

Will AI Save Humanity or End It? with Mustafa Suleyman
312 snips Sep 18, 2025
In this engaging conversation, Mustafa Suleyman, co-founder of DeepMind and head of AI at Microsoft, explores the dual-edged nature of artificial intelligence. He discusses the origins of DeepMind and the complexities of how models learn. Roberto highlights the need for humanist superintelligence and the ethical dilemmas that arise. He predicts AI breakthroughs could reduce energy costs, but warns of job displacement. Suleyman argues for robust governance to ensure AI benefits humanity while addressing deepfake challenges and social stability. He envisions a future where AI fosters abundance and meaningful work.
AI Snips
Chapters
Books
Transcript
Episode notes
Information Is The Core Of AI
- AI learns the structure of information by predicting relationships across data modalities.
- Scaling data and compute lets the same mechanism generate images, audio, text, and code.
Behavior Over Inner States
- Human judgments of
Start With Human Values
- Prioritize human wellbeing when building AI and measure net effects on suffering.
- Design toward humanist superintelligence that serves and protects human interests.