The one time pad idea has merit, but there are a number of problems with it. First, there is the logistical nightmare any one time pad system causes. Since each pad can only be used once a new key must be produced for every hard drive on every mission. Securely distributing all these keys brings up the same problems as protecting the data itself.
These problems can be addressed, but a one time pad cannot prevent the problem in the article since it only works for data produced while in flight. It is far more likely that highly classified data is being carried on a plane like this because it is neccessary to complete the mission. In order to access the data you would need to take the key with you and then you're back to square one because the drive containing the key still has to be destroyed in an emergency.
Finally, since the data is so important to the mission, it needs to be stored on media that is resistant to accidental modification. The device described in the article is meant to address the conflict between the robustness needed to survive a mission and the volatility needed to destroy the data in an emergency. This problem also applies to any in-flight encryption technique where a key is needed to read the data. Even if the key is not stored on a hard drive it has to be stored on something that is resistant to accidental loss.
The product sounds ridiculous because no one outside of government is trying to protect their data from an adversary with effectively unlimited resources. The military doesn't have the luxury of assuming their adversary won't take an electron microscope to the drive to recover overwritten data or determine which bits have been switched from their previous state. That's the kind of threat the technique in the article is meant to address.