<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: ykumards</title><link>https://news.ycombinator.com/user?id=ykumards</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Fri, 24 Jul 2026 03:14:11 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=ykumards" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[J-Space Oddity: Do VLMs Dream of Text Tokens?]]></title><description><![CDATA[
<p>Article URL: <a href="https://ykumar.me/blog/j-space-oddity/">https://ykumar.me/blog/j-space-oddity/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48995560">https://news.ycombinator.com/item?id=48995560</a></p>
<p>Points: 2</p>
<p># Comments: 0</p>
]]></description><pubDate>Tue, 21 Jul 2026 17:39:12 +0000</pubDate><link>https://ykumar.me/blog/j-space-oddity/</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=48995560</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48995560</guid></item><item><title><![CDATA[J-Space Oddity: Do VLMs Dream of Text Tokens?]]></title><description><![CDATA[
<p>Article URL: <a href="https://ykumar.me/blog/j-space-oddity/">https://ykumar.me/blog/j-space-oddity/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48897751">https://news.ycombinator.com/item?id=48897751</a></p>
<p>Points: 5</p>
<p># Comments: 0</p>
]]></description><pubDate>Mon, 13 Jul 2026 19:41:40 +0000</pubDate><link>https://ykumar.me/blog/j-space-oddity/</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=48897751</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48897751</guid></item><item><title><![CDATA[New comment by ykumards in "Autoresearch on an old research idea"]]></title><description><![CDATA[
<p>IMO it would be hard to reproduce the results using autoresearch setup.<p>To get CLIP to work properly we typically need large batch sizes. So the experiments in the original paper were quite heavy, and ran parallel across 8 GPUs.</p>
]]></description><pubDate>Tue, 24 Mar 2026 04:38:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=47498645</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=47498645</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47498645</guid></item><item><title><![CDATA[New comment by ykumards in "Autoresearch on an old research idea"]]></title><description><![CDATA[
<p>Sorry I could have worded this part better.<p>The docker container didn’t have network access. Claude didn’t have permission to execute anything other than the run.sh bash script, which would orchestrate the docker run</p>
]]></description><pubDate>Mon, 23 Mar 2026 22:21:17 +0000</pubDate><link>https://news.ycombinator.com/item?id=47495938</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=47495938</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47495938</guid></item><item><title><![CDATA[New comment by ykumards in "Autoresearch on an old research idea"]]></title><description><![CDATA[
<p>That’s true!
It felt a bit flippant to give medical data to an agent. Also, I wanted to see if the model would work in other domains!</p>
]]></description><pubDate>Mon, 23 Mar 2026 22:19:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=47495921</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=47495921</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47495921</guid></item><item><title><![CDATA[New comment by ykumards in "Autoresearch on an old research idea"]]></title><description><![CDATA[
<p>Thank you, glad you liked it!</p>
]]></description><pubDate>Mon, 23 Mar 2026 22:14:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=47495882</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=47495882</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47495882</guid></item><item><title><![CDATA[Autoresearch on an old research idea]]></title><description><![CDATA[
<p>Article URL: <a href="https://ykumar.me/blog/eclip-autoresearch/">https://ykumar.me/blog/eclip-autoresearch/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47493460">https://news.ycombinator.com/item?id=47493460</a></p>
<p>Points: 428</p>
<p># Comments: 95</p>
]]></description><pubDate>Mon, 23 Mar 2026 18:40:58 +0000</pubDate><link>https://ykumar.me/blog/eclip-autoresearch/</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=47493460</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47493460</guid></item><item><title><![CDATA[New comment by ykumards in "Show HN: Daily podcast of HuggingFace's Top voted papers"]]></title><description><![CDATA[
<p>I'm subscribed to Huggingface's Daily Papers newsletter but going through each paper individually was a chore.<p>So I created a GH workflow to pull the papers, use an LLM and TTS to generate the podcast daily.<p>Hope you like it!</p>
]]></description><pubDate>Sun, 24 Aug 2025 14:07:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=45004347</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=45004347</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45004347</guid></item><item><title><![CDATA[Show HN: Daily podcast of HuggingFace's Top voted papers]]></title><description><![CDATA[
<p>Article URL: <a href="https://pod.artok.app">https://pod.artok.app</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=45004337">https://news.ycombinator.com/item?id=45004337</a></p>
<p>Points: 4</p>
<p># Comments: 1</p>
]]></description><pubDate>Sun, 24 Aug 2025 14:05:49 +0000</pubDate><link>https://pod.artok.app</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=45004337</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45004337</guid></item><item><title><![CDATA[New comment by ykumards in "Show HN: Clauditate – Breathing exercises while Claude Code is thinking"]]></title><description><![CDATA[
