Business Pill · How much a model can hold in view
A model can read a long document and still get a question about it wrong. The reason is the context window: the amount of text a model can keep in view at one time.
A short explainer video, under a minute. It uses a story about a robot at a helpdesk and a 300-page contract.
At a glance
What the Context Window Is
The context window is the amount of text a model can keep in view at one time.
Everything Shares the Space
Everything shares that space: your instructions, the documents, the whole conversation, and the answer itself.
Why It Matters
A model only knows what is on its desk.
The Short Answer
A customer hands a robot a 300-page contract. The robot reads it all. Then she asks about a clause on page two, and the robot gets it wrong.
The robot’s desk only holds so many pages. When new ones arrive, the oldest fall off the edge. That desk has a name. It is called the context window.
What the Context Window Is
The context window is the amount of text a model can keep in view at one time.
In the story, the window is the robot’s desk. It has a fixed size, so something has to give when new pages arrive.
Everything Shares the Space
Everything shares that space: your instructions, the documents, the whole conversation, and the answer itself.
So a long document does not get the whole desk to itself. It sits next to everything else the model is holding.
The Idea on One Board
Everything shares the space
Bigger Windows Are Not a Free Fix
Windows keep getting bigger. But every page on the desk is paid for on every answer, and a crowded desk is harder to read.
More room helps, and it also adds cost and clutter. A bigger desk does not remove the need to choose what goes on it.
Why It Matters
A model only knows what is on its desk. So the choice of what goes there is a choice about what the model can answer.
The video’s question is what really needs to be on the desk.
The key insight: A model only knows what is on its desk. So the choice of what goes there is a choice about what the model can answer.
The honest limit
Windows keep getting bigger. But every page on the desk is paid for on every answer, and a crowded desk is harder to read.
More room helps, and it also adds cost and clutter. A bigger desk does not remove the need to choose what goes on it.
Where This Shows Up in AI Business Today
From our news coverage
Ai2 Open-Sources AstaBrief: 8B Model, 3.5x Faster Reports
Our report on Ai2’s AstaBrief says Ai2 released an 8B open-weights model that turns a research question and retrieved literature excerpts into a cited scientific report. What gets handed to a model in that way is what this lesson calls the desk.
The Business Engineer Lens
Where this idea sits in the Business Engineer library:
The One Question to Ask
- What really needs to be on the desk?
How This Concept Connects
Layer 3 of 8 · Inference & Context, on the Business Engineer Concept Map
- Transformers and Attention constrains Context Window. Attention weighs every word against every other, so doubling the text roughly quadruples the work a longer window must pay for.
- Context Window builds on Tokens. Instructions, documents, conversation and the answer all share the window, and all of it is counted in tokens.
- Context Window drives the cost of Inference. Every page on the desk is paid for on every answer, so a fuller window raises the cost of each inference call.
- Context Engineering builds on Context Window. Context engineering decides what goes on the desk, the fixed-size space the context window lesson describes.
- Multi-Agent Systems builds on Context Window. One agent’s desk fills and it forgets the first suppliers; splitting the work gives each helper a clean desk.
- Context, Not Capability builds on Context Window. Context has a limit: everything the model can consider must fit the window, so choosing what goes in is the work.
The Economics of an AI Prompt
Window contents are billed on every answer.
Get the full Concept Map as a free poster.
Related Business Pills
All Business Pills
- The Memory Wall: Why a Faster AI Chip Is Not Faster AI
- Tokens per Watt: What an AI Data Centre Actually Produces
- Why AI Can’t Be Both Instant and Cheap: Latency vs Throughput
- The Model and the Harness: Why Same-Model Products Differ
- Context, Not Capability: Why a Smart AI Model Gives Poor Answers
- Discardable Software: When Code Is Cheap Enough to Throw Away
- Human in the Loop vs Human on the Loop: Supervising AI Agents
- The AI Audit Problem: When Making Work Is Cheaper Than Checking It
- AI Evaluation as Acceptance Test: How to Know It’s Good Enough
- Extensibility Is Control: Who Holds the Power in an AI Product
- RLHF Explained: How an AI Model Learns What People Prefer
- Pretraining Explained: How an AI Model Learns Before Anyone Teaches It
- Fine-Tuning Explained: How to Adapt a General AI Model to One Job
- Tokens Explained: The Unit AI Reads, Writes and Bills In
- Embeddings Explained: How a Machine Compares Meaning
- RAG Explained: How a Model Answers From Your Documents
- AI Hallucination Explained: Why Models State False Things
- Distillation Explained: How a Small Model Learns From a Large One
- Reasoning Models Explained: What Changes When AI Thinks First
- Tool Use Explained: How an AI Model Goes From Text to Action
- Capex and Depreciation Explained: Why Chip Lifetime Drives AI Profits
- Run-Rate Revenue Explained: What an AI Company’s Number Means
- Backlog Explained: Revenue That Is Signed but Not Yet Earned
- Switching Costs Explained: Why It Is Hard to Leave an AI Supplier
- Temperature Explained: The Dial That Sets How Predictable an AI Model Is
- Guardrails Explained: Rules Enforced Around an AI Model, Not Inside It
- Benchmarks Explained: Why a Test Score Is Not Your Own Result
- Agent Memory Explained: How an AI Assistant Remembers You Between Chats
- MCP Explained: The Shared Plug That Connects AI Models to Tools
- Quantization Explained: How a Large AI Model Is Shrunk to Fit
- Parameters Explained: Where an AI Model’s Knowledge Lives
- Scaling Laws Explained: Why AI Labs Keep Building Bigger
- Synthetic Data Explained: When a Model Writes Its Own Training Examples
- Transformers and Attention Explained: How AI Reads Every Word at Once
- Prompt Injection Explained: When Text a Model Reads Gives the Orders
- Inference Explained: The Cost Paid on Every AI Answer
- Multimodal Explained: AI Models That Work With Images, Sound and Video
- AI Agents Explained: From Answering to Doing









