Although OpenAI has denied the release of its next major model, ‘Orion‘, this year, it is preparing to launch an AI agent called ‘Operator’ in January. According to Bloomberg, the Operator AI agent will be capable of autonomously performing tasks on a computer or web browser based on user input.
The report states that OpenAI informed its staff on Wednesday about plans to launch the Operator AI agent in January. Initially, the tool will be available as a research preview and through OpenAI’s API for developers.
Following the ChatGPT boom, AI companies are shifting towards action-driven agents. Recently, Anthropic launched its Computer Use tool, which can perform tasks on a computer based on user prompts.
I tested Anthropic’s Computer Use API via Docker and found it promising, though quite slow. The tool identifies click targets by analyzing what’s on the screen. However, it often gets stuck in a loop, and even the improved Claude 3.5 Sonnet (New) model struggles to resolve this issue.
Despite the challenges, 2025 is expected to be the year of AI agents. Google is also set to release its Jarvis AI agent in December. Available as a Chrome extension, Jarvis will perform actions based on user input and is described by Google as “a helpful companion that surfs the web with you.”
As LLM scaling has reached its limits, companies are focusing on creating new applications and use cases that leverage AI’s current strengths. In a recent Reddit AMA, OpenAI chief Sam Altman mentioned, “I think the thing that will feel like the next giant breakthrough will be agents.” With the rise of AI agents, it will be interesting to see which company emerges as the leader in this rapidly evolving space.
0 Comments