This paper discusses a new method of teaching programming using Large Language Models (LLMs) and Prompt Problems. LLMs have revolutionized computing education, enabling students to learn coding through code generation tools. The new teaching method involves students receiving a problem visually and translating that into a prompt for an LLM to decipher. The problem is considered solved when the student-generated code passes all test cases. The paper presents the design of this tool and discusses student interactions with it, offering insights into this new class of programming problems and the design tools that integrate LLMs.
Publication date: 19 Jan 2024
Project Page: https://doi.org/11.1111/1111111.1111111
Paper: https://arxiv.org/pdf/2401.10759
