No improvements have been made to addition; thus, it does not need to be replaced.

Meanwhile, 260 years of improvements have been made to Bayes, in particular with the addition of Confidence Intervals, which Bayes' Theorem is incapable-of. Why would anyone use a statistical tool that is so out-dated it can't give a confidence interval?

When better comes along, then yes, you become out-dated and silly for being old-fashioned.

Another example to prove your point wrong: neural networks recently found better ways to multiply matrices, and thus we SWITCH to the new one, instead of using the old, out-dated one. That *always* happens... IF an IMPROVEMENT has been made. So, your argument is based upon erroneous equality; addition is NOT like Bayes, because addition has not improved, while statistics HAS improved. Your argument demonstrates your quality of thought.