Haha, Sean Eddy used to complain about reproduction efforts in bioinformatics (specifically, people trying to benchmark HMMER and doing a bad job). Fortunately, my advisor helped me learn the techniques and Sean approved (he was also happy that HMMER beat BLAST for remote homolog detection).

I also left a postdoc position over my professor's decision to rewrite my paper to juice all the stats- not a specific error, but selectively interpreting the data to make the results look better than state of the art, when they were not.

I would absolutely love to have my "paper correctness AI" mark all those bad reproductions in the literature and outright misrepresentations- it's all too easy to rush to publish and get a lot of attention- especially if your advisor or coauthors are prestigious and mildly unethical.