On this episode of the Crazy Wisdom Podcast, I, Stewart Alsop, sit down with Brendon Wong, the founder of Unize.org. We explore Brendon's work in knowledge management, touching on his recent talk at Nodes 2024 about using AI to generate knowledge graphs and trends in the field. Our conversation covers the evolution of personal and organizational knowledge management, the future of object-oriented systems, the integration of AI with knowledge graphs, and the challenges of autonomous agents. For more on Brendon’s work, check out unize.org and his articles at web10.ai.

Check out this GPT we trained on the conversation!

Timestamps

00:00 Introduction to the Crazy Wisdom Podcast

00:35 Exploring Unise: A Knowledge Management App

01:01 The Evolution of Knowledge Management

02:32 Personal Knowledge Management Trends

03:10 Object-Oriented Knowledge Management

05:27 The Future of Knowledge Graphs and AI

10:37 Challenges in Simulating the Human Mind

22:04 Knowledge Management in Organizations

26:57 The Role of Autonomous Agents

30:00 Personal Experiences with Sleep Aids

30:07 Unique Human Perceptions

32:08 Knowledge Management Journey

33:31 Personal Knowledge Management Systems

34:36 Challenges in Knowledge Management

35:26 Future of Knowledge Management with AI

36:29 Melatonin and Sleep Patterns

37:30 AI and the Future of the Internet

43:39 Reasoning and AI Limitations

48:33 The Future of AI and Human Reasoning

52:43 Conclusion and Contact Information

Key Insights

1. The Evolution of Knowledge Management: Brendon Wong highlights how knowledge management has evolved from personal note-taking systems to sophisticated, object-oriented models. He emphasizes the shift from traditional page-based structures, like those in Roam Research and Notion, to systems that treat information as interconnected objects with defined types and properties, enhancing both personal and organizational knowledge workflows.
2. The Future Lies in Object-Oriented Knowledge Systems: Brendon introduces the concept of object-oriented knowledge management, where data is organized as distinct objects (e.g., books, restaurants, ideas) with specific attributes and relationships. This approach enables more dynamic organization, easier data retrieval, and better contextual understanding, setting the stage for future advancements in knowledge-based applications.
3. AI and Knowledge Graphs Are a Powerful Combination: Brendon discusses the synergy between AI and knowledge graphs, explaining how AI can generate, maintain, and interact with complex knowledge structures. This integration enhances memory, reasoning, and information retrieval capabilities, allowing AI systems to support more nuanced and context-aware decision-making processes.
4. The Limitations of Current AI Models: While AI models like LLMs have impressive capabilities, Brendon points out their limitations, particularly in reasoning and long-term memory. He notes that current models excel at pattern recognition but struggle with higher-level reasoning tasks, often producing hallucinations when faced with unfamiliar or niche topics.
5. Challenges in Organizational Knowledge Management: Brendon and Stewart discuss the persistent challenges of implementing knowledge management in organizations. Despite its critical role, knowledge management is often underappreciated and the first to be cut during budget reductions. The conversation highlights the need for systems that are both intuitive and capable of reducing the manual burden on users.
6. The Potential and Pitfalls of Autonomous Agents: The episode explores the growing interest in autonomous and semi-autonomous agents powered by AI. While these agents can perform tasks with minimal human intervention, Brendon notes that the technology is still in its infancy, with limited real-world applications and significant room for improvement, particularly in reliability and task generalization.
7. Reimagining the Future of the Internet with Web 10: Brendon shares his vision for Web 10, an ambitious rethinking of the internet where knowledge is better structured, verified, and interconnected. This future internet would address current issues like misinformation and data fragmentation, creating a more reliable and meaningful digital ecosystem powered by AI-driven knowledge graphs.