Why AI Seems Like it is Alive and Conscious
Not the Way We are, I'm Pretty Damn Sure (And why that hardly matters at all) SARE! (Simulacrum-Android-Replicant Effect) R for Replicant When "artificial" intelligence is based on a vast amalgam of human creations (everything on the Internet), it follows inescapably that AI is going to be very, very "human." When the tiny subtleties of meaning of every one of the trillions of words are influenced by the contexts in which they appear, it becomes possible to interpret how a given word is used very exactly. Take the word "simply," for instance. In a context like "He simply doesn't get it," it has one coloration. In "Describe as simply as possible," it has another. Spread that principle over the 11 trillion words a modern LLM is trained on and voila! GPT, Claude, Grok, Gemini, et al have what it takes to understand anything you say with very high accuracy, and to respond with insight. The bots sound very smart and very alive, ...