It all depends on the metric you are optimizing for. In academia, the metric would be grant money, directly influenced by number of published articles and citations.

From that perspective, the system works. We are making more articles, in more journals, there is also plenty of money thrown around. Unfortunately, there is no incentive for correctness, novelty or usefulness in this system.

Falsification of results, especially in the soft sciences, is relatively easy. Verification of results, doesn't give you any credit. So you can have people producing articles with blatantly misleading or false results for decades, all without any repercussion.

And it's not much better in the hard sciences either. Because verification of results there, is even more difficult and costly. And again, we are not incentivizing verification.