I esp. like the plausible deniability hack against bruteforce, the second to last story.
But realistically most cyphers were hacked with some plain text knowledge, esp. in the header/beginning, which would defeat this bruteforce claim.
But it would work with random text and without an AI which can derive a security context automatically. But apparently this exists. So you'd need to encode your text twice, first by using only innocent words, and using something like this link-grammar encode again.
But realistically most cyphers were hacked with some plain text knowledge, esp. in the header/beginning, which would defeat this bruteforce claim. But it would work with random text and without an AI which can derive a security context automatically. But apparently this exists. So you'd need to encode your text twice, first by using only innocent words, and using something like this link-grammar encode again.