Some problems with that: "discredited" isn't a binary; memes (in the original sense) from papers circulate easily even if nobody pays the publishers (the problem is arguably even worse, because people will read abstracts for free and distort the claims further); there might not be higher quality alternatives.