Reimagining service delivery in the agentic era with Google Public Sector
cloud.google.com·21h ago
Data Science
Simon Willison's Blog·Jun 10·5 min read
datasette-agent 0.2a0
TL;DR
Software engineers faced challenges in efficiently interacting with large language models (LLMs) for user queries. Simon Willison developed the ask_user() feature using a new LLM alpha model.
✦ Why It Matters
Engineers can implement the ask_user() feature to improve user interaction with LLMs in their applications.
Key Takeaways
- Implement ask_user() to enhance user interaction in your applications.
- Utilize the save_query tool for storing SQL queries with user consent.
- Ensure conversations persist through server restarts for better user experience.
- Explore the new LLM alpha for improved agent capabilities.
How It Works
The ask_user() feature allows tools to pause execution and ask users questions, utilizing a ToolContext object. This interaction is rendered as a form in the chat UI, and the conversation state is saved in the internal database, ensuring continuity even after server restarts.