These are pretty extraordinary claims and I'm curious to know whether they were ever tested in controlled conditions.
For example, how does one verify the recollection of the contents of "at least" 12,000 books? Was there an experiment performed, where he was asked questions about a sufficiently large, randomly chosen, sample of those 12,000 books?
I understand that Tim Peek made a living out of his mnemonic abilities and I think, in that context, skepticism is justified.