What's the difference between AI, machine learning, and deep learning in plain terms?

Think of them as nesting dolls. AI is the biggest doll — the broad idea of machines doing smart things. Machine learning is the middle doll — a specific way to achieve AI by letting machines learn from data instead of following explicit rules. Deep learning is the smallest doll — a particular type of machine learning that uses large, layered networks inspired loosely by brains. Here's a concrete example. Say you want a computer to recognize cats in photos. The old-school AI approach would be writing thousands of rules: pointy ears, whiskers, triangular nose, specific fur patterns. It fails constantly because cats sit weird and lighting changes. Machine learning flips this. You show the computer 10,000 labeled photos — "this is a cat," "this is not a cat" — and it figures out the patterns itself. No rules needed. Deep learning takes this further. Instead of you deciding what features matter (ears, whiskers), the system learns its own features layer by layer. The first layer might detect edges, the next layer combines edges into shapes, the next layer recognizes eyes or ears, and so on. You never told it what an ear looks like. It figured that out. In practice, if you're using a simple spreadsheet of numbers to make predictions, you're doing traditional machine learning. If you're working with images, audio, or messy text and using something like a neural network, you're in deep learning territory. The line blurs constantly. Don't stress about precise definitions — focus on what tool fits your problem. One thing that trips people up: all deep learning is machine learning, and all machine learning is AI, but not the other way around. A chess program that uses hardcoded rules is AI without any learning at all.