Did you use a LLM to help you write this article? It seems to possess the unusual blend of precisely-written prose and argumentation that sounds reasonable in detail, yet an almost complete lack of perspective that comes from having actually tried to utilize the methods, that LLMs often produce.

If not: the critique is embedded in the above query.

In particular, (1) if you simply use Bayesian methods on uniform prior hypotheses, you are reasoning from priors the way frequentists typically do; (2) most frequentist methods are only computationally efficient assuming distributions which are wrong (but yes, Bayesian methods can be more expensive because they model part of the scenario that is overlooked in a frequentist method); (3) you can correct for overfitting (e.g. BIC) just like you can with frequentist multiple comparisons; (4) diversity is a reason to be clear, not a reason to not use something; (5) misinterpretation is rife with any sort of statistics, alas, and the education needed to understand adequately is not that different; and (6) the set of problems for which Bayesian statistics can be an improvement is large and asking for it to be better everywhere is unreasonable.

In practice, for a lot of questions, the frequentist tools are common, good enough, and widespread, and you should reach for those; and for a lot of questions, a Bayesian approach is a very effective way to generate insight into what is likely to be the case and where you might be fooling yourself due to stochasticity.