New to Claude Cowork? Start Here.

The Evolution of AI: From Chat to Co-Work

The Limitations of Current AI Interactions

• Users often find that conversations with AI, like Chat GPT or Claude, lack tangible outcomes despite engaging discussions.

• There is a perception that current AI tools are limited to conversational roles, acting merely as assistants without executing tasks.

Introduction of Claude Code

• In February 2025, Anthropic launched Claude Code to enhance the capabilities of their AI beyond simple chatting.

• Developers began utilizing Claude for various tasks such as project management and research, expanding its use cases significantly.

Launch of Claude Co-work

• In January 2026, Anthropic introduced Claude Co-work, designed to make the power of coding more accessible for all users.

• This new mode allows users to collaborate with the AI on shared files that persist over time rather than just having conversations.

Features and Interface Changes in Co-work Mode

• The interface includes three tabs: Chat (beginner mode), Code (advanced mode), and Co-work (an enhanced version).

• Notably labeled as "new task" instead of "new chat," this change emphasizes a focus on actionable items from users' to-do lists.

Enhanced Collaboration Capabilities

• Users can grant the AI access to entire folders for comprehensive project work rather than selecting individual files. This facilitates deeper collaboration and exploration within a defined workspace.

• As an example, the speaker demonstrates how they allow access to their book manuscript folder for suggestions on improvements, showcasing real-time interaction with the AI's thought process and task management features.

AI-Powered Feedback and Planning Tools

Utilizing AI for Comprehensive Feedback

• The speaker discusses the ability of AI to analyze different parts of a book manuscript simultaneously, effectively "spawning" multiple AIs to tackle various sections in parallel.

• After a brief analysis period, the AI provides structured feedback on the manuscript, highlighting critical issues and suggesting improvements based on priority and impact.

• Key issues identified include an accessibility gap that may alienate readers, particularly in Chapter 12, along with structural pacing problems and vague guidance.

• The speaker notes that previous attempts to analyze the entire manuscript were limited by context; however, this new approach allows for comprehensive insights across all chapters at once.

• Emphasizing the importance of context, the speaker explains how having access to the full manuscript enables better advice from the AI.

Cost and Accessibility of AI Tools

• To utilize these advanced features, users need at least a paid standard plan costing $20 per month and access to a MacOS desktop app (Windows version forthcoming).

• Previously requiring higher-tier plans for similar capabilities, both Claude Code and Co-work are now accessible at this lower price point.

• The speaker recommends adopting Co-work even for everyday tasks due to its enhanced functionality compared to standard LLM interactions.

Real-Life Application: Planning a Family Road Trip

• The speaker shares their experience using Whisper Flow to plan a family road trip in central Mexico while considering specific needs such as traveling with young children and a pregnant spouse.

• They outline preferences for non-strenuous activities focused on nature and culture while avoiding overly touristy destinations.

• Turning off existing context from their book manuscript allows the AI to conduct web research tailored specifically for this task without prior biases influencing its suggestions.

• The AI performs extensive web research across multiple sites to gather information about family-friendly accommodations and unique experiences aligned with user preferences.

• A notable feature of Co-work is its "plan mode," which prompts follow-up questions from the user rather than making assumptions—enhancing interaction quality.

Travel Itinerary Planning with Co-Work

Initial Recommendations and Adjustments

• The speaker discusses a mixed recommendation for visiting Monarch Butterfly Sanctuaries, presenting a four-night itinerary that includes specific details about each day.

• The speaker notes the importance of iterating on the plan through conversation to build confidence in the recommendations, emphasizing their proximity to one of the sanctuaries.

• After revising, the itinerary excludes butterfly visits and suggests alternative locations like Morelia, indicating flexibility in travel plans based on preferences.

Finalizing the Itinerary

• The speaker requests a summarized day-by-day itinerary that is printable and easy to carry while traveling.

• A Word document is created automatically containing detailed information such as driving times, activities for morning and afternoon, accommodations, roads, weather, and food options—none of which were explicitly requested by the speaker.

Contextual Considerations in Co-Work

• It's highlighted that all work done with co-work is stored locally on one's machine; switching devices means losing access to previous conversations.

• The context management differs from traditional project file uploads; users maintain a folder where all ongoing work is stored. This requires an adjustment in mindset when using co-work tools.

Advantages of Co-Work Collaboration

• Despite requiring more consideration than typical chat or search functions, co-work represents significant progress for users of large language models (LLMs).

• The collaborative nature of co-work allows for dynamic interactions between users and LLMs as if they are working side by side on shared documents.

Additional Resources

• A free getting started guide has been created to help users navigate through co-work effectively after covering extensive content in this video.