The prior probability for a person being cis is obviously much higher than the prior for trans; looking at it one way, this quote is advising ignoring that prior.

Of course, looked at another way, it's specifically noting that there should be a large complexity penalty for the hypothesis "I’m just a cis person who has somehow managed to convince myself that I’m trans to the point that I’m having this kind of crisis" relative to the hypothesis "I'm trans". And also implicitly using the reversal test.