

Six: Richard Ngo on large language models, OpenAI, and striving to make the future go well
Sep 2, 2023
This podcast explores the understanding and potential risks of large language models like GPT-3 and ChatGPT. Richard Ngo from OpenAI discusses AI governance, concerns surrounding these models, and the challenges of AI behavior prediction. They also delve into the development of general AI, situational awareness in AI systems, and the need to study and modify goal formation in neural networks. The podcast concludes with discussions on the challenges of understanding AI behaviors, exploring utopia and the role of technology, and alternative history thought experiments.
Chapters
Transcript
Episode notes
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Introduction
00:00 • 4min
AI Governance and the Alignment Problem
03:41 • 9min
Comparing Governance Arrangements for Nuclear Weapons and Computer Chips
12:20 • 2min
Concerns Surrounding Large Language Models
13:58 • 18min
Concerns and Misconceptions about AI Behavior
31:35 • 3min
OpenAI's Work and Goals
34:41 • 18min
Exploring the Vast Space of Language Model Prompts
53:08 • 16min
Misconceptions and Insights into Machine Learning Models
01:09:17 • 4min
Development of General AI and Lab Priorities
01:13:18 • 13min
Situational Awareness in AI
01:26:02 • 10min
Understanding and Modifying Goal Formation in Neural Networks
01:35:59 • 14min
The Challenges of Understanding and Predicting AI Behaviors
01:49:31 • 23min
Exploring Conceptual Alignment Research on Defining Goals in Neural Networks
02:12:29 • 2min
Discussing Existing Books on the Alignment Problem
02:14:08 • 4min
Taking Action and Transitioning to Producing Knowledge
02:17:43 • 11min
Exploring Utopia and the Role of Technology
02:28:37 • 9min
Alternative History Thought Experiments and Contingencies in Evolution
02:37:22 • 2min
Science Fiction Books, Thought Experiments, and AGI Safety Fundamentals
02:39:28 • 5min