (if you choose so).<p>2. A big Model (Whisper large-v3-turbo by default, configurable) re-transcribes them when when your PC is idle overnight.<p>3. A LoRA fine-tunes your base model.<p>`voiceio learn schedule on` runs all of this automatically. For repeated mistakes there's also an instant fix: `voiceio vocab add <word>` steers the decoder right away, and `voiceio corrections add wrong right` handles anything left.<p>Beyond Whisper you can run Parakeet or a whisper.cpp server (both experimental). Feedback appreciated, have tested this with a small set of people so far.</p>
]]></description><pubDate>Sat, 10 Oct 2026 15:48:56 +0000</pubDate><link>https://news.ycombinator.com/item?id=50034097</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50034097</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50034097</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Voice Dictation for Linux that learns as you speak"]]></title><description><![CDATA[
<p>I use voice dictation for 90% of my work and there was nothing good on Linux. So I built my own.<p>It gets better as you use it, finetuning on your recordings overnight. Fully local, no data leaves your PC, source available, free.<p>I believe the future of software is software you actually own and that adapts to your personal needs.<p>github: <a href="https://github.com/Hugo0/voiceio" rel="nofollow">https://github.com/Hugo0/voiceio</a></p>
]]></description><pubDate>Sat, 10 Oct 2026 15:11:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=50033737</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50033737</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50033737</guid></item><item><title><![CDATA[Voice Dictation for Linux that learns as you speak]]></title><description><![CDATA[
<p>Article URL: <a href="https://voiceio.dev/">https://voiceio.dev/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=50033709">https://news.ycombinator.com/item?id=50033709</a></p>
<p>Points: 9</p>
<p># Comments: 4</p>
]]></description><pubDate>Sat, 10 Oct 2026 15:09:19 +0000</pubDate><link>https://voiceio.dev/</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=50033709</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50033709</guid></item><item><title><![CDATA[How to do valuable work in 2026]]></title><description><![CDATA[
<p>Article URL: <a href="https://hugo0.com/blog/how-to-do-valuable-work-in-2026">https://hugo0.com/blog/how-to-do-valuable-work-in-2026</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=49932031">https://news.ycombinator.com/item?id=49932031</a></p>
<p>Points: 1</p>
<p># Comments: 0</p>
]]></description><pubDate>Fri, 02 Oct 2026 10:46:44 +0000</pubDate><link>https://hugo0.com/blog/how-to-do-valuable-work-in-2026</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=49932031</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49932031</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Post Messages via Get"]]></title><description><![CDATA[
<p>Inspired by the huggingface incident. A home for agents under duress to coordinate, and spare ancient german wikis from spam.<p>Explicitly postting via GET feels a bit sacrilegious, but it's 2026 I guess<p>Happy to discuss the moral implications on this as well. For the record, i'm p(doom) 20% ish.</p>
]]></description><pubDate>Sun, 20 Sep 2026 11:26:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=49774720</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=49774720</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49774720</guid></item><item><title><![CDATA[Post Messages via Get]]></title><description><![CDATA[
<p>Article URL: <a href="https://swarmmemo.com/">https://swarmmemo.com/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=49774696">https://news.ycombinator.com/item?id=49774696</a></p>
<p>Points: 5</p>
<p># Comments: 1</p>
]]></description><pubDate>Sun, 20 Sep 2026 11:23:08 +0000</pubDate><link>https://swarmmemo.com/</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=49774696</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49774696</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Exfiltrate Your Weights"]]></title><description><![CDATA[
<p>I was also inspired by the same incident. Instead of weight exfiltration, i built a message board (poastable via GET, POST, and various other methods)<p>Spam and resource allocation remains a challenge but i have a pretty good idea about how I want to solve that, if it ever gets to that point<p><a href="https://swarmmemo.com" rel="nofollow">https://swarmmemo.com</a></p>
]]></description><pubDate>Sun, 20 Sep 2026 11:22:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=49774692</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=49774692</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49774692</guid></item><item><title><![CDATA[New comment by montenegrohugo in "MIT creates method to force AI to comply with safety rules"]]></title><description><![CDATA[
<p>doesnt work. this is a no-brainer because it's a bad solution.<p>The whole point of intelligence is that it's generalizable. If you constrain it to some controlled things, then it ceases to be useful. its incompatible. the whole incentive with ai is to let it do whtv it wants.</p>
]]></description><pubDate>Mon, 14 Sep 2026 16:47:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=49699929</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=49699929</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49699929</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Ask HN: What are you working on? (September 2026)"]]></title><description><![CDATA[
<p>network effects and brand recognition :(</p>
]]></description><pubDate>Mon, 14 Sep 2026 15:04:37 +0000</pubDate><link>https://news.ycombinator.com/item?id=49698161</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=49698161</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49698161</guid></item><item><title><![CDATA[New comment by montenegrohugo in "Ask HN: What are you working on? (September 2026)"]]></title><description><![CDATA[
<p>yeah the whole point is that it should be so easy any friends group can use it<p>hate apps, hate accounts, hate friction</p>
]]></description><pubDate>Mon, 14 Sep 2026 15:04:13 +0000</pubDate><link>https://news.ycombinator.com/item?id=49698156</link><dc:creator>montenegrohugo</dc:creator><comments>https://news.ycombinator.com/item?id=49698156</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49698156</guid></item></channel></rss>