The latest articles on DEV Community by Austin Johnson (@buoygg). https://dev.to/buoygg https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4148051%2Fd5c3c2f1-2aeb-48aa-9d43-70e04a161d32.jpg https://dev.to/buoygg en Austin Johnson Mon, 05 Oct 2026 19:53:20 +0000 https://dev.to/buoygg/jank-fighter-ii-claude-vs-codex-vs-grok-before-and-after-buoy-22mh https://dev.to/buoygg/jank-fighter-ii-claude-vs-codex-vs-grok-before-and-after-buoy-22mh <p>I built <a href="https://buoy.gg/?utm_source=devto&amp;utm_medium=social&amp;utm_campaign=jank-fighter" rel="noopener noreferrer">Buoy</a> — debugging and optimize tooling for React Native, Expo, Flutter, Swift, web, and more. We put three coding models through the same prompt on a janky React Native Skia scene rendered with React Native Web in Chrome, then ran Optimize and measured again.</p> <p> <iframe src="https://www.youtube.com/embed/NTY6YrGmPUs" width="710" height="399"> </iframe> </p> <h2> Results (FPS, before → after) </h2> <ul> <li> <strong>Claude:</strong> 975,800 → 6,123,350 (<strong>6.3×</strong>)</li> <li> <strong>Codex:</strong> 155,450 → 444,100 (<strong>2.9×</strong>)</li> <li> <strong>Grok:</strong> 202,100 → 577,350 (<strong>2.9×</strong>)</li> </ul> <p>Lights held around <strong>55 fps</strong>.</p> <p>Same prompt, same scene, same harness — the gap is what Optimize changed in the app under test.</p> <p>Full write-up with screenshots and method: <a href="https://buoy.gg/blog/ai-models-before-and-after-buoy?utm_source=devto&amp;utm_medium=social&amp;utm_campaign=jank-fighter" rel="noopener noreferrer">AI models before and after Buoy</a></p> <p>Questions welcome — especially if you want details on the Skia/Web setup or how we scored the runs.</p> reactnative ai performance webdev Austin Johnson Mon, 28 Sep 2026 21:26:50 +0000 https://dev.to/buoygg/introducing-buoygg-in-app-tools-for-mobile-web-apps-5do5 https://dev.to/buoygg/introducing-buoygg-in-app-tools-for-mobile-web-apps-5do5 <p>When something breaks in production, the usual loop is slow: reproduce locally, dig through logs, guess at what the user saw, then ship a fix and hope.</p> <p><strong>Buoy</strong> is built to shrink that loop. It’s a toolkit of in-app tools for mobile and web apps — aimed at people who ship and support products day to day: developers, QA, and support.</p> <h2> What it is </h2> <p>Buoy helps you <strong>see what your app actually did</strong> — not just what you think it did. The idea is simple: put the right diagnostics and tooling <em>inside</em> the product experience so you can understand real sessions faster.</p> <p>We’re building for teams who care about:</p> <ul> <li>Faster debugging when users hit weird states</li> <li>Clearer signal for QA and support without a full engineer dig</li> <li>Better visibility into performance and behavior on real devices</li> </ul> <h2> Who it’s for </h2> <ul> <li> <strong>Indie hackers and small teams</strong> shipping React Native / mobile / web apps</li> <li> <strong>Dev and QA</strong> who need faster feedback loops</li> <li> <strong>Support</strong> who want better context when something goes wrong</li> </ul> <h2> Try it </h2> <p>Website: <a href="https://buoy.gg" rel="noopener noreferrer">https://buoy.gg</a></p> <p>If you’re building mobile or web apps and want fewer “works on my machine” mysteries, take a look and tell us what would help most.</p> <p>We’re building in public — feedback welcome.</p> opensource reactnative mobile webdev