OpenAI Astra Model Draws Safety Concerns

TechCrunch reports OpenAI’s Astra model will use “recurrent depth,” a technique that allows the model to operate outside of the sequential thinking that characterizes most reasoning models. Company statements claim advances in computer and browser use plus speed, accuracy, and safety. Neowin reports performance gains over prior versions. The Information quotes an OpenAI leader stating the looping technique makes chains of thought less likely to appear in output, which could complicate researcher oversight. TechRadar calls for caution over the resulting opacity.