Really? Why not explain why this is wrong instead of just attacking. I could very well randomly generate material "owned" by someone. As the size increases the chance goes down to being very low of course.
if I understand the concept correctly, basically it's wrong because your index/seed values will be much more complex than the data you are supposedly extracting. So basically you have to already know the data you want in order to "find" it in pi.