JPEG XL Is a Bad Fit for the Web, and the Data Shows It
The case against JPEG XL
A compression engineer explains why JPEG XL, despite its technical strengths, is the wrong choice for the Web. He argues that libjxl's lossy compression lags behind AVIF in both perceptual quality and decode speed, its lossless advantage is negligible, and its extreme flexibility creates security risks like JXL-bombing. The piece offers a detailed, empirical comparison of coding tools and performance.
I believe Web codecs should be purpose-built, efficient, and narrowly scoped to the needs of the Web.