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.

35 min · Reviewed 2026

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

Training data density: plentiful topics are well-learned, rare ones are not
Tokenization: character-level tasks fight the model's native unit
Reasoning depth needed: multi-step chains are fragile
Mismatch between training distribution and the specific task

How to use AI on a jagged frontier

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.
— Ethan Mollick, Co-Intelligence (2024)

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 check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-creators-jagged-frontier

What is the core idea behind "The Jagged Frontier of AI Capabilities"?
1. 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.
2. AISI
3. category
4. Probabilities over all outcomes sum to 1
Which term best describes a foundational idea in "The Jagged Frontier of AI Capabilities"?
1. capability surface
2. jagged frontier
3. tokenization
4. demo-reality gap
A learner studying The Jagged Frontier of AI Capabilities would need to understand which concept?
1. jagged frontier
2. tokenization
3. capability surface
4. demo-reality gap
Which of these is directly relevant to The Jagged Frontier of AI Capabilities?
1. jagged frontier
2. capability surface
3. demo-reality gap
4. tokenization
Which of the following is a key point about The Jagged Frontier of AI Capabilities?
1. Writes a credible sonnet but cannot reliably reverse a string
2. Solves olympiad geometry but misses simple arithmetic edges
3. Generates fluent Finnish but confabulates Finnish history
4. Codes a React app but cannot tell you which character is the 5th of a given word
Which of these does NOT belong in a discussion of The Jagged Frontier of AI Capabilities?
1. Solves olympiad geometry but misses simple arithmetic edges
2. Writes a credible sonnet but cannot reliably reverse a string
3. Generates fluent Finnish but confabulates Finnish history
4. AISI
Which statement is accurate regarding The Jagged Frontier of AI Capabilities?
1. Tokenization: character-level tasks fight the model's native unit
2. Reasoning depth needed: multi-step chains are fragile
3. Training data density: plentiful topics are well-learned, rare ones are not
4. Mismatch between training distribution and the specific task
Which of these does NOT belong in a discussion of The Jagged Frontier of AI Capabilities?
1. Tokenization: character-level tasks fight the model's native unit
2. Reasoning depth needed: multi-step chains are fragile
3. Training data density: plentiful topics are well-learned, rare ones are not
4. AISI
What is the key insight about "Not a single axis" in the context of The Jagged Frontier of AI Capabilities?
1. There is no IQ score that captures a model's ability. It is good at some things, bad at adjacent-seeming things, and the…
2. AISI
3. category
4. Probabilities over all outcomes sum to 1
What is the key insight about "The danger of demos" in the context of The Jagged Frontier of AI Capabilities?
1. AISI
2. Vendor demos land inside the capability zone. Your actual problems may land just outside it.
3. category
4. Probabilities over all outcomes sum to 1
What is the recommended tip about "Ground your practice in fundamentals" in the context of The Jagged Frontier of AI Capabilities?
1. AISI
2. category
3. Every AI capability has an underlying mechanism. Understanding that mechanism tells you where it'll fail — which is more…
4. Probabilities over all outcomes sum to 1
Which statement accurately describes an aspect of The Jagged Frontier of AI Capabilities?
1. AISI
2. category
3. Probabilities over all outcomes sum to 1
4. Ethan Mollick coined 'jagged frontier' to describe the weird reality that modern AI can write a competent legal brief in 30 seconds but fail…
What does working with The Jagged Frontier of AI Capabilities typically involve?
1. The big idea: AI ability is not a single number on a scale. It is a jagged coastline.
2. AISI
3. category
4. Probabilities over all outcomes sum to 1
Which best describes the scope of "The Jagged Frontier of AI Capabilities"?
1. It is unrelated to foundations workflows
2. It focuses on AI is amazing at things that should be hard and terrible at things that should be easy. That jaggedn
3. It applies only to the opposite beginner tier
4. It was deprecated in 2024 and no longer relevant
Which section heading best belongs in a lesson about The Jagged Frontier of AI Capabilities?
1. AISI
2. category
3. Examples of the jaggedness
4. Probabilities over all outcomes sum to 1

← Back to interactive lesson

Tendril · Creators · AI Foundations

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.

