Programming with LLMs

A deeper dive into the things you can do with LLMs when you’re programming with them that are harder to do in a chat interface.

Slides

View slides in full screen

Resources

• (15m) Choosing a model
  • Overview of major providers: OpenAI, Anthropic, Google, ollama
  • Tradeoffs: capability, context length, speed, cost, intelligence
  • Activity: same question, change one string to switch models, e.g. chat("openai"), chat("anthropic").
• (15m) Multi-modal input (vision, PDF)
  • Activity: images of food and ask for recipes
  • Activity: take a PDF of a recipe, turn it into markdown
• (15m) Structured output
  • Explain ellmer::type_*() or pydantic model in chatlas
  • Activity: Extract rich data from the recipe PDF
  • Note use_attribute_docstrings
• (10m) Parallel/batch calls
  • Activity: Extract recipe data in parallel or batch