According to a report by Reuters, the code name for the covert initiative is Strawberry.
OpenAI is reportedly working on a secret project that is being code-named “Strawberry” to build technology that will add sophisticated reasoning capabilities to its artificial intelligence models, according to a new story from Reuters. According to a document that was seen by Reuters from within the corporation, one of the objectives of the project is to make it possible for the artificial intelligence models of the company to independently search the internet in order to “prepare ahead” for more difficult tasks. The project was formerly known as Q* (pronounced “Q star”), and demonstrations of it demonstrated earlier this year that it was capable of answering “tricky science and math questions,” according to Reuters, which cites unnamed sources who witnessed the demonstrations. There were also demonstrations of the project.
Some things regarding Strawberry are still unknown at this point, such as how far along in the development process it is and whether or if it is the same system with “human-like reasoning” skills that OpenAI purportedly exhibited at an employee all-hands meeting earlier this week, as reported by Bloomberg. In contrast to what is now accessible, however, the capability of the company’s artificial intelligence to carry out “deep research,” which is rumored to be the goal of Strawberry, would represent a significant step forward.