OK, but it's been broken for ages. There is a simple python script floating around that you can run on Kindle books to turn them into .html files. The decryption key is just your Kindle's serial number (which I believe is available via USB, so typing it in is not strictly necessary).
If you don't have a Kindle, that might be a problem though. But I think Amazon lets you "register" a Kindle, so you can just make up a serial number, buy books, and decrypt them.
Basically, the DRM scheme is just a joke to keep publishers happy. If your eyes can see something, it can be copied.
Actually, K4PC uses a different scheme - the book key is not encrypted in the same way as for normal kindle books. I believe each book uses a different Mobipocket PID, and what this new script does is extract the PID for a particular book and then invoke the original script using that PID.
Also, the PID is not the kindle serial number, but can be calculated from it. There's another script around to do that.
If you don't have a Kindle, that might be a problem though. But I think Amazon lets you "register" a Kindle, so you can just make up a serial number, buy books, and decrypt them.
Basically, the DRM scheme is just a joke to keep publishers happy. If your eyes can see something, it can be copied.