Just as a discussion point (and going off topic and being maybe too pedantic) that is not how I understood the XKCD. The difference is that the one interpretation only takes into account the odds of the machine lying, the bayesian approach also takes into account the odds of the sun actually going nova, which is still neglible.

Yes as a second consideration you can also take the view of the world will end, but that is not the point of the XKCD.

And then drawing it to this discussion, yes the odds that Trump would be willing to start a nuclear war might be high, but combined with the odds of an actual nuclear war occurring the odds of it happening is still neglible.