[[!meta title="From FRDCSA to FLP2: Building AI-Powered Life Planning Systems in Emacs - A Journey from Research to Real-World Impact"]] [[!meta copyright="Copyright © 2025 Andrew John Dougherty"]] [[!inline pages="internal(2025/info/life-nav)" raw="yes"]] # From FRDCSA to FLP2: Building AI-Powered Life Planning Systems in Emacs - A Journey from Research to Real-World Impact Andrew John Dougherty (he/him) - Pronunciation: Andrew DOW-er-tee, IRC: aindilis, Website: GitHub: , [[!inline pages="internal(2025/info/life-before)" raw="yes"]] Six years after my EmacsConf 2019 presentation on the Formalized Research Database (FRDCSA), I return to share the evolution of this ambitious project into the Free Life Planner (FLP) - a comprehensive AI-powered system designed to help people manage the complex logistics of daily life, from basic survival to thriving in challenging circumstances. This talk chronicles the development journey from FRDCSA's initial vision of collecting and integrating all free software AI systems, to FLP's practical focus on solving real-world problems like homelessness, medical crises, and financial hardship. I'll demonstrate how Emacs serves as both the development environment and user interface for a system that now encompasses over 1000 subsystems including meal planning, financial forecasting, medical management, and emergency preparedness. Key highlights include: Technical Architecture: How we've built FLP using SWI-Prolog as the core reasoning engine, with Emacs providing the primary interface through 4680+ custom ELisp functions. I'll show the integration of modern LLMs (Llama3, DeepSeek-Prover-V2, etc) with classical AI planning systems (PDDL planners, behavior trees, agent-based systems) all orchestrated through Emacs. Security Evolution: The transition from the original internet-connected FRDCSA/FLP to air-gapped systems, and the development of FRKCSA/FLP2 as open-source, internet-facing reboots. How we balance powerful AI capabilities with privacy and security concerns. Real-World Impact: Using generative AI and our PresGen presentation system, I'll showcase documented case studies where FLP has helped users navigate homelessness, medical crises, and benefit loss scenarios. These aren't hypothetical use cases - they represent the system's practical application to urgent human needs. Emacs Integration: Deep dive into our Emacs-based development workflow, including the CLEAR book reader for academic research, Verber planning domain editor, UniLang integration for multi-system communication, and our Academician mode for synchronized document processing. How Emacs serves as mission control for a complex AI ecosystem. AI Planning Systems: Demonstration of how we've integrated automated planning, temporal reasoning, and contingency planning into daily life management. Users can ask "How do I get through the next month if I lose my job?" and receive detailed, executable plans. Community Building: How FLP facilitates mutual aid networks and community resilience, turning individual survival into collective thriving. The system doesn't just help people - it helps them help each other. Future Directions: The roadmap for FLP2, including improved security models, enhanced AI integration, and broader accessibility. How we're preparing for a world where AI-powered life assistance could become a virtual social safety net. This presentation will use our PresGen system to generate slides dynamically, incorporating the latest AI developments and real user feedback. Attendees will see both the technical implementation and the humanitarian vision that drives this project - demonstrating how Emacs can be the foundation for transformative social technology. About the speaker: Andrew Dougherty is the creator of the FRDCSA (Formalized Research Database: Cluster, Study and Apply) and Free Life Planner projects. For over 25 years, he has worked on integrating AI systems to solve real-world problems, with a focus on helping vulnerable populations navigate complex life challenges. His work spans automated planning, knowledge representation, and practical AI applications. Andrew presented FRDCSA at EmacsConf 2019 and has continued developing these systems with Emacs as the primary development and deployment environment. He believes in using AI technology to strengthen communities and provide practical assistance to those who need it most. [[!inline pages="internal(2025/info/life-after)" raw="yes"]] [[!inline pages="internal(2025/info/life-nav)" raw="yes"]]