OpenAI Is Launching An Agent In January That Will Control Computers

OpenAI is expected to launch its successor to GPT-4 in December, according to reports. However, in January, they will have another release. This time, they plan to launch an internal software tool called “Operator.” Operator can take direct actions on a computer. Initially, they intend to launch Operator as a research preview through its developer API, similar to how they launched their search engine earlier this year.

Operator will compete with Anthropic’s Computer Use feature and may also face competition from Google’s version of the same feature, which could launch next month when the next Gemini model is announced. Reports suggest that Operator could be a general-purpose tool that executes tasks in the web browser.

This news aligns with Sam Altman’s claims that we will have increasingly sophisticated models, which were revealed during a Reddit AMA session. At their Dev Day event last month, Kevin Weil, OpenAI’s chief product officer, predicted that 2025 will be the year when agent systems finally become mainstream.

This news comes at a time when OpenAI released its policy paper, which outlines its US AI strategy. The paper proposes that the federal government create AI-focused economic zones that allow work to be done on building AI infrastructure and establish a bloc of US allies to compete with China in the AI field.