Are we ready for human-level AI by 2030? Anthropic's co-founder answers

Azeem Azhar interviews Jared Kaplan, co-founder and chief scientist of Anthropic, about the rapid advancements in AI and its potential impact on society. Kaplan discusses the expanding capabilities of AI models like Claude, including their increasing "horizon" for complex tasks and the factors driving this progress, such as model intelligence, context length, and reinforcement learning. They explore the challenges of scaling AI, the importance of "test time scaling," and the competitive landscape, including the progress of Chinese AI firms like DeepSeek. The conversation also covers the safety and governance of increasingly intelligent AI systems, the role of interpretability, and the potential economic and social disruptions caused by the fast deployment of AI. Kaplan emphasizes the need for scalable supervision of AI to ensure its beneficial use and expresses excitement about the potential for AI to contribute to solving complex problems.