Lesson 265 of 2116
The Jagged Frontier of AI Capabilities
AI is amazing at things that should be hard and terrible at things that should be easy. That jaggedness is the key to using it well.
Lesson map
What this lesson covers
Learning path
The main moves in order
- 1Smart Here, Blind There
- 2jagged frontier
- 3capability surface
- 4Mollick
Concept cluster
Terms to connect while reading
Section 1
Smart Here, Blind There
Ethan Mollick coined 'jagged frontier' to describe the weird reality that modern AI can write a competent legal brief in 30 seconds but fail at counting the letter R in 'strawberry.' The frontier of ability is not a smooth wall. It is a jagged coastline.
Examples of the jaggedness
- Writes a credible sonnet but cannot reliably reverse a string
- Solves olympiad geometry but misses simple arithmetic edges
- Generates fluent Finnish but confabulates Finnish history
- Codes a React app but cannot tell you which character is the 5th of a given word
Why the frontier is jagged
- 1Training data density: plentiful topics are well-learned, rare ones are not
- 2Tokenization: character-level tasks fight the model's native unit
- 3Reasoning depth needed: multi-step chains are fragile
- 4Mismatch between training distribution and the specific task
How to use AI on a jagged frontier
Compare the options
| Do | Do not |
|---|---|
| Test the specific task with the specific model | Assume 'it is smart' generalizes |
| Use AI inside its clear capability zone | Push it into adjacent tasks without checking |
| Build a personal map of what each model is good at | Trust one global leaderboard as definitive |
| Verify anything important | Assume fluency means correctness |
“AI is like a coworker with a jagged skill set — superhuman at some things and oddly incompetent at others.”
Key terms in this lesson
The big idea: AI ability is not a single number on a scale. It is a jagged coastline. Learning its shape is the difference between skilled use and frustration.
End-of-lesson quiz
Check what stuck
15 questions · Score saves to your progress.
Tutor
Curious about “The Jagged Frontier of AI Capabilities”?
Ask anything about this lesson. I’ll answer using just what you’re reading — short, friendly, grounded.
Progress saved locally in this browser. Sign in to sync across devices.
Related lessons
Keep going
Creators · 45 min
What Is Intelligence, Really? A Working Framework
Before we can judge whether an AI is intelligent, we need a framework for what intelligence even means. Draw on Chollet, Dennett, and modern evals.
Creators · 45 min
The Economics and Ethics of Training Data
Data is the strategic asset of AI. Understand the supply chain, the legal fight, and the philosophical stakes before you build anything on top.
Creators · 45 min
Emergence, Capability Forecasting, and Safety
Emergent abilities make AI both more exciting and more dangerous. How do labs forecast what the next model will do — and what happens when they are wrong?