Plaudit is nothing more than like an upvote, right? I don't think this is really helpful. We already have a very similar metric, which is citation count. Which is definitely a useful metric.
At first, I thought that Plaudit adds some post peer review platform, or discussion forum, for each paper. Similar to OpenReview but for any paper. I think this might actually be a great idea.
It's not a metric, it's an endorsement. Think of it like being published in a journal: it's not about in how many journals an article gets published, but in which ones. Likewise, what's important here is who has endorsed it. Instead of a journal name acting as a proxy for "someone I trust thinks this work is worth reading", you can directly see the name of that person.
(That said, a big challenge with citation counts is that they take a long time to accumulate.)
As I mentioned, I think the three functions should be separated. Although Plaudit by itself does not facilitate giving feedback, it does not prevent it either. Ideally, if someone comes across e.g. an error in an article, they provide that feedback to the author (e.g. through an email, of using Hypothes.is - a project that I think you might like). There's no need to only contribute to improving scientific literature through a formal journal-assisted peer review process.
And who knows: if the author incorporates that feedback, the giver of the feedback might decide to endorse that article using Plaudit :)
At first, I thought that Plaudit adds some post peer review platform, or discussion forum, for each paper. Similar to OpenReview but for any paper. I think this might actually be a great idea.
I think that https://fermatslibrary.com/librarian actually provides a similar functionality. But it has not seen much adoption.