For starters, anyone that ever worked with a codec, will know that you don't compare them with ONE SINNGLE IMAGE.
This whole basic idea of the blog post is just to generate more whining and clicks and not to actually make a comparison between formats that's worth a basic smell test.
This cuts against WebP more: all of Google’s marketing was “it’s a third smaller!!!!” and then when you looked they were comparing it to unoptimized libjpeg outout and using computational metrics like SSIM which only crudely approximate what humans notice about image quality.
I did the same comparison the author did when WebP came out but used an optimized JPEG encoder and found the same conclusion: when you produced subjectively equivalent images, the savings were more like -10% to +15% and for web sites which didn’t get Google-scale traffic the performance impact was negative since it made caching less effective and you had to support an entire new toolchain.
Criticism of cherry-picking cuts against WebP because the marketing campaign for that codec relied on cherry-picking both the least optimized JPEG codec and the most favorable metrics for comparison. If you had humans comparing images or enabled JPEG optimization you saw far less exciting numbers for WebP - usually under 10% savings, not uncommonly negative – and there were other formats which consistently outperformed it. You can see the mood around that time here:
Even a decade later, however, Google repeats the 25-34% claim and their performance tools tell developers they should use a modern format, which by sheer coincidence means the one they invented rather than the best ones on the market.
Except the problem isn't in a single image, it is a pattern that is frequently there and the image was only used to demonstrate it. WebP has this problem way back as one of the reason others were hesitant to support it except Google.
This whole basic idea of the blog post is just to generate more whining and clicks and not to actually make a comparison between formats that's worth a basic smell test.