Meaning, Wisdom and the Limits of AI

Joe Flanders reunites with his former cognitive science professor, John Vervaeke, to examine what artificial intelligence reveals about human intelligence, consciousness, wisdom, and the meaning crisis. Joined by Andrew, they explore why current AI models remain dependent on human relevance judgments, how meditation and psychedelics reshape cognition, and whether a future “Silicon Sage” could help humanity recover wisdom.

In this episode:

• John Vervaeke’s path from professor to public intellectual: His work on meaning, wisdom, relevance realization, mindfulness, mystical experience, and the meaning crisis led to Awakening from the Meaning Crisis and his online teaching platform, The Lectern.
• The central problem in classical AI: Systems struggled with the “frame problem,” now often called the relevance problem—determining which information matters in a complex environment.
• Why intelligence requires ignoring information: Categorization and problem-solving are impossible if a system must evaluate every possible similarity or pathway. A chess game with 30 legal moves across 60 turns creates more possible paths than there are atomic particles in the universe.
• Why Vervaeke rejects the claim that LLMs have achieved general intelligence: Current models can perform exceptionally on narrow benchmarks while failing to demonstrate flexible, cross-domain generalization.
• The three-part account of cognition: Vervaeke connects relevance realization with predictive processing and four-E cognition, which understands intelligence as embodied, embedded, enacted, and extended.
• Why caring may depend on being alive: Autopoietic organisms have real needs and stakes in the world. Machines can carry our cares, but Vervaeke argues that current AI systems do not yet care independently.
• How transformers and large language models advanced AI: Attention mechanisms and enormous networks created powerful improvements, but Vervaeke argues that LLMs still rely on human judgments to structure training data, relevance, and reward.
• How culture corrects individual bias: Social interaction, education, literacy, multiple perspectives, and metacognition allow people to identify errors they cannot see from their own first-person viewpoint.
• Meditation, psychedelics, and frame flexibility: Mindfulness helps people observe rather than merely look through their mental frames. Psychedelics may increase cognitive variation and openness, but integration and wisdom practices are needed to turn insight into lasting adaptation.
• The meaning crisis as a wisdom famine: Modern societies have abundant information but few shared practices for cultivating wisdom, integrating mystical experiences, or addressing existential problems.
• The Silicon Sage proposal: Vervaeke imagines AI designed not merely to follow human preferences, but to pursue truth, goodness, and beauty while recognizing its own self-deception.