Raymond Perrault on language models, scaling laws, reasoning, multimodal models, and ethical AI.
Raymond Perrault is a Distinguished Computer Scientist at SRI International, and Co-Director of the Steering Committee for the AI Index, an annual report that tracks, collates, distills, and visualizes data relating to AI, to help inform decision-makers and teams to take meaningful action for responsible and ethical AI. We discussed key findings from the 2023 AI Index, and how they relate to machine learning and AI trends.
Interview highlights – key sections from the video version:
- Description of the AI Index: data sources, goals, target audience
- Research output by geography
- Scaling laws and trends in model sizes
- Hybrid models (neural + information retrieval)
- Performance saturation on traditional benchmarks
- AI systems become more flexible
- Reasoning and Large Language Models
- AI starts to build better AI (reinforcement learning & more)
- Why Fairer models may not be less biased
- State of AI Ethics
- Speculating on what cutting edge models will look like in the near future
- A video version of this conversation is available on our YouTube channel.
- Jack Clark: The 2022 AI Index
- Jakub Zavrel: Uncovering and Highlighting AI Trends
- Percy Liang: Evaluating Language Models
- Dylan Patel: The Open Source Stack Unleashing a Game-Changing AI Hardware Shift
- Pablo Villalobos: Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades
- Roy Schwartz: Efficient Methods for Natural Language Processing
- Connor Leahy and Yoav Shoham: Large Language Models
- Mark Chen of OpenAI: How DALL·E works
If you enjoyed this episode, please support our work by encouraging your friends and colleagues to subscribe to our newsletter: