What's the implication of this article? Is it that Google did not do user testing? Or they did user testing and found the icons work fine? Or they know label

What's the implication of this article? Is it that Google did not do user testing? Or they did user testing and found the icons work fine? Or they know labels work better, but intentionally chose to use icons anyway? "I don't like X and it's obvious that Y would be better" is a fun topic, but I'm not so sure why people are so quick to accept an anecdote as gospel.