I thought the point was that the comment showed how the arguments, which we've gotten used to and don't fully question anymore, would look ridiculous when applied in a different context. (It was a pretty effective demonstration for me - the same responses did look far less convincing when they were put in the mouth of Leverage Research people rather than LW users..)

Exactly right.

Some remarks:

• I don't think the arguments LW/SI uses against its opponents are wrong but that reality is more complex than the recitation of a rationality mantra.
• If you want to discuss or criticize people who are not aware of LW/SI then you should commit to an actual discussion rather than telling them that they haven't read the sequences.
• There is no reason for outsiders to suspect that LW/SI has any authority when it comes to arguments about AI, quantum physics or whatever.
• If you want to convince outsiders then you should ask them questions and voice your own opinion. You should not tell them that you have it all figured out and that they just have to read those blog posts you wrote.
• You should not portray yourself as the single bright shining hope for the redemption of the humanities collective intellect. That's incredible arrogant and cultish.
• You have to distill your subject matter and make it more palatable for the average person who really doesn't care about being part of the Bayesian in-crowd.