SAN FRANCISCO - Artificial intelligence lab Anthropic on Wednesday announced a new capability for its Claude AI system, a feature it calls 'dreaming.'
The tool, available now as a research preview, is designed for self-improvement. It allows AI agents to review their own work between sessions, identify patterns, and update internal files that store user preferences and other contextual data.
Anthropic unveiled the feature at its developer conference in San Francisco, part of a broader push to win business customers. The company's AI-powered coding agent has seen a surge in popularity.
On Tuesday, the startup launched 10 financial AI agents at an event in New York. According to Anthropic, the tech sector is its largest source of enterprise revenue, followed by financial institutions.
The Google and Amazon-backed startup's aggressive moves have weighed on software-as-a-service stocks, as investors anticipate that AI will disrupt legacy business models.
Anthropic also announced wider availability for additional capabilities, including one that enables an AI agent to break down and delegate tasks to other specialist agents.