You're right about the efficiency of analyzing the simpler hypothesis first. This kind of reasoning reminds me more of Pascal's Wager than Occam's Razor, but it

You're right about the efficiency of analyzing the simpler hypothesis first. This kind of reasoning reminds me more of Pascal's Wager than Occam's Razor, but it's all good. We can use decision theory to prove that it makes sense. However, you can't go from "according to Occam's Razor, let's test the simpler stuff first" to "according to Occam's Razor, the simple stuff is more likely to be true". And that's exactly what you did when you wrote "Occam's Razor would say that the code was modified by a disgruntled employee just before leaving, no?". > Both of these are big conspiracy theories. Proven to be true, in the past, by whistleblowers. There's also the choice of an encryption scheme that no sane actor would choose on its own. So the likeliness of a malicious state-level attack is higher than the baseline you probably assign to accident / disgruntled employee scenarios.