
It was that slashdot article that lead me to E2 originally; interesting that gaining a million nodes since then sounds like a lot, until you think of it as merely doubling in size. Of course, there's a question as to how many writeups specifically there are, as 'everything is a node'; and whether that's more than double what we had in 2001 (or have kept from then, perhaps).
You consider obtaining the data content and algorithms by physically dismantling the chip, but I think the whole point of PUFs is that the physical structure of the chip is part of its data content, and thus would also have to be recorded in a reproducible fashion in order to clone.
There was a talk on this at last year's Elliptic Curve Crypto workshop by Pim Tuyls of Philips, but the slides aren't available online (the problem of working in industry instead of academia I guess- everyone elses are available). Doesn't take long to google up papers on the theoretical basis, but I can't rememeber how far along they were at the hardware level unfortunately.
The computer is to the information industry roughly what the central power station is to the electrical industry. -- Peter Drucker