Latest Interviews
Showing 76–77 of 77 interview transcripts.
Clear all filters- Sequoia Capital50 min
LangChain’s Harrison Chase on Building the Orchestration Layer for AI Agents | Training Data
Harrison Chase, Sonya Huang, Pat Grady
Harrison Chase positions Langchain as a critical orchestration layer for the "middle ground" of agent autonomy, prioritizing production-grade reliability over the volatile hype of fully autonomous systems. The company addresses this shift by deploying LangGraph for complex, stateful workflows and LangSmith for observability, enabling organizations to build custom cognitive architectures that balance flexibility with necessary human-in-the-loop controls. As the industry transitions from static chains to dynamic agents, these tools facilitate the move from customer support automation to software development integration while redefining testing and user experience paradigms.
- Sequoia Capital32 min
Trust, reliability, and safety in AI ft. Daniela Amodei of Anthropic and Sonya Huang
Anthropic, a Public Benefit Corporation founded three years ago to prioritize trustworthy generative AI, recently launched its Claude 3 model family featuring the high-complexity Opus, the cost-efficient Haiku, and the enterprise-focused Sonnet. These models deliver state-of-the-art coding proficiency and reduced hallucination rates, enabling adoption across sectors from healthcare's Dana-Farber Cancer Institute to financial firms like Bridgewater. Despite ongoing challenges in fully autonomous agent behavior, the company continues to advance safety through its pioneering Constitutional AI techniques and proactive Responsible Scaling Policy to align technical capability with human values.