The constraints on the preservation must be either real (Google's fault), legal (Google's fault, or privacy), or Google's decision (Google's fault) Reasoning: I'm guessing it's cheaper and easier for Google to provide IA with the data directly from the backend rather than scrape.

Surely access is not the issue here; it must violate DMCA and or the 5th. You can't have it both ways

Ummm. It's beta. or didn't you notice that.