A former co-worker of mine recounted an experience with a person hired by his previous company who had written a book on software development methodology, Scrum or something. Evidently the way the author's employment worked is the company got the reputation benefit from having a famous expert on staff. In exchange he gave talks but didn't have to do any actual work for the company, and the company didn't have to follow his methodological orthodoxies.
And that's how you end up with speakers saying ridiculous things that don't work in practice. Because they come up with some idea (or more likely, a very minor tweak to an existing idea) but because nobody has to actually do it, and all their time is spent speaking, nobody uncovers the critical flaws.
