Google's JPEG XL victory lap, complete with an AI-reconstructed whiteboard scene

Google writes the official history of JPEG XL

Google's open-source blog published a retrospective arguing that JPEG XL is the cumulative result of years of public, iterative compression research. It surfaced now as a kind of victory lap — and the thread spent most of its energy on everything around the history rather than the history itself.

A decade of experiments compressed into one lineage

The blog's thesis is that no single breakthrough made JPEG XL; a chain of open-source projects each tested one idea. WebP Lossless (2011) introduced the "entropy image" concept, using a secondary image to choose entropy codes. Realizing that math metrics like PSNR fail in color-rich scenes, the authors built Butteraugli and the XYB color space to mimic human vision; Guetzli (2016) used it to make legacy JPEGs "20-30% smaller," while Brunsli enabled lossless recompression of existing JPEGs. Those converged into PIK, which Google proposed to ISO, and merged with Cloudinary's FUIF — combining "PIK's faster-to-decode distribution selection with FUIF's sophisticated context trees." The graph shows three eras of building blocks stacking into the standard, which ISO pushed down to 0.06 bits per pixel — roughly 35x compression of internet-quality images. The credited authors are Jyrki Alakuijala, Zoltán Szabadka and Luca Versari.

The thread reads the byline, not the bytes

The top reaction wasn't about codecs at all — it was about an AI-generated illustration in a technical post:

"In this Gemini-reconstructed scene … I'm generally pretty pro-AI, but I find this icky … the whiteboard drawing seemed not quite right, so I'll probably be fooled in the future." — a self-described pro-AI reader

The whiteboard photo, captioned as a discussion of a "VarDCT Block Joining Strategy," drew a longer broadside about the whole article's polish:

"The sloppa in this article is... offending. The obvious AI headings, pointless genned image of people … this is a technical blog, not an investor sales pitch! … If even Google, previously a gold standard of technical writing, is falling prey to this." — a reader who'd followed JXL standardization for years

A Forbes-veteran engineer flagged a substantive omission in Google's lineage:

"Weird they don't name Jon Sneyers - a person pivotal in creation of JPEG XL" — the creator of Video Hub App, ex-engineer at Forbes

And a JXL fan pointed at the elephant in the room — Google narrating a format it once dropped from Chrome:

"coming from the google blog I was hoping for some news about how they are pushing the format forward … Android is the only mainstream OS that does not support JPEG XL right now." — a self-described huge fan of JXL

A real history, told by the party that stalled it

The underlying chronicle is genuine and impressive — Butteraugli, Brunsli and PIK are real milestones, and the 0.06-bpp target is a serious engineering bar. The problem is the messenger and the medium: the company telling this triumphant origin story is the same one that pulled JPEG XL from Chrome and still won't ship it in Android, and it chose to illustrate a decade of careful research with a synthetic photo of people who, as far as the thread can tell, were generated rather than photographed.