Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
OpenAI has started previewing a new tool called Operator that can run in a browser. According to the blog post published on Thursdaythe software is managed by what the company calls the Desktop User Assistant. “CUA is trained to interact with graphical user interfaces (GUIs)—the buttons, menus, and text that people see on screen—as humans do,” says OpenAI of the model. “This provides the ability to perform digital tasks without using OS- or web-specific APIs.”
The latest release of Operator builds on OpenAI’s GPT-4o model. It combines the capabilities of algorithmic vision with “superior thinking” taught through reinforcement learning. Employees can “divide tasks into multiple plans and manage themselves when problems arise.” According to OpenAI, this capability represents the next step in the development of AI.
Like previous observations, OpenAI warns that the operation is “still early and has limitations,” and that it “does not perform reliably in all situations.” For example, depending on the complexity of the task and the features involved, the agent will benefit greatly from the user taking a few minutes to write a detailed description. Per SeasideThe users will give the user the power if they are stuck on the job. It will also provide authority whenever a website asks for personal information, including login information. The company says it designed the tool to “reject malicious requests and prevent unauthorized access.”
OpenAI is making Operator initially available to its users for $200 per month ChatGPT Pro subscription. It also cooperates with companies such as Instacart to provide support for their platforms, even then you will need a ChatGPT Pro subscription to test the integration.
The assistant joins a growing list of AI assistants that can target a browser or an entire operating system. Anthropic was the first to provide the capability with its release Claude 3.5 Sonnet example in Octoberfollowed soon by Google and its Gemini 2.0 model is Project Mariner.
If you buy something through a link in this article, we may get a job.