OpenAI is Working on the Revolutionary AI Strawberry

OpenAI, the developer of the popular ChatGPT, is actively working on a secret project named Strawberry. The team involved aims not just to create another neural network for generating user query responses but to develop a system that approaches artificial general intelligence (AGI). This system would be capable of performing a wide range of tasks at or even beyond human levels. At the company’s recent internal meeting, a demo version was presented that demonstrated its ability to reason at a human level. OpenAI hinted that the company would soon announce the readiness of the new technology, although project details remain strictly confidential even among employees.

The strategic goal of the Strawberry project is to create a neural network that can not only respond to inquiries but also independently conduct its own research online. Such an approach should allow AI from OpenAI to act and think at a significantly higher level, analyzing all possible scenarios and autonomously exploring cyberspace. This would enable it to perform complex tasks requiring planning over extended periods.

Experts believe that the ability to reason is a key aspect in achieving AI intelligence at or above human levels. Current LLMs are effective at generating and summarizing texts for various needs. At the same time, they often struggle to adequately solve tasks that are intuitively understandable to humans and may make logical errors. This flaw leads to unexpected "hallucinations."

Initially, the Strawberry project was named Q. In 2023, OpenAI called this model groundbreaking as it was already successfully solving complex scientific and mathematical problems that other models could not handle. This was made possible by applying learning after initial data processing.