This story is too funny. A little background: James “The Amazing” Randi has a brilliant, longstanding million-dollar prize for anyone who can demonstrate paranormal abilities — whatever ability they choose, shown via any method they like, only requiring that all involved agree their method would demonstrate it under controlled circumstances. Why is it brilliant?

Many have tried, all have failed.

So a while back Randi revised the focus of the program toward challenging high-profile ‘psychics’ by name (maybe defending the honor would entice them the way a cool million doesn’t?) and he is also now only considering applicants who have gotten some media attention as well as convinced some academic — probably because the endless stream of crackpot losers he has examined is boring him to death. Okay, maybe it is really a matter of resources: the crackpot next door doesn’t influence the culture like those who get media attention and confuse academics.

Anyway, today I saw this story talking about how Randi had set up a remote-viewing test and (to prove no cheating on his part) had published an enciphered clue as to what was in the box. Turns out a cryptographer easily broke his amateur cipher and was able to pass the test, beating out the psychics.

The moral of the story is actually a message Randi routinely gives to scientists, who are notoriously prone to being duped while testing paranormalists: if you are in waters where people often fool themselves and others, have a magician on staff dammit! Well, if you are in the realm of hiding information in plain sight, have a cryptographer on staff dammit!

Or at least consult someone who has some dealings with security. Even a moderately knowledgeable software engineer would have immediately recognized this as an important problem that has been studied hard by cryptographers. Randi could have written a detailed description and run it through some common and well-designed cryptographic hash function to produce a mathematical “fingerprint” which he could safely publish. Good hash functions are very chaotic and one-way, making it infeasible (read: essentially impossible) to use the fingerprint to figure out the original string as happened in this story. Then when the box is opened, he need only show his description matching the contents and the hash function generating the published fingerprint.

While I chuckle at the story, in all seriousness my hat is WAY off to James Randi and his long record of kicking butt and making the world a saner place. At the top of his list of achievements is his bringing about that stark fact which deflates so much silliness: many have tried, all have failed — and on their own terms.