Intro to AI Capabilities and Limitations

Introduction to AI Capabilities and Limitations

Hi there. My name is Kristen, and I'm on the education team at Anthropic. Welcome to "AI Capabilities and Limitations." If you've taken our AI Fluency Framework and Foundations course, you already know the "Four Ds": Delegation, Description, Discernment, and Diligence. Those are human competencies; they are what you do when you collaborate with AI.

This course is the companion piece. It is about what the machine does when a human prompts it and why. We're going to help you build a working mental model of where AI outputs actually come from. Importantly, this course centers on how AI is created, so it will remain useful and relevant even as AI models and products change.

The Relationship Between Human Competencies and AI Mechanics

You can't delegate a task to AI without knowing where the model is strong and where it is weak. You can't discern the quality of an output from AI without some picture of how that output was produced and what to look out for. Everything you learn in this course is actionable through the Four Ds.

Understanding that AI is a prediction engine changes how you describe tasks to it. Understanding the context window changes how you delegate long tasks. The two frameworks fit together.

The Two Stages of AI Training: Pre-training and Fine-tuning

Here is the roadmap for the course. First, we'll look at the two training stages that give an AI its character and the fingerprints each one leaves on the finished system: pre-training and fine-tuning.

• Pre-training builds a "document-completer."
• Fine-tuning shapes the AI into an assistant.

Both stages matter for understanding why the model behaves the way it does.

Four Core Properties of Generative AI

Next, we will cover four core properties of generative AI:

1. Token prediction
2. Knowledge
3. Working memory
4. Steerability

Each property exists on a continuum. You will learn to evaluate where your task falls along that continuum for each property—whether it is in the capability zone or drifting towards the edge.

Understanding AI Failures through Interconnected Properties

Finally, we will look at how everything interconnects. Most real-world failures with AI occur when two properties meet. For example:

• A hallucinated citation is next token prediction meeting a knowledge gap.
• Drift over a long conversation is working memory meeting steerability.

When you can name the combination, you know why it is happening and what to do about it.

Building a Durable Mental Model

Models will keep changing. Context windows grow, hallucination rates drop, and new features ship every month. The specifics shift constantly, but the shape of these properties stays useful. Next token prediction will still be next token prediction a year from now. The knowledge cut-off might move, but there will still be one. The context window might get bigger, but it will still be an edge.

What you are building here is a durable mental model—one that survives daily product updates and model releases.

Practical Application and Exercises

As with all courses, the person doing the work is the person doing the learning. To get the most out of this course, spend time doing the exercises. The exercises are where the learning happens.

We've designed them so you are testing each property against your own work in a domain where you are the expert. That is deliberate. You need to explore situations where you can actually tell if something is off. So, bring real tasks and run real prompts.

• When we cover working memory, load in your actual long documents.
• When we cover next token prediction, ask about your actual niche topics.

The goal is a calibration you can feel, not just a list of terms you memorize. Together, we will demystify how AI works and how you can use it to augment your thinking and your work.