<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: liamlaverty</title><link>https://news.ycombinator.com/user?id=liamlaverty</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sun, 21 Jun 2026 13:04:33 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=liamlaverty" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by liamlaverty in "Rich Sutton on AI creativity and discovery"]]></title><description><![CDATA[
<p>>But there's no reason you couldn't add an Artistic Director layer which has been trained on emotional and cultural signifiers and to direct the pastiche and remixing.<p>I gave this approach a shot over the first few months of this year[1] (although my director didn't have any custom training). The results were interesting, but I'd not call them "art", since they're low-quality derivative pieces. With reasoning traces enabled, you can see that there's not much intent going on. Though they do attempt to include "incidental objects" to reinforce meaning, like in this jungle scene[2].<p>[1] <a href="https://news.ycombinator.com/item?id=48105385">https://news.ycombinator.com/item?id=48105385</a><p>[2]<a href="https://www.liamlaverty.com/paint-by-language-model/inspect/patchwork-elephant-line-arc-polyline-circle-claude-opus-4-6" rel="nofollow">https://www.liamlaverty.com/paint-by-language-model/inspect/...</a></p>
]]></description><pubDate>Wed, 10 Jun 2026 11:08:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=48474532</link><dc:creator>liamlaverty</dc:creator><comments>https://news.ycombinator.com/item?id=48474532</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48474532</guid></item><item><title><![CDATA[New comment by liamlaverty in "Finding Security Bugs in OSS with LLMs on a Budget"]]></title><description><![CDATA[
<p>Anthropic have been using their Claude 4.5 & 4.6 series models to automate searching for security issues. Their approach seems pretty expensive, so I've demonstrated getting a lot of the same benefits on a much lower budget in the Umbraco-CMS repository.</p>
]]></description><pubDate>Mon, 25 May 2026 08:14:04 +0000</pubDate><link>https://news.ycombinator.com/item?id=48264592</link><dc:creator>liamlaverty</dc:creator><comments>https://news.ycombinator.com/item?id=48264592</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48264592</guid></item><item><title><![CDATA[Finding Security Bugs in OSS with LLMs on a Budget]]></title><description><![CDATA[
<p>Article URL: <a href="https://www.etive-mor.com/blog/carlini-style-vulnerability-hunting-on-a-budget/">https://www.etive-mor.com/blog/carlini-style-vulnerability-hunting-on-a-budget/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48264591">https://news.ycombinator.com/item?id=48264591</a></p>
<p>Points: 2</p>
<p># Comments: 1</p>
]]></description><pubDate>Mon, 25 May 2026 08:14:04 +0000</pubDate><link>https://www.etive-mor.com/blog/carlini-style-vulnerability-hunting-on-a-budget/</link><dc:creator>liamlaverty</dc:creator><comments>https://news.ycombinator.com/item?id=48264591</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48264591</guid></item><item><title><![CDATA[New comment by liamlaverty in "Can a Language Model Paint?"]]></title><description><![CDATA[
<p>I've been trying to get some language models to paint one stroke at a time for a few months now. I thought this community would be interested to see the results.<p>The article runs through my findings, and there's a linked technical rundown of how the app was built. There's also an interactive gallery [0] of my attempts. You can point an agent at the API docs [1], and they might (ymmv) do a painting themselves.<p>[0] <a href="https://www.liamlaverty.com/paint-by-language-model/" rel="nofollow">https://www.liamlaverty.com/paint-by-language-model/</a>
[1] <a href="https://www.liamlaverty.com/paint-by-language-model/draw/api" rel="nofollow">https://www.liamlaverty.com/paint-by-language-model/draw/api</a></p>
]]></description><pubDate>Tue, 12 May 2026 07:47:37 +0000</pubDate><link>https://news.ycombinator.com/item?id=48105405</link><dc:creator>liamlaverty</dc:creator><comments>https://news.ycombinator.com/item?id=48105405</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48105405</guid></item><item><title><![CDATA[Can a Language Model Paint?]]></title><description><![CDATA[
<p>Article URL: <a href="https://www.etive-mor.com/blog/can-a-language-model-paint/">https://www.etive-mor.com/blog/can-a-language-model-paint/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48105385">https://news.ycombinator.com/item?id=48105385</a></p>
<p>Points: 27</p>
<p># Comments: 8</p>
]]></description><pubDate>Tue, 12 May 2026 07:44:19 +0000</pubDate><link>https://www.etive-mor.com/blog/can-a-language-model-paint/</link><dc:creator>liamlaverty</dc:creator><comments>https://news.ycombinator.com/item?id=48105385</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48105385</guid></item></channel></rss>