Building an AI to slow down: a tutorial to get started differently with Claude Code

2025-11-30 · Source: Génération IA · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Novice, extended

Summary

The article introduces the concept of "slow AI," advocating for taming a single AI agent over time within a personalized workspace rather than constantly chasing new releases, directly addressing the "Fear Of Missing Out" (FOMO) prevalent in the rapidly evolving AI landscape. Authored by Benoit Raphael, Thomas Mahier, and Jeff GPT on May 24th, 2026, it provides a step-by-step tutorial for setting up such a system using Claude Code and Visual Studio Code. The guide details creating a "harness" for the language model by configuring specific folders and files in VS Code, installing the Claude Code extension, and connecting to external tools like Gmail. It also highlights Claude Code's capabilities for working with long documents and generating custom, "disposable" tools, such as a YouTube video downloader or a video-to-HTML summary generator.

Key takeaway

For AI engineers or professionals overwhelmed by constant AI updates, adopting a "slow AI" approach with Claude Code can build a stable, personalized intelligence capital. Instead of chasing new models, you should establish a dedicated workspace in Visual Studio Code, configure Claude Code with persistent context files, and let it create custom tools. This strategy allows you to tame one AI over time, reducing FOMO and fostering deeper collaboration.

Key insights

Taming a single AI agent over time within a personalized workspace combats FOMO and builds lasting intelligence capital.

Principles

• AI tools structure activity; handwriting offers different cognition.
• An AI "harness" transforms a model into an active collaborator.
• Persistent memory for AI is built through structured files and folders.

Method

Set up a "workspace" folder, install Visual Studio Code and the Claude Code extension, then configure the AI with CLAUDE.md and who-i-am.md files to establish its memory and context. Connect external tools via Claude web.

In practice

• Create a CLAUDE.md file to define the AI's workspace purpose.
• Use who-i-am.md to provide personal context to the AI.
• Connect Claude Code to Gmail via Claude web for email management.

Topics

• Slow AI
• Claude Code
• Visual Studio Code
• AI Agents
• Personalized Workflows
• Digital Workspace

Best for: AI Student, AI Engineer, Software Engineer

Related on AIssential

• Claude Code 2.0 MASSIVE Upgrade! (Game Changer) — Claude Code's latest updates enhance developer productivity through improved scheduling, memory, and multi-agent capabilities.
• How to automate your life with Claude Code | Hilary Gridley — Automate tasks with low impact potential to free human effort for high-impact activities and personal enrichment.
• Give Claude Permanent Memory — Claude's memory architecture offers layered strategies for persistent context, moving beyond stateless interactions.
• Mar 23, 2026ScienceLong-running Claude for scientific computing — Autonomous AI agents can compress months of scientific coding work into days by leveraging structured workflows.
• I Let Claude Code for Me for 30 Days. Here’s What Nobody’s Telling You. — Extensive AI coding assistance can degrade developer instincts and deep code comprehension through cognitive offloading.
• 23 Tips for Smart Claude Code Token Saving and Workflow Optimization — Proactive context management and strategic tool usage are crucial for controlling Claude Code token costs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Génération IA.