4 stories
Archibald tested avifenc, libjxl, jixel and cwebp on an M4 Pro, scoring with SSIMULACRA2. JPEG XL lost on lossy photos and screenshots (30-90% larger), but dominated lossless and can recompress existing JPEGs 9-20% smaller.
Commenters disputed whether the benchmark or Chrome's earlier decision was the real story.
The top HN threads: facts from the discussions, both sides, insider voices. Email, Telegram, or RSS.
The jump is not a single bug-finding demo: Chrome 149 and 150 shipped 1,072 fixes after the team connected model-assisted discovery to triage, patching, testing, and release work.
The uncomfortable constraint is delivery. A public fix can become an attacker’s roadmap before it reaches users, so Google is also piloting two security releases a week and exploring dynamic patching.
His demos, built with Claude, include a markdown notes app and a node-based video compositor, all editing real local files in-page. The trick is window.showDirectoryPicker(), which lets the user's own cloud-synced folder become the sync layer. The catch the post skips: all three editors of the draft spec are Google engineers, and Jake Archibald showed up to debunk the demo's scariest fear.
Kage drives headless Chrome to render a page the way a human sees it, snapshots the settled DOM, deletes every script, downloads CSS, images and fonts locally, then bundles the mirror into a standalone binary or a ZIM archive.
The first reply pointed to SingleFile, which packs a page into one self-contained HTML file with no server needed. A security researcher caught Chrome launched with --no-sandbox; others called the README LLM slop and assumed the code was too.
