Fact checking Moravec’s paradox
AI 摘要：This famous aphorism is neither true nor useful
AI Won’t Automatically Make Legal Services Cheaper
AI 摘要：Applying the AI as Normal Technology framework to lega...
New Paper: Towards a science of AI agent reliability
AI 摘要：Quantifying the capability-reliability gap
Open-world evaluations for measuring frontier AI capabilities
AI 摘要：Introducing CRUX, a new project for evaluating AI on l...
Do AI Risks Require Extraordinary Government Intervention?
AI 摘要：Let’s not skip the hard work of AI governance
Did Google’s AI agents really build an operating system for $916?
AI 摘要：The importance of independent evaluation
A guide to understanding AI as normal technology
AI 摘要：And a big change for this newsletter
Could AI slow science?
AI 摘要：Confronting the production-progress paradox
AGI is not a milestone
AI 摘要：There is no capability threshold that will lead to sud...
AI as Normal Technology
AI 摘要：A new paper that we will expand into our next book