On January 23, OpenAI announced a new tool called Operator – an agent that performs online tasks on behalf of a user. It can open websites, fill out forms, scroll through pages, and even create memes. Currently, Operator is available to Pro subscribers in the U.S. as part of a testing phase,
writes the OpenAI portal. The tool is based on the Computer-Using Agent (CUA) model, which combines GPT-4's capabilities in image recognition and logic. The agent can view websites through screenshots, use a keyboard and mouse. If it encounters difficulties, it analyzes its actions or requests assistance from the user, for instance, for authentication or entering payment information. Operator is capable of performing various tasks, such as filling out forms, ordering products, adjusting flight parameters, and multitasking. Users can also provide custom instructions or save recurring requests. This agent simplifies the use of digital services for individuals and businesses. OpenAI is already collaborating with DoorDash, Uber, OpenTable, and others to enhance their convenience. In the public sector, for example in Stockton, the tool assists with registration in municipal programs. Operator has limitations, but its functionality is expected to expand. In the future, the tool will be available to a broader audience and will be integrated into ChatGPT.
Previously, we wrote about how the music of the future: artificial intelligence signed the first contract in history with a studio
Additionally, Znai.ua reported that Razer introduced Project AVA: an AI that will teach you to play better
Our portal also informed that GTA fans uncovered Rockstar's secret: what "goodies" developers are hiding along with modders
