Perplexity Launches Computer Use Feature, Raising Comparisons to Anthropic's Claude

Summary

Perplexity has introduced a new computer control feature that allows its AI to interact with users' computers, prompting immediate comparisons to Anthropic's Claude Computer Use capability. The feature, which appears to be positioned as a potentially safer alternative, enables the AI assistant to perform tasks directly on a user's machine. While specific technical details remain limited, the announcement has sparked discussion in the AI community about the safety implications of giving AI systems direct computer access.

The reference to 'OpenClaw' in discussions appears to be a play on 'Claude,' Anthropic's flagship AI model that pioneered computer use capabilities for consumer AI assistants. Anthropic's Computer Use feature, launched in late 2024, allows Claude to control a computer by viewing screens, moving cursors, clicking buttons, and typing text. This capability has been both celebrated as a breakthrough in AI autonomy and scrutinized for potential security risks.

Perplexity's entry into computer control capabilities represents a significant expansion of the company's AI assistant beyond its core search and answer engine functionality. The company has been rapidly adding features to compete with larger AI labs, including real-time information retrieval and multi-modal capabilities. How Perplexity's implementation differs in terms of safety guardrails, permissions models, and use cases compared to Anthropic's approach will be critical for user adoption and regulatory scrutiny.

• Safety and security concerns around giving AI systems computer access remain a key consideration for both companies and users