"Peer reviewed" as a binary concept might be as well, given that incentives have aligned to greatly reduce its filtering power.
They might both be examples of metrics that became useless as a result of incentives getting attached to them.
Right now Elsevier is by far the biggest offender and also happens to the be the topic of the conversation and the article.
That being said, I'm happy to encourage open access.