<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Webgpu on MacWorks</title><link>https://macworks.dev/tags/webgpu/</link><description>Recent content in Webgpu on MacWorks</description><generator>Hugo</generator><language>en</language><atom:link href="https://macworks.dev/tags/webgpu/index.xml" rel="self" type="application/rss+xml"/><item><title>2026-06-22</title><link>https://macworks.dev/docs/week/simonwillison/simonwillison-2026-06-22/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://macworks.dev/docs/week/simonwillison/simonwillison-2026-06-22/</guid><description>&lt;h1 id="simon-willison--2026-06-22"&gt;Simon Willison — 2026-06-22&lt;a class="anchor" href="#simon-willison--2026-06-22"&gt;#&lt;/a&gt;&lt;/h1&gt;
&lt;h2 id="highlight"&gt;Highlight&lt;a class="anchor" href="#highlight"&gt;#&lt;/a&gt;&lt;/h2&gt;
&lt;p&gt;Simon&amp;rsquo;s success in porting a PyTorch machine learning model to a browser-based WebGPU application entirely through &amp;ldquo;vibe coding&amp;rdquo; highlights a fascinating shift in developer workflows. It demonstrates how autonomous agents like Claude Code can now bridge significant gaps in domain knowledge, allowing developers to orchestrate the deployment of complex client-side AI tools while actively writing code for entirely different primary projects.&lt;/p&gt;
&lt;h2 id="posts"&gt;Posts&lt;a class="anchor" href="#posts"&gt;#&lt;/a&gt;&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Porting the Moebius 0.2B image inpainting model to run in the browser with Claude Code&lt;/strong&gt; · &lt;a href="https://simonwillison.net/2026/Jun/22/porting-moebius/#atom-everything"&gt;Source&lt;/a&gt;
Simon successfully ported the Moebius 0.2B lightweight image inpainting framework to run locally in the browser, relying purely on &amp;ldquo;vibe coding&amp;rdquo; with Claude Code. While waiting for Codex Desktop to complete mid-sized refactors for a new Datasette table UI, he instructed Claude in a terminal window to convert the original PyTorch model to ONNX, publish the 1.24GB converted weights to Hugging Face, and build a user interface hosted on GitHub Pages. To prevent the application from downloading the massive 1.3GB model on every single page load, he pointed a Claude subagent at a Whisper Web demo to successfully implement browser caching via the CacheStorage API. The core takeaway is the impressive capability of Opus 4.8 to act as a full-stack ML engineer—handling format conversion, model deployment, and front-end development—proving that heavy client-side AI is feasible today if users tolerate the initial download. After completing the project without writing a single line of code himself, Simon used Claude.ai to study his newly generated repository and gain a deeper technical understanding of how ONNX files bundle computation graphs and model weights together.&lt;/p&gt;</description></item><item><title>Simon Willison</title><link>https://macworks.dev/docs/week/simonwillison/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://macworks.dev/docs/week/simonwillison/</guid><description>&lt;h1 id="simon-willison--week-of-2026-06-18-to-2026-06-25"&gt;Simon Willison — Week of 2026-06-18 to 2026-06-25&lt;a class="anchor" href="#simon-willison--week-of-2026-06-18-to-2026-06-25"&gt;#&lt;/a&gt;&lt;/h1&gt;
&lt;h2 id="highlight-of-the-week"&gt;Highlight of the Week&lt;a class="anchor" href="#highlight-of-the-week"&gt;#&lt;/a&gt;&lt;/h2&gt;
&lt;p&gt;This week&amp;rsquo;s absolute standout is the launch of the &lt;code&gt;datasette-apps&lt;/code&gt; plugin, which fundamentally transforms how we build micro-applications over local databases. By utilizing tightly constrained iframe sandboxes and Content-Security-Policy headers, developers and LLMs alike can safely run custom HTML/JS interfaces against a persistent Datasette backend. It brilliantly merges Simon&amp;rsquo;s ongoing experiments with AI-assisted &amp;ldquo;vibe coding&amp;rdquo; and robust security architectures into a core ecosystem feature, effectively bridging the gap between Claude Artifacts and secure data environments.&lt;/p&gt;</description></item></channel></rss>