June 3, 2026
Pixels, pettiness, and payback
Journey to JPEG XL: open-source experiments shaped the future of image coding
Google celebrates JPEG XL’s rise — and the comments instantly turn into a roast
TLDR: Google looked back at the long road to JPEG XL, a newer picture format designed to make web images smaller and better-looking. But readers hijacked the moment with accusations of hypocrisy, complaints about AI-made visuals, and anger over who did—or didn’t—get credit.
Google’s big nostalgia trip about how JPEG XL was built was supposed to be a victory lap for a newer image format that makes pictures look better while taking up less space. Instead, the real fireworks broke out in the comments, where readers treated the post less like a history lesson and more like a public cross-examination. The loudest reaction? A very pointed “That’s rich coming from the company that tried to kill it” from one user, summing up the mood of people who still remember Google stepping back from JPEG XL support in the past. Ouch.
The article itself is a tour through years of Google experiments: improving old JPEGs, building smarter ways to squeeze image files smaller, and eventually helping shape JPEG XL into a serious next-generation format. For non-image nerds: this matters because the web is full of pictures, and smaller files can mean faster sites without making images look ugly. But the community was far less interested in the lab story than in who got credit, who got ignored, and whether the post itself looked suspiciously AI-made.
That’s where the drama escalated. One commenter called it flat-out “AI slop article,” while another said an AI-reconstructed scene felt “icky” and only got caught because a whiteboard looked off. Then came the name-check controversy: why, asked one reader, was JPEG XL figure Jon Sneyers barely acknowledged, linking to his own blog post. In other words: the format may be about image quality, but the comment section was all about receipts, side-eye, and wounded trust.
Key Points
- •The article presents JPEG XL as the outcome of roughly a decade of experiments in image compression, psychovisual modeling, and optimization driven by modern display requirements such as HDR and wide color gamut.
- •Early precursor work included WebP Lossless and Brotli, where techniques such as the entropy image concept and data-driven context modeling were developed and later reused.
- •Butteraugli and the XYB color space were created to better align image compression with human visual perception than traditional metrics like PSNR and SSIM.
- •Guetzli and Brunsli explored the limits of legacy JPEG by delivering, respectively, high-density perceptual encoding and lossless recompression of existing JPEG files.
- •PIK combined earlier research into a proposal for ISO, and the push for extremely low bit rates led to the VarDCT architecture that the article says remains central to JPEG XL.