Lesson 1096 of 1455
Fine-Tuning vs Prompting: When You Actually Need to Train
Most people who think they need fine-tuning just need better prompts and a few examples. Real fine-tuning is rare.
Builders · Model Families · ~4 min read
The big idea
Fine-tuning means actually retraining a model on your data so it learns your style or domain. It's expensive and slow. 90% of the time, few-shot prompting (giving examples in the prompt) gets you the same result without training. Reach for fine-tuning only when prompts can't reach the quality you need at scale.
Some examples
- A startup using GPT-5 with 5 example outputs in every prompt — prompting wins.
- A medical company training Llama on their own clinical notes — real fine-tuning.
- A side project that wants the AI to talk like a pirate — system prompt, not training.
- A legal-tech tool needing exact citation format on millions of queries — fine-tune for that one task.
Try it!
Take a task you wished an AI did better. List 5 strong examples in your prompt. See if you actually needed to fine-tune.
Key terms in this lesson
End-of-lesson quiz
Check what stuck
8 questions · Score saves to your progress.
Lesson help
Questions are best handled with a grown-up here.
For this age range, Tendril keeps freeform AI chat paused until parent/guardian consent and child-safe moderation are fully verified. Use the quiz, notes, and related lessons below, or ask a parent, guardian, teacher, or librarian to work through the question with you.
Progress saved locally in this browser. Sign in to sync across devices.
Related lessons
Keep going
Builders · 40 min
Claude vs ChatGPT for Teens: Quick Comparison
Both are great chatbots but they have different vibes. Knowing which to pick saves time.
Builders · 40 min
Context Windows: How Much AI Can 'Remember'
Each AI has a 'context window' — how much it can hold in memory. Knowing this matters for big tasks.
Builders · 40 min
AI model families: Meta's Llama (open source)
Understand why Llama matters as a free, open AI model anyone can run.