File-native agents: when reading files backfires
A 9,649-experiment study found file-native retrieval lifts frontier-model accuracy 2.7% and drops open-source accuracy 7.7%. Match architecture to the model.
A 9,649-experiment study found file-native retrieval lifts frontier-model accuracy 2.7% and drops open-source accuracy 7.7%. Match architecture to the model.
AI notetakers ship transcripts, but downstream work needs decisions, drafts, or handoffs. The artifact gap is a context engineering problem, not transcription.
Constraint decay: AI coding agents lose 30 points of accuracy under architecture and database rules. New EURECOM study explains why and where it hurts most.
Context offloading keeps an AI agent's working context window small by moving state to a destination outside it. Three patterns, and what each one costs.
Connecting an MCP server is easy. Getting an agent to call its tools on the first relevant turn is where teams lose, and the cause is context.
OX Security's April 2026 advisory traces 14 MCP CVEs and 200,000 exposed servers to a single design choice: STDIO as the default local transport.
Anthropic's 2026 trilogy on context engineering, tools, and code execution with MCP each assume the same missing layer: the substrate where context lives.
An MSR 2026 study of 466 open source projects maps the five modes developers use to write AGENTS.md context, and what 50% file staleness reveals about practice.
Memory consolidation fixes one specific failure: agents writing the same claim dozens of times into a flat scratchpad. When it helps and where it breaks.
GitHub's MCP costs tens of thousands of tokens before any work begins. We compare MCP, Claude Skills, and CLI by context cost, not by user preference.