OpenAI reportedly discovered a new optimization method that can reduce inference costs by over 50% Odaily Planet Daily News: OpenAI's engineering team recently revealed to some colleagues that the company has found a new system optimization method that can reduce the "inference" cost of AI models by more than half. Inference cost refers to the computational resource cost consumed by the model in actual operation and response to user requests. This optimization mainly comes from improving the utilization efficiency of existing server resources, rather than relying on the investment of new computing chips. This progress reflects that AI companies are continuously competing for computing resources while also improving the efficiency of existing infrastructure through software and system layer optimization to alleviate the pressure of rapidly growing model operating costs. （The Information）