Why does your chatbot forget?

Context, temperature and cost

Tap to start →

A whiteboard

The model only sees what fits in its context window. Old text falls off.

The randomness dial

Low: same safe answer. High: more creative, more risky.

You pay per token

Input and output tokens both count. Long chats cost more.

Pick by the job

Facts and code: low temperature. Ideas: higher.

The mistake

Pasting huge documents every turn. You pay for those tokens every single time.

Can you get it?

Tap an answer. The full video explains why.

Cheat sheet

• Context window = memory
• Temperature = randomness
• Low for facts and code
• Pay per token in and out
• Shorter prompts cost less

Want the full lesson?

Free Generative AI course. Subscribe to Compiled Canvas on YouTube.