New arXiv paper quantifies CoT necessity via opaque serial depth amid OpenAI architecture rumors
ArthurConmy · x · 2026-09-04
Former mech interp researcher Arthur Conmy argues the discourse around The Information's report on OpenAI's alleged looping-transformer-like architecture was poor, pointing to a new arXiv paper (2603.09786) for technical grounding:
- The paper formalizes "opaque serial depth": the longest computation a model can perform without interpretable intermediate steps like chain of thought.
- It computes numeric upper bounds on Gemma 3 models and asymptotic results for other architectures.
- The team open-sourced an automated method to bound opaque serial depth for arbitrary neural networks, showing MoE models likely have lower depth than dense ones.
- Takeaway: there is likely no binary divide between "neuralese" and non-neuralese models — serial depth is a spectrum, and keeping it small preserves CoT monitorability.
Relevant to the ongoing debate about whether new recursive architectures could undermine reasoning monitorability.
Related event: Looped Transformer Rumors Spark Fierce Debate Over CoT Monitorability(7 posts)→
More from Safety
- CVSS 10.0 RCE in PHP core: critical patches land Oct 21, PHP 8 only — cyb3rops · 2026-10-21
- Anaconda ships MCP to stop coding agents from hallucinating malicious dependency packages — anacondainc · 2026-10-21
- Skip the open access fees: NIH-funded papers must go on PMC with no embargo — anshulkundaje · 2026-10-21
- Creed: LLM code auditing infers developer beliefs to find semantic security flaws — mboehme_ · 2026-10-21
- Spotify removed 50M+ AI slop tracks from recommendations in a year to back human artists — moustaki · 2026-10-21
- METR signs off on Anthropic's risk report but experts demand quantified uncertainty and neutral evals — DKokotajlo · 2026-10-21