OpenAI has indefinitely postponed the release of its highly anticipated open-source AI model, originally set for this summer but now delayed for additional safety reviews and risk evaluations, according to CEO Sam Altman. He announced on X (formerly Twitter) that the company is exercising caution due to the model’s unprecedented nature, emphasizing the need for thorough safety tests. This release aims to make the model fully downloadable, contrasting with the tightly regulated GPT-4 and upcoming GPT-5. While the model promises unrestricted access to essential parameters, concerns about potential misuse are significant. Aidan Clark, OpenAI’s VP of Research, stated the model’s capabilities are exceptional, but a higher standard for open-source release necessitates more time. Competitors like Moonshot AI are gaining ground, presenting challenges in an increasingly competitive landscape that includes tech titans such as Google and Anthropic. Developers eagerly await the release, which Altman implies may hold significant breakthroughs.
Source link
Share
Read more