Last quiz of AI Beginner. Ten questions covering multi-turn conversation, system prompts, token awareness, refusal detection, and the 5-primitive synthesis. 80% to pass.
When you pass, you have the AI primitives kit. Calling, prompt principles, the four task verbs, structured output, multi-turn, system prompts, cost awareness, refusal handling. Whatever LLM script you read or write from here, you'll recognise the shape.
| Lesson | Concept |
|---|---|
| 15 | Conversation as a list — pass message_history to run_sync |
| 16 | Multi-turn context — accumulate history across N turns |
| 17 | System prompt — persistent rules across the whole conversation |
| 18 | Token cost — result.usage().total_tokens |
| 19 | Refusals — heuristic detection + route around |
| 20 | Synthesis — system prompt + multi-turn + JSON + parse + assert |
Agent(model).run_sync(prompt).outputtry/except json.JSONDecodeError + retryusage().total_tokens to see costAI Intermediate (deferred for v1) adds: tool calling, chain-of-thought, output validation, eval suites, agent loops. AI Advanced adds: embeddings, RAG, model routing, caching, production guardrails. Auto Advanced adds: webhooks, observability, queues. All of them sit on top of these primitives.
Create a free account to get started. Paid plans unlock all tracks.
Last quiz of AI Beginner. Ten questions covering multi-turn conversation, system prompts, token awareness, refusal detection, and the 5-primitive synthesis. 80% to pass.
When you pass, you have the AI primitives kit. Calling, prompt principles, the four task verbs, structured output, multi-turn, system prompts, cost awareness, refusal handling. Whatever LLM script you read or write from here, you'll recognise the shape.
| Lesson | Concept |
|---|---|
| 15 | Conversation as a list — pass message_history to run_sync |
| 16 | Multi-turn context — accumulate history across N turns |
| 17 | System prompt — persistent rules across the whole conversation |
| 18 | Token cost — result.usage().total_tokens |
| 19 | Refusals — heuristic detection + route around |
| 20 | Synthesis — system prompt + multi-turn + JSON + parse + assert |
Agent(model).run_sync(prompt).outputtry/except json.JSONDecodeError + retryusage().total_tokens to see costAI Intermediate (deferred for v1) adds: tool calling, chain-of-thought, output validation, eval suites, agent loops. AI Advanced adds: embeddings, RAG, model routing, caching, production guardrails. Auto Advanced adds: webhooks, observability, queues. All of them sit on top of these primitives.