I actually want to win every debate. Winning is hitting delta less than something small. The whole point is for each utterance to increase information gain. Aumann's Agreement Theorem always applies so if you get IG = 0 at any stage it is worth considering if the discussion is worth it.
Additionally, verbal and textual conversations lack the depth to transmit the full state of the probabilities in your model.
Consequently, it is often useful, as a participant, to break off a miniclone of oneself to perform the information interchange and then integrate it into the rest afterwards. You can't tabula rasa the miniclone easily because of the bootstrap time but you can't retain elasticity of your true mind. A more plastic form of yourself can work well.
The problem is that this approach is susceptible to information contagion.
Additionally, verbal and textual conversations lack the depth to transmit the full state of the probabilities in your model.
Consequently, it is often useful, as a participant, to break off a miniclone of oneself to perform the information interchange and then integrate it into the rest afterwards. You can't tabula rasa the miniclone easily because of the bootstrap time but you can't retain elasticity of your true mind. A more plastic form of yourself can work well.
The problem is that this approach is susceptible to information contagion.