Business Pill 24 · Adapting a model to one job
Fine-tuning is extra training on your own examples. It can make a smaller model good at one job, but it is rarely the first thing to try.
A short explainer video. The numbers in it are round numbers for illustration.
The Short Answer
A general model is like a well-educated graduate. It knows a little about everything, but it does not know your products, your tone, or your way of doing things.
Fine-tuning is extra training on your own examples. You collect a few thousand cases of the task done well. The model studies them, and its internal settings shift slightly toward your way of doing it.
What You Gain
The gains are real. The model follows your format without long instructions. It handles your special cases. And often a small fine-tuned model can match a large general one on that task, at a fraction of the cost.
What It Costs
You need good examples, and those take work to collect. When a new base model arrives, you have to tune again. And the model becomes narrower: better at your task, worse at others.
When to Use It
Start simpler first. Write better instructions. Then give the model the right documents at the moment of the request. Fine-tune only when the task is stable and high in volume.
Do not train what you can simply tell. Fine-tune when telling is no longer enough.
Three Questions to Ask
- Is the task stable?
- Do we have enough good examples?
- Have we tried the simpler options first?
More 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
- 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









