ChatGPT may have introduced many users to AI, but the browser is fast becoming its next major frontier. What sets the browser apart is its direct access to real-time user data — from logged-in accounts to financial tools — offering the contextual awareness today’s AI needs to perform real actions.
According to The Verge, two recent products hint at this shift: OpenAI’s ChatGPT Agent and Perplexity’s Comet browser. ChatGPT Agent can browse the web but remains limited to read-only access. Comet, on the other hand, is designed to go deeper by logging into websites and attempting task completion through large language models.
Neither tool currently delivers on all its promises. ChatGPT Agent is slow and unable to act on logged-in sites. In one case, it took 50 minutes to search Etsy and failed to add an item to a cart, despite claiming success. Comet performs faster, but still struggles with reliability and overpromises its capabilities.
Perplexity CEO Aravind Srinivas believes in future improvements. “We’re betting on progress in reasoning models to get us there,” he said. OpenAI’s custom reasoning model for its agent is a step in that direction.
The browser may soon become AI’s most powerful workspace.
Article updated 3 days ago. Content is written and modified by multiple authors.