On September 2, OpenAI's Chief Scientist, Jakub Pachocki, posted on a social media platform, responding to the controversy surrounding the 'unmonitorable' nature of the new model, Astra. The controversy arose on September 1 when news emerged that the Astra model employed recurrent deep reasoning technology, which allows the same set of Transformer layers to be computed repeatedly, with part of the reasoning process completed internally within the model rather than being fully presented in the form of a chain of thought, hence the term 'opaque loops.' Pachocki stated that he hopes to avoid sparking an arms race toward an unmonitorable state, where major labs compete to develop models that are too powerful for humans to monitor or audit. He emphasized that the computational graph depth of the Astra model differs by no more than a factor of two from that of GPT-4, and there has been no leap in architectural complexity. To address safety risks, OpenAI will deploy additional chain-of-thought monitoring for Astra to swiftly detect and curb potential misbehavior.