How AI can help us share intuitive knowledge | Dan Shipper

Episode Overview

• This episode explores the historic limitation of sharing "intuitive" or tacit knowledge, which cannot be easily written down or codified.
• It highlights how society has historically progressed by prioritizing explicit, written, and scientific explanations because they are easy to spread.
• It explains how neural networks and AI offer a paradigm shift by allowing us to capture, package, and share human intuition.
• It frames the potential impact of this technology through the lens of medicine, showing how AI can democratize access to the world's best, unwritten clinical expertise.

Key Concepts

• The Explicit Knowledge Bottleneck: Human progress has historically been limited to what can be documented. When knowledge cannot be explicitly written down, collaborating on it, scaling it, or making collective progress becomes incredibly difficult.
• AI as a Medium for Intuitive Knowledge: While traditional computer programs require explicit rules, neural networks learn from patterns. This allows them to capture the "intuitive experience" that humans build up through practice, effectively digitizing tacit knowledge.
• Democratizing Uncodifiable Expertise: Many expert fields, like clinical medicine, rely on intuitive judgments that experts cannot easily explain or turn into a set of rules. AI can capture these complex, sub-conscious decision-making patterns, allowing anyone in the world to access top-tier expertise.

Quotes

• At 0:05 - "because if you can write it down, you can spread it. And that becomes... the way that society makes progress is sort of like these spreading explanations." - Explaining why humanity has historically favored explicit, scientific definitions over intuitive knowledge.
• At 0:25 - "And what neural networks allow us to do is to take some of that... intuitive experience or intuition that we might have built up ourselves and put it into a machine that we can pass around." - Explaining the core technological breakthrough that AI brings to human knowledge sharing.
• At 0:43 - "The best clinicians in the world know something about how to deal with patients that they can't write down, they can't embody in a set of rules... But language models and AI in general allows us to put that kind of intuition into a tool." - Illustrating how AI can solve the problem of "trapped" expert knowledge in critical fields like healthcare.

Takeaways

• Shift your documentation strategy: Instead of trying to write rigid, rule-based manuals for complex, intuitive tasks, focus on gathering diverse data and examples that can train AI models to mimic expert decision-making.
• Identify "trapped" expertise within your organization: Look for highly skilled roles where performance relies on "gut feeling" or unwritten experience, and target these areas for AI integration to help scale that tacit knowledge.
• Prepare for democratized expertise: Position your organization to leverage AI tools that grant access to top-tier, specialized intuition, shifting your focus from acquiring scarce expert knowledge to effectively applying it.