In most cases, it's the same thing.
Making code look like it does something else and making code difficult to understand rely on the same principle. Ultimately, the goal of both is to hide the real purpose of the code.
Case in point, your self-answer to the challenge that spiked this discussion looks like a tutorial about pseudo-random number generation, yet produces a specific output.
That doesn't mean that all challenges that involve obfuscation are necesarily off topic. For example, we had cops and robbers contests where the robbers had to find certain inputs that produced a certain output. That's objectively measurable (shortest code for cops, most cracks for robbers), and doesn't suffer from the same problems as obfuscation popularity contests.
For the latter kind, all the reasons that drove the community to blacklist underhanded challenges still apply. They're essentially code trolling in disguise, don't add any quality content to the site, and are too broad and too subjective for the scope of PPCG.