hslu-ki-h2501

Part 2 — Human Learning: The Failed Animal

Learning materials

• 📘 NotebookLM: Human Learning — The Failed Animal:
  • Video overview
  • Notebook
• 🎤 Podcast: Charles Pépin — The Virtues of Failure:
  • Official page
  • Audio: French / English
  • English Transcript

2.1 Conceptual Introduction — Unnatural Animals

Humans differ from other animals through their lack of instinct and dependence on learning. Where most species act effectively from birth, humans begin in weakness and failure — yet this deficiency becomes the foundation of our creativity and freedom.

Guiding question:.
If humans learn through imperfection and effort, how does that differ from the efficiency of AI learning?

2.2 The Failed Animal — From Instinct to Learning

• Key idea: Humans are “denatured animals” (Charles Pépin) — our instincts are too weak to guide us.
• Example: A foal walks within minutes; a baby fails about 2,000 times before succeeding.
• Insight: Human learning emerges through trial, error, and perseverance, not pre-programmed success.

Activity idea 1 — The Foal vs. the Baby.
Compare the instinctive efficiency of the foal with the slow, failure-ridden learning of the baby.
→ Discuss: If AI is designed for speed and optimization, what do human failures reveal about the nature of deep learning?

2.3 The Methodology of Failure — Learning by Inefficiency

• Key idea: Because we fail, we improvise, adjust, and discover new paths.
• Humans learn through tâtonnement (“trial and error”) and correction — a dynamic process that builds adaptability.
• Contrast: AI minimizes error statistically; humans extract meaning from error existentially.

Activity idea 2 — Virtues in Failure.
Groups brainstorm what kinds of knowledge or insight arise only through failure.
→ Compare: What does AI miss if it never “feels” the cost of failure?

2.4 Compensation and Culture — Learning from Others

• Key idea: Our weak instincts force us to depend on culture, education, and relationships.
• Humans compensate for biological limits by learning socially — through teaching, imitation, empathy, and love.
• Contrast: AI absorbs data but lacks emotional and relational context.

Activity idea 3 — The Compensation Audit.
Classify human compensations into:

1. Data / Information (e.g., reading, tutorials)
2. Relational / Social (e.g., empathy, support)

→ Discuss: Can AI replace the second category, or is social learning uniquely human?

2.5 Freedom and Invention — The Outcome of Failure

• Key idea: Because we are “failed animals,” we are not bound by instinct — we can invent ourselves.
• Freedom arises from this indeterminacy: humans are defined by possibility, not programming.
• Philosophical link: Existence precedes essence (Sartre).

Activity idea 4 — The Freedom Inventory.
Reflect: If AI’s goals and functions are fixed, can it experience human-like freedom or creativity born from uncertainty?
→ Debate whether freedom itself depends on being “imperfect.”

2.6 Synthesis — AI vs. Human Learning

Aspect	Human Learning	AI Learning
Origin	Weak instincts → learning through trial and error	Strong optimization → learning through data efficiency
Method	Experiential, social, embodied	Computational, statistical
Relation to Error	Failure is essential and meaningful	Error is minimized and avoided
Outcome	Freedom and self-invention	Performance and prediction

Takeaway: Humans learn because they fail; AI learns to avoid failure.
Our weakness is the source of our adaptability, creativity, and freedom.