In the article, there are a couple of passes that mention “benchmarks” where Firefox was faster than Chrome. It sounds like those benchmarks, albeit I assume they were correct, were actually providing a distorted reality of what “fast” actually means, because I distinctly remember people (including myself) switching to Chrome in early days because it was simply objectively faster. It sounds like the existence of those benchmarks have had a net negative on Firefox: it provided a false confidence that the browser was “fast”, and possibly delayed the refactorings that eventually managed to really speed up Firefox again.
You mention that Chrome used “UX tricks” but I assume what this means is that Chrome developers were actually driven by more correct benchmarks: the ones that actually translate to a better experience for the end user.
Since I’m always intrigued by the topic of false benchmarking and how complex it is to do correct benchmarks that don’t fall into traps, I’d love to hear what those benchmarks you mention were measuring, and if you agree that they gave a false assurance to the Firefox team.