August 28, 2024 – On August 7 local time, Altman teased a mysterious “Strawberry” project by sharing a photograph of strawberries, sparking widespread discussion across the internet.
According to The Information, as OpenAI seeks to raise additional funds, its researchers are working on launching a new AI product designed to tackle thorny issues more effectively than existing offerings.
It is reported that OpenAI has already demonstrated this technology to U.S. national security officials this summer. Sources with direct knowledge of these meetings, which were previously unreported, have confirmed this information.
Insiders have revealed that the team aims to introduce the new AI project, codenamed “Strawberry” (previously referred to as Q*), as part of a chatbot, possibly ChatGPT, by fall this year.
Remarkably, “Strawberry” can solve mathematical problems it has never encountered before, a feat that is challenging for existing AI systems. Although it was trained to address programming-related issues, it has the capability to answer any technical question.
According to informed sources, the “Strawberry” project is designed to enhance OpenAI’s models’ reasoning abilities, improving their capacity to handle complex scientific and mathematical problems. This advancement allows large models not only to generate answers to queries but also to plan ahead, enabling autonomous and reliable internet browsing for “deep research” as defined by OpenAI.
