Anthropic quietly redacts Claude Code telemetry logs

Anthropic's unannounced changes to telemetry logs highlight the ongoing tension between AI transparency, user privacy, and model security.

• –Removing model reasoning and prompts from traces reduces visibility for developers who rely on these logs for debugging and optimization.
• –The lack of release notes or proactive communication about the changes risks eroding trust among the developer community.
• –The move might reflect Anthropic prioritizing user privacy by minimizing logged data, or potentially tightening control over the exposure of their model's internal reasoning pathways.