- Date: September 26th
- Time: 15:35 - 15:45
Participants need access to https://www.cursor.com/ along with an OpenAI or Anthropic API key. (It would be better if you get the PRO plan on cursor to save costs.)
For detailed information and reference materials discussed in the presentation, please see the Workshop Handout.
- Broad knowledge of how LLMs work (no need for deep technical details)
- Understanding of transformer attention, particularly autoregression
- Familiarity with tokens and context window size
- Knowledge of RLHF instruct vs "pure" completion models, and function calls encoding
- DSL design pattern (knowledge of Common Lisp or Ruby is beneficial)
- Explore Simon Willison's resources on programming with LLMs: simonwillison.net
- "Computational Logic and Human Thinking" by Robert Kowalski
- LLMs Will Fundamentally Change Software Engineering
- Centaurs and Cyborgs on the Jagged Frontier