You've covered a lot of ground. Before moving on, it's worth pausing to see how the pieces fit.
Artificial intelligence is the broad project of building systems that can perform tasks that typically require human intelligence. But the version that exists today doesn't think the way people think. It doesn't reason from principles or draw on lived experience. What it does is recognize patterns, at enormous scale, across enormous amounts of data.
Machine learning is the mechanism that makes this possible. Instead of following rules a programmer wrote, a machine learning system is trained on examples. It adjusts itself, iteratively, until it gets good at the task. The "learning" is real, even if it doesn't look like human learning.
Neural networks are the architecture that underlies most of what you'll encounter. Loosely inspired by the structure of the brain, they process information through layers, each one building on the last. Deep learning is what you get when those networks go many layers deep, which is what unlocked the capabilities that made AI feel genuinely new: the ability to recognize images, understand speech, generate language, and more.
And yet: these systems fail in predictable ways. They hallucinate. They're brittle at the edges of their training. They reflect the biases in the data they learned from. These aren't bugs to be patched; they're structural properties of how the systems work, and they're worth keeping in mind every time you use one.
That's the mental model. It won't answer every question you'll have about AI, but it will make the rest of the learning center make sense. The sections ahead get more specific: the types of AI you'll actually encounter, how these systems are built and trained, what they're being used for, and what it means to use them well. You're ready for all of it.
