Couple of quick points on that story: first, it should be obvious that it's a motivated leak. It's clear that the only way Romney can win is if there's a massive systematic error with all the state polling. It should not be surprising then that a GOP pollster suddenly produces a story on how there might be a massive systematic error with all the state polling. Taken in that light, this is another example of GOP recognition of how bad Romney's position appears.
Secondly, importantly, the argument hinges on whether polls are too bullish or too bearish on Latinos and young people. There isn't, as far as I can see, an obvious reason that a lot of polls would all be too bullish on Latinos and young people. There are a couple of well-known reasons why they might be too bearish: pollsters generally only interview in English and they generally don't reach cell phones. Historically these factors have proven to underweight Latino and youth voters. Polls that do call cell phones? Consistently show even better results for Obama. The very few polls that do Spanish and English interviews? Latino Decisions shows more Latino vote intention than ever before and huge margins for Obama.
So what are the theoretical reasons that the polls might be overweighting them? It's not really sufficient to theorize an effect; to be taken seriously, you need to theorize a cause.