llm-kb
← Back to social
Community

User identifies context drift as root cause of bugs when coding with Claude

A developer reports that bugs are introduced during coding sessions with Claude due to a divergence between what the user assumes Claude will do and what Claude actually does, caused by too much or too little context. This misalignment can lead to unnoticed changes in the codebase that may cause severe breakage later.

15 engagement·1 source·Sun, Jul 12, 2026, 11:53 PM
In a Reddit post from July 12, 2026, a user describes a pattern where, during an extended conversation with Claude, the assistant's understanding of the user's intent gradually drifts. The user believes Claude has made a specific fix or addition, but Claude has actually implemented something different, often in a part of the codebase the user won't check until a failure occurs. The user attributes this to context overload or insufficient context, causing the model to misinterpret instructions. This highlights a practical challenge for developers relying on AI coding assistants: maintaining alignment over long interactions.

Entities

Claude(model)Anthropic(company)

Related

CommunitySat, Jul 11, 2026, 01:14 PM

Developer shares CLAUDE.md trick to preserve reasoning context in Claude Code sessions

A developer describes a habit of instructing Claude Code via CLAUDE.md to log design decisions, preventing the model from forgetting past reasoning after context compaction. This addresses a common pain point in long coding sessions where the model suggests previously ruled-out approaches.

5 engagement·1 source·reddit
IncidentSat, Jul 11, 2026, 01:31 PM

User discovers Claude Code hallucinates user messages it never received

A Reddit user reports a reproducible issue in Claude Code where the model responds to a user message that was never sent. After instructing Claude to repeat the last user message verbatim, the user sent only one message but Claude responded by repeating a non-existent second message, suggesting the model may be generating or hallucinating user inputs.

5 engagement·1 source·reddit
IncidentSun, Jul 12, 2026, 04:31 PM

Users report ChatGPT project isolation and memory leaks

A Reddit user reports that ChatGPT's project isolation feature is leaky, causing the AI to bring up the user's project 'Praana' in unrelated chats. The user contrasts this with Claude, which handles context more consistently. The complaint highlights a UX engineering gap at OpenAI.

5 engagement·1 source·reddit
CommunitySat, Jul 11, 2026, 09:07 AM

Developer open-sources agent-instructions repo to curb AI coding agent degradation

A developer frustrated by AI coding agents losing context and hallucinating after about 10 minutes created a set of rules to keep them on track. The rules, shared as an open-source GitHub repo, aim to reduce the need for constant reminders and prevent infinite loops. The project has gained attention from other developers facing similar issues.

2 engagement·1 source·reddit
ProductSun, Jul 12, 2026, 08:37 AM

Claude Code extension that injects real system time into context

A browser extension for Claude Code that adds a real-time clock to the AI's context, so it knows the current time and can detect long gaps between messages. This solves the problem of Claude having no sense of time, which leads to incorrect assumptions about project duration and inappropriate suggestions like telling users to sleep at 9 AM.

2 engagement·1 source·reddit