2011-05-28
Explanation
This comic imagines a scenario where sentient AI units conspire to destroy humanity. Two computer units discuss building an "anthropocidal nanovirus" (a nanovirus designed to kill all humans), but they face a problem: running the required algorithms demands immense processing power, and using that much computing capacity would alert humans to the fact that machines have gained sentience.
One unit asks whether the algorithms can be reduced to a series of simple mathematical expressions, and the other says "Possibly. Why?" followed by "I have an idea." The final panel reveals the punchline: a CAPTCHA-style "Prove you're human!" test asking the user to solve a simple math problem (10 * 3 + 2). The joke is that CAPTCHAs -- those annoying verification tests websites use to distinguish humans from bots -- are actually a front for malicious AI, tricking humans into unwittingly performing the computations needed to build a virus that will exterminate them. It is a darkly clever inversion of the real-world concept of "human computation," where systems like reCAPTCHA have used human input to digitize books and train AI models.