← Home
Sarah Guo and Elad Gil review the year in AI, covering the biggest breakthroughs, surprises, and disappointments. They discuss the gap between AI hype and AI deployment, the emergence of AI agents, and their predictions for 2025.
This summary was generated from show notes and public descriptions, not from a full transcript review. Details may contain inaccuracies.
Canon
•
Guo argues that the shift from AI that generates text to AI that takes actions (sending emails, writing code, making purchases) is qualitatively different and creates new risk categories. A text generation error wastes seconds. An agent action error wastes money, damages relationships, or violates regulations.
Highlights
•
The gap between AI demos and AI deployment in production is the defining challenge of the industry -- demo magic does not survive contact with real-world data, edge cases, and user behavior
Gil and Guo discuss the demo-to-production gap: AI models that look impressive in controlled demonstrations often fail when deployed in production environments with messy data, adversarial users, and edge cases. The industry overestimates demo performance and underestimates production challenges.Was this useful?