The misunderstanding about XP and "doing the simplest thing" is more often than I'd like taken to absurd extremes. For example I encountered a developer recently repeating a claim that an XP team would not use a framework that had been built to simplify a certain style of software development.
Certainly the XP team would want to understand their experience with the framework, the maturity of the framework, and the nature of the problem's suitability for the framework. But not use it, period?
So how far does this go?
"I'm not sure we need a computer for this project, but when we do we'll dig up the coal to forge the steel to make the rack to mount the motherboard."
"I'm not sure we need a high-level programming language. We'll use assembly language and see what patterns emerge, then let the higher-level language evolve."
I think there is a spirit and a letter to the law, and the same goes for XP values. Just work in small pieces with concrete feedback. Avoid too much prediction and too many assumptions. Common sense works wonders.