Overview

• The Learning mode, initially exclusive to Claude for Education, is now available in the style dropdown for all Claude.ai users to adopt a Socratic approach.
• Claude Code’s new output styles include Explanatory mode for narrating AI decision-making and Learning mode that pauses with “#TODO” prompts to invite developer input.
• Users can enable the new coding styles by updating Claude Code, running the /output-styles command and selecting their preferred interaction preset.
• Anthropic implements these teaching behaviors via modified system prompts rather than fine-tuning models to allow rapid iteration based on real-time feedback.
• Anthropic frames learning modes as a human-centric response to student concerns about ‘brain rot’, with similar offerings rolling out from OpenAI and Google.