You might as well ask, "What's wrong with putting all the functionality of my program into one giant class?" On the one hand, there's nothing wrong; it can be made to work. On the other hand, everything is wrong: if your program is one big ball of wire, you have to straighten all of it out in order to change any of it.
Look at a Windows Forms programmer written by a beginner. You'll find all of the business logic in the buttons' Click event handlers. What's wrong with that? Don't you want that logic to be executed when the user clicks the button?
This is essentially what you're proposing doing within the world of WPF. Will it work? Sure. For trivial projects, it may even work well. You're accumulating technical debt, though, and when the time comes, you'll have to pay it off by refactoring your code into something manageable.