OpenAI The company published a blog post yesterday (January 23) announcing the launch of a new product called "Operator" an AI intelligence that uses its own browser to perform tasks for users, is only available to US Pro subscribers at this stage.
1AI cites a blog post that describes Operator as using its own browser, Operator can use the Internet to perform a variety of tasks just like a human would, by opening a browser, clicking buttons on a page and typing in content. All those things that a human user would do online, such as booking flights, hotel reservations, planning shopping orders and completing online purchases, can be done by Operator on their behalf.
Operator is available to Pro subscribers in the U.S. at operator.chatgpt.com, with subsequent extensions to Plus, Team and Enterprise users and future integration of these features into ChatGPT.
Operator is driven by a new model called Computer-Using Agent (CUA), which combines the visual capabilities of the GPT-4 with advanced reasoning capabilities gained through reinforcement learning, and is trained to interact with the GUI, the buttons, menus, and text fields one sees on the screen.
Operator can "see" browser content through screenshots and "interact" with all the actions allowed by the mouse and keyboard, allowing it to take action on the web without custom API integration.
