It's funny how the author uses published studies to support his claim that science is broken. What's statistically more likely, that 80% of articles conflate "statistical significance" with colloquial "significance", or that
one study misinterpreted many of the others. That requires bias on the part of only one team of authors. Nah, the studies that support my thesis must be true.
But if those studies (which reviewed other studies) are true, it's a nice example of the self-correcting nature of science. We've gotten this far. To say that science is fundamentally broken is a bit of an exaggeration.