Alright, I’ll admit it: I feel like I’m missing something fundamental. Every article or product description throws around "AI" and "machine learning," but I still don't get what it's actually doing. Is it just a fancy set of rules, or is it actually "thinking"? I keep hearing about models being "trained," but on what? Can someone break down how does ai works in a way that focuses on the core concept, not the complex computer science? What's the basic mental model I should have?

That's a great question, and you're right that the term gets used so loosely it loses meaning. The simplest mental model is to think of most modern AI not as "thinking" but as advanced, automated pattern recognition. It doesn't follow a fixed rulebook written by a human. Instead, it's a program that learns patterns from enormous amounts of data.
Here's a common analogy: Imagine you want to teach a computer to identify spam emails. Instead of programming thousands of specific rules ("if it says 'Nigerian prince,' flag it", you'd feed it millions of examples of emails that are marked as "spam" and "not spam." The AI algorithm sifts through this data, finding subtle statistical patterns in the words, phrases, and sender info that correlate with spam. It "trains" itself to recognize these patterns. After training, when a new email arrives, it predicts the probability that it's spam based on those learned patterns. This same principle applies to recognizing images, translating languages, or suggesting what to watch next. The "intelligence" comes from finding correlations in data, not from conscious thought. For a really clear, straightforward walkthrough of this process, I found this article super helpful: how does ai works. It explains training, models, and predictions without the heavy jargon.