Online activists are silencing us, scientists say

People need to accept that both they, and others, are allowed to be wrong. They need to be able to express that without backlash.

Also, this assumes that all "errors" are alike, but a methodological error in a study is very different from "new evidence has come to light that questions the hypothesis", "we now know that these things have a third common cause", "we discovered in subsequent research that there are confounding factors we did not know to control for in the original study", etc.

IMHO, this comment badly misunderstands the point of scientific inquiry. It isn't to produce a final and immutable result, it's to make observations to rationally assess a hypothesis, and then to repeat the process. This means that science necessarily represents the best interpretation we have for existing observations, with the understanding that subsequent observation might show us confounding factors we couldn't have thought of before.

To put this another way: what if Newton had articulated his understanding of physics in the internet age, and then relativity came along? Do we now tar and feather Newton because he failed to anticipate the contradictions between his position and subsequent observations around the speed of light? That seems like an impossibly high standard to hold scientists to, and that to me is the crux of the problem: we should be expecting scientists to get things progressively more right and to explain why, not to get it completely right the first time.

If a paper turns out to be flawed, then you retract the paper. In order to retract it, someone needs to figure out that it was in fact flawed. Most scientists don't knowingly publish bad research, they just make mistakes that get missed in peer review. Keeping research hidden away in paywalled journals isn't going to improve the quality of science one iota.

Expecting scientists to maintain radio silence is roughly analogous to security by obscurity, with the same obvious shortcomings. The literature is full of widely-cited studies with significant methodological flaws, because they're useful enough to cite but not important enough to warrant serious scrutiny. If we're serious about building a credible scientific literature, we need to subject papers to the disinfectant of sunlight. We need to leverage the power of social media to make connections between scientists, iterate faster, bring informed laypeople into the conversation and find those bad papers.

All scientific studies have errors. Science itself is the process of finding errors and correcting them. It is a asymptotic journey of discovery that tries to obtain perfect knowledge but will never quite reach it.