Why AI Chatbots Forget the Conversation: The LLM Context Window Explained
An LLM doesn't keep memory between requests. With every new message, the model rereads the entire conversation from scratch. This "visibility box" is called the context window: it holds a limited number of tokens, and every token is part of your chat history. Anything that falls outside the window is simply invisible to the model. That's why long conversations start to "forget" details from the beginning.
AI-processed from Habr AI; edited by Hamidun News
A neural network's context window is an architectural trait that explains why AI doesn't work like a person with memory. Instead, it works like an expert who rereads the entire correspondence from scratch every time before writing a reply.
Why does AI forget parts of a conversation?
AI doesn't forget in the usual sense — it has neither working memory nor long-term memory. The model receives the whole conversation every time, but the amount of information it can process at once is capped by the context window. Once a conversation grows too long, older messages fall outside the window and the model can no longer see them.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.