Can we really trust polls?

This article does not include information I find critical in determining if we should value this polls information or not, for instance how many people were polled, what did they state as their political allegiance if they have one, and what was the age group of those polled. The ultimate question though is even if that information was provided can we still trust this poll, or any other poll?

In the article itself it states that "The problem with interpreting the generic ballot is that a national poll cannot produce the same results that 435 separate polls in 435 House districts would produce," so does that mean we shouldn't pay much attention to these polls that claim to be the "poll of polls" and what not?