absolutely does not follow from

You seem to be confused. It isn't supposed to follow - it is meant as a contrast! Kompo estimated 1 bit of evidence of crackpotness is embedded in prominent misuse of language. He then reaffirms that despite this he isn't saying that singinst is a crackpot institute... that is what declaring a one bit update on a very low prior means and there is no evidence suggesting that kompo intended anything else. He is making a general gesture of respect to the institute so it is clear that he isn't using this issue as an excuse to insult the institute itself.

I assumed you were, because you spoke of evidence likelihoods and Bayesian evidence in favour of propositions: but now I fear those are just locally high-status words you were using, because when you take a low prior and update on 2:1 evidence you are left with a low prior.

He knows this, has used bayesian reasoning correctly in the past in his posts and has not made a mistake here.