Failure to Reproduce Modern Paper Claims [D]

I have tried to reproduce paper claims that are feasible for me to check. This year, out of 7 checked claims, 4 were irreproducible, with 2 having active unresolved issues on Github. This really makes me question the current state of research.