If something is difficult, the best possible solution to that difficulty is to just not do it... said no one who accomplished things.
Yes, choosing the right clustering keys is hard. It's not however impossible. Further, sometimes having a bad clustered index can hurt performance, but frequently, it's just meh, it's not hurting, but it's not helping. In general, when you consider how SQL Server works, I'd err on the side of having a clustered index on every table, even if it's not perfect, with some exceptions. This is as opposed to, all the tables are heaps, because picking the clustered index is hard, and a few exceptions where the clustered key was clear have clustered indexes. I'm positive, based on years of experience and testing, that the clustered approach is going to work better than the heap approach.
I've read that persons blogs before. They have more than a few controversial approaches. I think it's primarily to gin up hits as opposed to actually teaching useful methodologies.