Context Rot
Context rot is the decline in AI reasoning and rule following as its context window fills up.
“Context rot is the slow erosion of model reliability, where early instructions are forgotten as conversational memory fills up.”
Context rot causes autonomous agents to forget critical safety instructions and operational constraints. This phenomenon leads to hallucinated API calls and severe breaches of deterministic governance protocols.
Reverse Citations: Implemented & Audited Across Platform
Richard Ewing’s Research Thesis
We cannot rely on long context windows to enforce complex rules. To combat context rot, we must implement stateless tool calls and deterministic governance architectures that validate constraints outside the LLM context.
Latest Publications & Research Activity
How to Prevent Memory Loss in AI Applications
Giving an AI a bigger memory window is like giving a confused worker a bigger inbox.
Claude Search Fails: Prompting Kills Adoption
Frequently Asked Questions
Q:What causes context rot?
The dilution of attention across a large volume of tokens in a long interactive session.
Inspectable Evidence Ledger
Classified evidence items supporting, extending, or refining this canonical research specification.
| Evidence Item | Publisher | Evidence Type | Strength | Role | Action |
|---|---|---|---|---|---|
| Context Window Efficacy | AI Research Metrics | Research Note | ★★★★ | Origin | Inspect ↗ |
Recommended Citation
Ewing, R. (2026). "Context Rot." Richard Ewing Research Canon. Available at: https://www.richardewing.io/concepts/context-rot
@article{ewing_context_rot,
author = {Ewing, Richard},
title = {Context Rot},
journal = {Richard Ewing Research Canon},
year = {2026},
url = {https://www.richardewing.io/concepts/context-rot}
}