<p>True :)<p><a href="https://xkcd.com/303/" rel="nofollow">https://xkcd.com/303/</a></p>
]]></description><pubDate>Thu, 31 Jul 2025 15:29:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=44746676</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=44746676</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44746676</guid></item><item><title><![CDATA[Show HN: Clauditate – Breathing exercises while Claude Code is thinking]]></title><description><![CDATA[
<p>Hello HN!<p>I realized how often my mind wanders while waiting for Claude to finish "thinking". So I built Clauditate, a lightweight meditation menubar app that helps you with some breathing exercises.<p>It uses Claude Code hooks and pops up with a breathing guide. I added some heuristics to make it less "annoying", for example if you dismiss it twice within 30 mins, it won't show up for another 2 hours. You can also tweak this behavior (and more) in the settings.<p>It's early, but I'd love to hear your feedback!<p>Github: <a href="https://github.com/ykumards/clauditate">https://github.com/ykumards/clauditate</a></p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=44745279">https://news.ycombinator.com/item?id=44745279</a></p>
<p>Points: 6</p>
<p># Comments: 2</p>
]]></description><pubDate>Thu, 31 Jul 2025 13:10:27 +0000</pubDate><link>https://github.com/ykumards/clauditate</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=44745279</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=44745279</guid></item><item><title><![CDATA[New comment by ykumards in "Show HN: ArTok is TikTok for research papers"]]></title><description><![CDATA[
<p>Thank you so much! :)<p>I’ll definitely try to add audio. Local TTS models are getting better and I guess I can feed the headline to create a podcast-like feed</p>
]]></description><pubDate>Tue, 22 Apr 2025 04:21:20 +0000</pubDate><link>https://news.ycombinator.com/item?id=43759047</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=43759047</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43759047</guid></item><item><title><![CDATA[New comment by ykumards in "Show HN: ArTok is TikTok for research papers"]]></title><description><![CDATA[
<p>That’s a great point! I could add some “roles” that people can choose for the kind of papers they’d like to see.</p>
]]></description><pubDate>Tue, 22 Apr 2025 04:18:17 +0000</pubDate><link>https://news.ycombinator.com/item?id=43759032</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=43759032</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43759032</guid></item><item><title><![CDATA[Show HN: ArTok is TikTok for research papers]]></title><description><![CDATA[
<p>Hello everyone!<p>I always found it hard to find new research papers outside of my usual bubble. I thought a random feed (with no recommendation algorithms) might be a fun way to explore. But I also didn’t want to waste time on completely unrelated stuff — so the idea of a fast, swipeable format came to mind.<p>Wikitok was a real inspiration! But Arxiv and Open Review APIs weren’t as robust as Wikipedia, so I pulled the papers into a Postgres backend.<p>Right now, I’ve indexed papers from a few recent ML conferences to see if this might be useful for others too.<p>No signups required and it’s totally free. You can mark your favorites and add text annotations, which are saved on your device.<p>Would love to hear your feedback!</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=43755405">https://news.ycombinator.com/item?id=43755405</a></p>
<p>Points: 8</p>
<p># Comments: 5</p>
]]></description><pubDate>Mon, 21 Apr 2025 19:13:51 +0000</pubDate><link>https://artok.app</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=43755405</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43755405</guid></item><item><title><![CDATA[Steinbeck used the diary as a tool of discipline (2015)]]></title><description><![CDATA[
<p>Article URL: <a href="https://www.themarginalian.org/2015/03/02/john-steinbeck-working-days/">https://www.themarginalian.org/2015/03/02/john-steinbeck-working-days/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=37723310">https://news.ycombinator.com/item?id=37723310</a></p>
<p>Points: 3</p>
<p># Comments: 0</p>
]]></description><pubDate>Sun, 01 Oct 2023 06:32:22 +0000</pubDate><link>https://www.themarginalian.org/2015/03/02/john-steinbeck-working-days/</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=37723310</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=37723310</guid></item><item><title><![CDATA[New comment by ykumards in "The dots do matter: how to scam a Gmail user"]]></title><description><![CDATA[
<p>That was exactly my reaction after reading the first few paragraphs. “How did Eve circumvent the email verification step?”<p>Netflix should totally fix this, even my local public library asks for email verification after signing up.</p>
]]></description><pubDate>Mon, 09 Apr 2018 06:04:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=16790757</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=16790757</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=16790757</guid></item><item><title><![CDATA[New comment by ykumards in "Get your loved ones off Facebook"]]></title><description><![CDATA[
<p>That's totally fine, as long as you are aware of the problem and decide that it doesn't matter to you. 
