Anthropic has reportedly removed or obscured the "thought process" feature in its Claude models, starting around the release of Claude Opus 5. Users are now seeing "Thought process is unavailable" messages, and newer documentation suggests that even summarized reasoning may be omitted by default. While Anthropic may be doing this to prevent model distillation or the exposure of sensitive information, critics argue that visible reasoning is crucial for auditing AI decisions, understanding model interpretation, and identifying errors, especially in agentic systems. AI
IMPACT Reduced transparency in AI models may hinder auditing and understanding of AI decision-making processes.
RANK_REASON User-generated discussion and opinion piece about a product change, not an official announcement.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
