Unsupervised Learning with Jacob Effron Ep 58: Google Researchers Noam Shazeer and Jack Rae on Scaling Test-time Compute, Reactions to Ilya & AGI
145 snips
Mar 17, 2025 
Noam Shazeer, co-inventor of the Transformer, and Jack Rae, Research Director at DeepMind, dive into the future of AI. They discuss the groundbreaking capabilities of Gemini 2.0 for reasoning and creativity. The duo explores the complexities of AI evaluation metrics and the evolving role of test-time compute, emphasizing efficiency over traditional methods. They also reflect on the philosophical challenges of AGI, the rise of vision-based models, and AI's transformative impact on healthcare and education, highlighting the balance between innovation and safety.
AI Snips
Chapters
Transcript
Episode notes
Reinforcement Loops
- Reinforcement loops are key for AI development, particularly using AI for increased productivity.
- Data flywheels and overall excitement further accelerate this process, leading to rapid advancements.
Mom Vibe Check
- Jack Rae's mother used the Gemini app, even asking philosophical questions like the meaning of life.
- She appreciated the model's thought process, demonstrating the model's potential for broader audiences.
Test-Time Compute Scaling
- Test-time compute has a high ceiling due to the low cost of inference compared to other tasks.
- This allows for using more compute to improve model performance, pushing current scaling curves further.