In a world already accustomed to the lightning speed of technological progress, OpenAI has made a statement that could reorder the entire scientific paradigm. During a livestream on Tuesday, the company’s CEO, Sam Altman, revealed an ambitious and aggressive timeline: creating a “full-fledged autonomous AI researcher” by 2028.
OpenAI’s deep learning roadmap is divided into two phases that demonstrate the exponential leap the company expects:
2026 (September): Systems should reach the level of a research intern, capable of performing certain functions in scientific projects.
2028: The goal is to achieve the level of an autonomous, full-fledged AI researcher – a system capable of independently initiating and executing large-scale research projects.
OpenAI's chief scientist, Jakub Pachotski, explained that "AI researcher" is not just a next-generation chatbot, but a system that can autonomously work on complex, critical tasks. Pachotski added that the company is confident that superintelligent systems that will surpass human intelligence on a wide range of critical tasks can be created using deep learning methods in less than a decade.
To realize this impressive leap, OpenAI will focus on two key technology directions: Continuous algorithmic improvements, developing more efficient and powerful deep learning algorithms, and Radical Increase in computational resources: Increasing power, especially in the “inference“ phase – the time it takes for the model to "think" and process a given task.
Pachocki gave an eloquent example of current capabilities: current models can run for about five hours and achieve results comparable to those of medalists at the International Mathematical Olympiad. With the expected expansion of computing power, this "thinking time" will quickly expand, opening the door to solving incomparably more complex problems.
The ambitious schedule was announced on the same day that OpenAI finalized its restructuring - transitioning from a non-profit organization to a public benefit corporation. This change frees the company from the constraints of its original charter, opening up new horizons for raising capital for scientific expansion.
The new structure ensures that scientific progress remains the top priority. The OpenAI Foundation (a non-governmental organization) will receive a 26% stake in the commercial division. The foundation will determine the direction of research and oversee safety initiatives for AI.
In its strongest commitment yet, the Foundation has committed $25 billion to using AI to treat disease, underscoring AI’s ambition to accelerate scientific discoveries in medicine, physics, and technology that humans cannot achieve in the short term.
In conclusion, OpenAI is not just creating better algorithms; it is reshaping the very future of scientific discovery. It remains to be seen how this technology will address ethical and practical challenges as we move toward 2028.