{"slug": "disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation", "title": "DISCO: Distributed Long Context Scaling with Grounding-Reasoning Disaggregation", "summary": "A paper titled DISCO proposes distributed long context scaling with grounding-reasoning disaggregation to address \"context rot,\" the collapse in reasoning quality that occurs as LLM inputs grow toward advertised million-token context windows. The work attributes the failure to structural entanglement in monolithic architectures, where the search burden of contextual grounding is combined with reasoning.", "body_md": "While Large Language Models (LLMs) advertise million-token context windows, reasoning quality often collapses as inputs grow -- a phenomenon termed context rot. This failure stems from a structural entanglement in monolithic architectures, where the massive search burden of contextual grounding exha", "url": "https://wpnews.pro/news/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation", "canonical_source": "https://aiflash.com/news/128512/", "published_at": "2026-09-29 13:00:26+00:00", "updated_at": "2026-09-29 13:21:03.699339+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "artificial-intelligence", "machine-learning"], "entities": ["DISCO"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation", "markdown": "https://wpnews.pro/news/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation.md", "text": "https://wpnews.pro/news/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation.txt", "jsonld": "https://wpnews.pro/news/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation.jsonld"}}