I think I agree with everything here. Would it be fair to summarize this as:
Proposals such as mine won’t do any good, because this is fundamentally a cultural problem not a methodological one
People know what “math” looks like but they don’t understand Bayes (in your isotope example)
Yea… well with math in general, you can quite effectively mislead people by computing some one out of context value which grossly contradicts their fallacious reasoning. Then the fallacious reasoning is still present and still doing strong, and something else gives in to explain that number.