But they have the moral(legal?) obligation to let us know how they're using our data.</p>
]]></description><pubDate>Mon, 22 May 2017 19:21:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=14396207</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=14396207</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=14396207</guid></item><item><title><![CDATA[New comment by ykumards in "Get your loved ones off Facebook"]]></title><description><![CDATA[
<p>I share your concerns about losing Facebook data of the past, but frankly you didn't address any of the concerns that were listed in the article.<p>The issue is what they do with our data and whether you'd be comfortable letting them use your data this way. It's also due to their shady practices like not disclosing changes to privacy policy. This is especially worse with non-technical people like my mom using their platform. Who knows what kind of bs they're peddling onto her wall.<p>I think it's awesome that people are finding ways to get off Facebook. It's like getting out of a heroine addiction or something:)</p>
]]></description><pubDate>Mon, 22 May 2017 17:14:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=14395279</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=14395279</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=14395279</guid></item><item><title><![CDATA[New comment by ykumards in "Ask HN: What are the most useful Chrome extensions in 2016?"]]></title><description><![CDATA[
<p>* uBlock Origin - Lightweight adblocker<p>Link: <a href="https://chrome.google.com/webstore/detail/ublock-origin/cjpalhdlnbpafiamejdnhcphjbkeiagm" rel="nofollow">https://chrome.google.com/webstore/detail/ublock-origin/cjpa...</a><p>* Tab Snooze - Removes clutter by hiding a tab for later. We can pick the time & date. Can use this as a reminder.<p>Link: <a href="https://chrome.google.com/webstore/detail/tab-snooze/pdiebiamhaleloakpcgmpnenggpjbcbm" rel="nofollow">https://chrome.google.com/webstore/detail/tab-snooze/pdiebia...</a><p>* One Tab - Push all current tabs into one tab. I use this with Tab Snooze, mostly. All tabs related to a project are snoozed till I get to work on them again.<p>Link: <a href="https://chrome.google.com/webstore/detail/onetab/chphlpgkkbolifaimnlloiipkdnihall" rel="nofollow">https://chrome.google.com/webstore/detail/onetab/chphlpgkkbo...</a><p>* Papier/Workflowy New Tab - Have a text document in the new tab page. I usually have to-dos or jot down some quick notes here.<p>Link1: <a href="https://chrome.google.com/webstore/detail/papier/hhjeaokafplhjoogdemakihhdhffacia" rel="nofollow">https://chrome.google.com/webstore/detail/papier/hhjeaokafpl...</a><p>Link2 : <a href="https://chrome.google.com/webstore/detail/workflowy-new-tab/nfmnfkpjdcgodkcdjhhgfefpjbcnkfhi" rel="nofollow">https://chrome.google.com/webstore/detail/workflowy-new-tab/...</a></p>
]]></description><pubDate>Sun, 01 Jan 2017 16:25:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=13296175</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=13296175</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=13296175</guid></item><item><title><![CDATA[New comment by ykumards in "Ask HN: What are the most useful Chrome extensions in 2016?"]]></title><description><![CDATA[
<p>If you have a workflowy account, this syncs well and is similar to Papier.<p><a href="https://chrome.google.com/webstore/detail/workflowy-new-tab/nfmnfkpjdcgodkcdjhhgfefpjbcnkfhi?hl=en" rel="nofollow">https://chrome.google.com/webstore/detail/workflowy-new-tab/...</a></p>
]]></description><pubDate>Sun, 01 Jan 2017 16:11:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=13296133</link><dc:creator>ykumards</dc:creator><comments>https://news.ycombinator.com/item?id=13296133</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=13296133</guid></item></channel></rss>