There is (at least ;) ) one really good challenge listed: The usability test. First, it's surprisingly fun. And second you usually get great results. A while back I studied HCI and did a usability test of a web project of mine. Testing it with an user test was just a small part of the project, but it was a great practice. I learned that a lot of the things I considered to be easy weren't easy, and the other way around, stuff I thought complicated actually worked fine. All while already being aware of a lot of design principles, usability knowledge and heuristics. But really testing it still uncovered a lot.
Example: That site used browserid as a login. Back when I tested that initially it was completely easy. But in the meantime it had become persona, and each and every of my testers failed to log in with their email address because they did not get directed towards my site, but towards the persona account management.
Definitely fun and worth the time investment!
The challenge site for that is listing conflicting views on how to do a usability test though. That's not surprising, because it's such a broad field with many conflicting views. And it shouldn't be surprising that an UX site is not that clear about usability (it's not close to the same thing after all). But maybe it could be a bit clearer? The site should convey:
1. The target of an usability test is to collect critical incidents, blockers that occur while the user is using an app/site.
2. For those to occur the user has to have a task to finish. Usability is about effectiveness and efficiency in a given context while achieving a given task. So the user has to have a clear task. Don't let him just browse.
3. Contrary to what some of the linked resources claim, do not ask questions like "what would you prefer." That's participiatory design or part of user requirement research. Do that in the design phase, however yours looks like, not while evaluating usability. Instead, observe the critical incidents, ask what he thinks when he is blocked/confused, and guide him through subtasks if he can't continue on its own. All while keeping notes on what happens.
The end result of an usability test is a list of criticial incidents: The description of what went wrong ("The user is confused; He wants to do X, but the button is not recognised as a button, making the task X impossible"), which dialog principle was not followed, and optionally a suggestion on how to fix the problem ("do not use flat design for buttons").
Would be great if the site would link to exactly one resource that correctly describes this. The nielsen article might be the best option there, but it's a bit broad and theoretical.