The text discusses the shift in the approach to AI from data-centric models to the importance of expert prompts. While previous AI models relied on large amounts of data, the rise of Large Language Models (LLMs) has highlighted the significance of prompts that encode expertise. The author emphasizes the need for prompts that go beyond simple explanations and instead engage learners in a meaningful way. The text provides examples from education, where prompts are designed to create a tailored learning experience for students. The author encourages the development of prompt libraries that can be freely accessed and used by individuals and organizations.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| Rise of Large Language Models (LLMs) | Shift from data-centric AI to expertise-centric AI | More emphasis on expert prompts and spellbooks | Desire for more accurate and tailored AI outputs |
| Importance of expert prompts | Focus on creating prompts that encode expertise | Increased development of prompt libraries and spellbooks | Need for personalized and effective AI interactions |
| Building effective prompts | Emphasis on expertise, experience with AI models, and clear goals | Improved AI interactions and tailored feedback | Desire for better AI performance and user experience |
| Grimoires over data | Shift from data-centric to expertise-centric AI approach | Creation of prompt libraries for scaling practices | Empowerment of individual workers and innovation |
| Large-scale public libraries of prompts | Availability of expert-written prompts for public use | Promoting innovation and learning through LLMs | Grassroots prompting effort and ethical discussions |