Hayden Field, senior AI reporter at The Verge, guest hosted an episode of Decoder discussing the advancements in the AI industry, specifically focusing on a new model released by Anthropic called Claude Sonnet 4.5. Field interviewed David Hershey, the lead of the applied AI team at Anthropic, who collaborates with startups to implement Anthropic’s technology and tests new AI models to assess their capabilities.
Claude Sonnet 4.5 was recently introduced and is intended to enhance autonomous AI functions, particularly in coding tasks. Anthropic claims that this model can operate for up to 30 hours continuously on a single task, such as developing a software application without human assistance.
In the past year, multiple companies, including Anthropic, Microsoft, and OpenAI, have been promoting agentic technology as the next significant evolution in AI, following the rise of general-purpose chatbots. They suggest that such advancements could unlock the full potential of generative AI, and incremental progress has been made.
Despite these claims, deployment of agentic AI remains limited. Currently, most users are not utilizing these systems for extensive autonomous tasks, such as those requiring hours of independent operation. Nonetheless, companies are hopeful that agentic technology may contribute to significant productivity enhancements by either supplementing or replacing human labor.
Field aimed to understand the current status of AI agents by discussing with Hershey the practical applications of models like Claude Sonnet 4.5 beyond programming tasks and exploring the future trajectory of agentic technology.
For further insights from the episode, listeners can refer to related links provided.
Source: https://www.theverge.com/podcast/789772/ai-agents-anthropic-claude-sonnet-autonomous-coding