9
Phase 9 · ongoing
Using AI in your apps
The fun, useful part of AI — and the part you can actually do now. The goal here is to understand how these tools work and add an AI feature to your own app. The deep "build AI from scratch" stuff is a separate, much-later road.
What to learn
- First, understand the basics in plain words: what an LLM is (= a big AI model that reads and writes text, like ChatGPT or Claude), what a token is (= the small chunks of text an LLM reads and writes, a bit like syllables), and roughly how it works (it predicts the next token, over and over).
- What these models are good at and where they go wrong — why they sometimes confidently make things up (hallucinate), and why clear instructions matter.
- Prompt engineering (= writing good instructions so the AI gives good answers). Small wording changes make a big difference.
- How to call an AI from your code: send a prompt to an LLM’s API and show the answer in your app — for example, add a simple chatbot or “summarise this” button to your Phase 8 project.
Want to go deeper into building AI itself — training your own models, neural networks, the maths? That’s a real path, but it comes much later and needs the Phase 3 maths first. For now, focus on using AI well. The optional links below are there for when you’re ready.
Videos to watch
Practice on your own
- Take one task and write it 5 different ways as prompts — compare how the answers change.
- Get an API key from a free/cheap LLM and make a tiny script that sends a prompt and prints the reply.
- Add one small AI feature to your Phase 8 app — a chatbot, or a “summarise this” button.
Build this
Add a real AI feature to your Phase 8 app — something a user can actually click and use. That’s the goal for now. (Building and training your own models is a later, optional road.)
Move on when: You can explain in simple words what an LLM and a token are, and you’ve shipped one working AI feature in an app.