28
Lesson 28 · 2 min
Fine-tuning vs prompting
Before you train your own AI, try a better prompt and your own documents; that solves most problems faster and cheaper.
There are three main ways to make AI fit your needs. Prompting: give instructions and examples each time. RAG: let it look up your documents. Fine-tuning: train an existing model further on your own examples.
Fine-tuning changes the model itself. It is good for a very consistent style or format, or a narrow task done millions of times. It needs good data, money and skills, and must be redone when better models arrive.
Prompting and a good system prompt cost almost nothing and can be changed in minutes. RAG is best when the problem is missing knowledge, like prices or policies.
A simple rule: start with prompting, add RAG for knowledge, and consider fine-tuning only when those clearly fall short.
Do it now
Write a reusable system prompt
- 1.Most assistants let you set custom instructions or create a custom assistant.
- 2.Fill in the template below and save it there.
- 3.Use it for a week and adjust it. That is often all the customisation you need.
You help me with [MY JOB OR TASK]. I am [SHORT DESCRIPTION]. Always: [RULES, e.g. use British spelling, keep answers under 150 words, ask before assuming]. Never: [RULES]. When unsure, say so.
Words from this lesson
- Fine-tuning
- Extra training of an existing model on your own examples so it gets better at one narrow kind of task.
- System prompt
- Background instructions set by the app maker, hidden from you, that tell the AI how to behave in every conversation.
- RAG
- Having an AI look up relevant passages from your documents first, then answer using them, so replies are more accurate.
- Prompt
- The instruction or question you give an AI. Clearer prompts with context usually get much better answers.
- Model
- The trained file of learned patterns that does the actual work inside an AI product. One model can power many apps.