Machine Learning Street Talk (MLST) #81 JULIAN TOGELIUS, Prof. KEN STANLEY - AGI, Games, Diversity & Creativity [UNPLUGGED]
14 snips
Nov 20, 2022 
Julian Togelius, an NYU Associate Professor and co-founder of model.ai, teams up with Ken Stanley from OpenAI, who leads research on open-endedness. The duo dives into the intersection of AI and games, emphasizing how games serve as a testing ground for AGI. They discuss the need for diversity in AI, tackling challenges in integrating different learning approaches. The conversation also touches on the balance between creativity and technology, and the philosophical debates shaping AI's future.
AI Snips
Chapters
Books
Transcript
Episode notes
Challenges in AI and Game Development
- Julian Togelius's dream of symbiotic AI and game development faced challenges in practical application.
- Game testing and procedural content generation benefited from AI, but creating generally intelligent agents for better gameplay proved more complex.
Limitations of AlphaZero
- AlphaZero, while impressive, isn't truly intelligent due to its narrow, domain-specific nature.
- Formal definitions of intelligence can be inspirational but won't fully capture the multifaceted nature of intelligence.
Critique of AGI
- The term "Artificial General Intelligence" (AGI) is problematic, especially due to its association with the flawed superintelligence argument.
- AGI’s focus on broader capabilities beyond specific tasks is valuable, but true generality in intelligence is likely unattainable.
