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

Outline

• (15m) Providers and models
  • Overview of providers and models
  • Tradeoffs: context, speed, cost, intelligence, and capabilities
  • Your Turn _exercises/07_models: compare responses from two models
• (15m) Multimodal input
  • Images and PDFs from local files and URLs
  • Your Turn _exercises/08_vision: show images of food to an LLM
  • Your Turn _exercises/09_pdf: convert a recipe PDF to Markdown
• (15m) Structured output
  • ellmer::type_*() and Pydantic models in chatlas
  • Your Turn _exercises/10_structured-output: extract structured recipe data
• (10m) Parallel and batch calls
  • Differences between parallel requests and provider batch APIs
  • Your Turn _exercises/11_parallel: extract all recipes in parallel and open them in the recipe app