35 min · Reviewed 2026

Smart Here, Blind There

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

Training data density: plentiful topics are well-learned, rare ones are not
Tokenization: character-level tasks fight the model's native unit
Reasoning depth needed: multi-step chains are fragile
Mismatch between training distribution and the specific task

How to use AI on a jagged frontier

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.
— Ethan Mollick, Co-Intelligence (2024)

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 check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-creators-jagged-frontier

What is the core idea behind "The Jagged Frontier of AI Capabilities"?
1. 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.
2. AISI
3. category
4. Probabilities over all outcomes sum to 1
Which term best describes a foundational idea in "The Jagged Frontier of AI Capabilities"?
1. capability surface
2. jagged frontier
3. tokenization
4. demo-reality gap
A learner studying The Jagged Frontier of AI Capabilities would need to understand which concept?
1. jagged frontier
2. tokenization
3. capability surface
4. demo-reality gap
Which of these is directly relevant to The Jagged Frontier of AI Capabilities?
1. jagged frontier
2. capability surface
3. demo-reality gap
4. tokenization
Which of the following is a key point about The Jagged Frontier of AI Capabilities?
1. Writes a credible sonnet but cannot reliably reverse a string
2. Solves olympiad geometry but misses simple arithmetic edges
3. Generates fluent Finnish but confabulates Finnish history
4. Codes a React app but cannot tell you which character is the 5th of a given word
Which of these does NOT belong in a discussion of The Jagged Frontier of AI Capabilities?
1. Solves olympiad geometry but misses simple arithmetic edges
2. Writes a credible sonnet but cannot reliably reverse a string
3. Generates fluent Finnish but confabulates Finnish history
4. AISI
Which statement is accurate regarding The Jagged Frontier of AI Capabilities?
1. Tokenization: character-level tasks fight the model's native unit
2. Reasoning depth needed: multi-step chains are fragile
3. Training data density: plentiful topics are well-learned, rare ones are not
4. Mismatch between training distribution and the specific task
Which of these does NOT belong in a discussion of The Jagged Frontier of AI Capabilities?
1. Tokenization: character-level tasks fight the model's native unit
2. Reasoning depth needed: multi-step chains are fragile
3. Training data density: plentiful topics are well-learned, rare ones are not
4. AISI
What is the key insight about "Not a single axis" in the context of The Jagged Frontier of AI Capabilities?
1. There is no IQ score that captures a model's ability. It is good at some things, bad at adjacent-seeming things, and the…
2. AISI
3. category
4. Probabilities over all outcomes sum to 1
What is the key insight about "The danger of demos" in the context of The Jagged Frontier of AI Capabilities?
1. AISI
2. Vendor demos land inside the capability zone. Your actual problems may land just outside it.
3. category
4. Probabilities over all outcomes sum to 1
What is the recommended tip about "Ground your practice in fundamentals" in the context of The Jagged Frontier of AI Capabilities?
1. AISI
2. category
3. Every AI capability has an underlying mechanism. Understanding that mechanism tells you where it'll fail — which is more…
4. Probabilities over all outcomes sum to 1
Which statement accurately describes an aspect of The Jagged Frontier of AI Capabilities?
1. AISI
2. category
3. Probabilities over all outcomes sum to 1
4. Ethan Mollick coined 'jagged frontier' to describe the weird reality that modern AI can write a competent legal brief in 30 seconds but fail…
What does working with The Jagged Frontier of AI Capabilities typically involve?
1. The big idea: AI ability is not a single number on a scale. It is a jagged coastline.
2. AISI
3. category
4. Probabilities over all outcomes sum to 1
Which best describes the scope of "The Jagged Frontier of AI Capabilities"?
1. It is unrelated to foundations workflows
2. It focuses on AI is amazing at things that should be hard and terrible at things that should be easy. That jaggedn
3. It applies only to the opposite beginner tier
4. It was deprecated in 2024 and no longer relevant
Which section heading best belongs in a lesson about The Jagged Frontier of AI Capabilities?
1. AISI
2. category
3. Examples of the jaggedness
4. Probabilities over all outcomes sum to 1

← Back to interactive lesson