I don't have flash installed at the moment so I didn't watch the video, but I'm jumping in anyway.
I think it depends on the nature of the result descriptions. I've seen unofficial MBTI test summaries that were incredibly vague or that were based solely on the 4 letters, and so in those cases, a wrong results are (statistically speaking) going to be about half right for anyone. If all the descriptions have a positive spin on top of that (which a lot do), then people are going to go along with it thinking "yeah, this is pretty close and it makes me seem pretty awesome, so I'd say, 'sure, that's me'". But if they saw one for their actual type to compare it to, there's a good chance that many people would change their minds.
That's why I prefer to see type descriptions that either get very specific, or that include negative traits. They are more likely to throw up red flags for people when the test results are wrong.
Also, I've always been a fan of the concept of subtypes (especially for SJs).