--- format: "aidr-story-markdown/v1" id: "dea3f4530bc0ba186f3ef600c2ff4bd73fab2d1ce86f2c7ad99970f6afa6c588" canonical_url: "https://aidr.today/dea3f453?lang=en" title: "OpenAI Adopts Loop Transformers in Astra to Block Chain of Thought Monitoring" lang: "en" requested_lang: "en" available_langs: ["en","vi"] translation_fallback: null fallback_fields: [] published_at: "2026-09-02T02:57:49.000Z" category: "Research" topics: ["openai","astra","llm","safety","reasoning"] source_urls: ["https://huggingnews.com/ai/update-openai-adopts-loop-transformers-in-astra-to-block-chain-of-though-cb949eef","https://x.com/steph_palazzolo/status/2094954680765829533","https://x.com/amir/status/2094953820464046312","https://x.com/_NathanCalvin/status/2094957301564092914","https://x.com/_NathanCalvin/status/2094961768774647958","https://x.com/ZeffMax/status/2094879845780173283","https://x.com/wallstengine/status/2094879253871694136","https://x.com/ZeffMax/status/2094879848179363880"] summary: "The upcoming Astra AI model from OpenAI employs a new architectural approach known as recurrent depth to improve processing costs and performance. This method, also described as loop transformers, obscures the model's internal reasoning process, which makes it more difficult for safety researchers to monitor. A report from The Information states that OpenAI is currently limiting the use of this technique in Astra to maintain some level of oversight. AI safety experts describe the ability to monitor the chain of thought as a \"fragile opportunity\" for alignment and control. The shift toward non-human-readable \"neuralese\" logic could trigger a competitive race among labs to prioritize efficiency over transparency. These concerns emerge as OpenAI prepares the imminent release of Astra, which the company previously disclosed has reached a \"critical\" cybersecurity threshold." --- # OpenAI Adopts Loop Transformers in Astra to Block Chain of Thought Monitoring > [Open the canonical story]() **Published:** 2026-09-02T02:57:49.000Z **Category:** Research **Topics:** openai, astra, llm, safety, reasoning ## Summary The upcoming Astra AI model from OpenAI employs a new architectural approach known as recurrent depth to improve processing costs and performance\. This method, also described as loop transformers, obscures the model's internal reasoning process, which makes it more difficult for safety researchers to monitor\. A report from The Information states that OpenAI is currently limiting the use of this technique in Astra to maintain some level of oversight\. AI safety experts describe the ability to monitor the chain of thought as a "fragile opportunity" for alignment and control\. The shift toward non\-human\-readable "neuralese" logic could trigger a competitive race among labs to prioritize efficiency over transparency\. These concerns emerge as OpenAI prepares the imminent release of Astra, which the company previously disclosed has reached a "critical" cybersecurity threshold\. ## Sources - [Story source]() - [Story source]() - [Supporting source]() - [Supporting source]() - [Supporting source]() - [Story source]() - [Supporting source]() - [Supporting source]()