The NSA has a history of keeping very serious faults in security to themselves [1].
It's like the government detected a structural problem in bridges and then decides not to fix it because that could make it harder to destroy other countries' bridges.
This is absolutely false. The NSA asked IBM to not reveal that security against differential attacks was a design goal. They did not, as you suggest, allow the weakness to be present in DES (go read the Wikipedia article you cite again).
The NSA did make changes to the S-boxes used in DES at the last minute, and would not comment on the nature of the changes. Many people speculate this was the introduction of a backdoor. Last year at RSA a speaker from the NSA revealed that the (now 35 year old) changes served two purposes: 1. it actually fixed a weakness they had found, but did not want to disclose the details of. 2. scared the Russians into not trusting the now hardened algorithm and instead relying on older systems that they had attacks against.
This is very silly. As Cody would tell you, the Onity flaw he found was so basic (it is the electronic equivalent of the Bic cap trick that unlocked Kryptonite locks) that any EE grad working at DoD or NSA would have had it instantly. The idea that there'd be some huge conspiracy involving the government reaching out to private firms to enable them to break into trivially breakable locks rings false.
You should simply assume that the government has always, always, always been technically capable enough to break into hotel rooms undetected.
Two years before Cody's talk, an unknown entity (assumed by local police to be Mossad), used a third party device to reprogram VingCard hotel door locks in the field as part of the assassination of Mahmoud Al-Mabhouh in Dubai.
I imagine this is even somewhat of a litmus test for a powerful govt.-like organization in any area- the ability to find out where a foreign representative is staying, know when they're out, and search their room.
