You've Been Using AI the Hard Way (Use This Instead)

You've Been Using AI the Hard Way (Use This Instead)

by NetworkChuck · View the channel on PSI

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 NetworkChuck 👥 5.4M 📅 October 28, 2025 ⏱ 33 min 👁 2.2M 📄 tutorial 🧭 2026-09-16
Available in: English (current) Français

Keywords

terminal AICLI toolscontext filesagentsmulti-AI workflow

Summary

The video, hosted by NetworkChuck, argues that using AI in the terminal (via tools like Gemini CLI, Claude Code, and OpenCode) is far more productive than traditional browser-based chat interfaces. It begins by illustrating the frustrations of managing multiple browser chats and lost context. The presenter then installs Gemini CLI and Claude Code, demonstrating key features such as project-aware context files (e.g., GEMINI.md, CLAUDE.md), agents for delegating subtasks, output styles for customizing AI persona, headless mode for scripting, and the ability to run multiple AI tools simultaneously on the same project. He emphasizes that this approach gives users ownership of their data and context, preventing vendor lock-in. The video also includes a sponsored segment on Twingate for secure remote access, tied to the theme of granting AI access to your computer. A daily workflow is shown, using Git for version control and synchronization of context files. The creator finishes with a personal note and an ethical reminder.

157 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video offers substantial practical value, demonstrating concrete steps to set up and use terminal-based AI tools. The argumentation is logical and persuasive, contrasting the inefficiencies of browser chats (lost context, scattered notes) with the streamlined workflow of terminal tools that can read/write files, manage context via files, and delegate tasks to agents. The presenter’s enthusiasm is backed by live demonstrations, which adds credibility. However, the claims of being ‘10x faster’ are anecdotal and not measured; the video does not provide comparative benchmarks or test results. The demonstration of multi-agent tasks and headless operation is compelling, but the security implications of granting AI full file system access are only briefly addressed. Overall, the video’s argument is solid for its intended audience but lacks empirical evidence.

Scientific Rigor, Source Quality, Title Accuracy

The video is a tutorial and opinion piece rather than a research presentation. It references official documentation and product pages (Gemini API, Claude Code, OpenCode) via links in the description, but within the video itself it does not cite external studies or independent reviews. The title is well-matched to the content. The sponsor segment (Twingate) is clearly disclosed and relevant to the topic of secure access. The creator’s expertise is evident from his established tech education channel, and the demonstration is reproducible. However, the absence of critical evaluation of potential drawbacks (e.g., token costs, security risks) weakens the scientific rigor. The comments section shows a largely positive reception, with many viewers expressing gratitude and asking for follow-up security guidance, indicating the content resonates with the audience.

266 words

Title / Content Match

The title accurately reflects the content: the video argues that using AI in the terminal is more efficient than browser-based usage and demonstrates the alternative tools and workflows.

Quality & Reliability

7/10

The video provides practical, hands-on demonstrations of terminal AI tools with clear setup steps and workflow examples. However, it relies heavily on personal experience and does not cite formal studies or independent benchmarks. The sponsored segment (Twingate) is clearly labeled and does not undermine core content. Overall, the information is reliable but not scientifically rigorous.

Chapters

Cited Sources

  • Gemini API — Official documentation for the Gemini CLI and API, referenced as the tool used in the first segment.
  • Claude Code — Official product page for Claude Code, the terminal tool demonstrated extensively in the video.
  • OpenCode — Official site for OpenCode, another terminal AI tool mentioned and demonstrated.
  • NetworkChuck's AI-in-the-Terminal Walkthrough — GitHub repository containing the commands and walkthrough used in the video.

External References

Contribution & Novelties

The video’s original contribution lies in demonstrating a practical multi-AI terminal workflow, combining Gemini CLI, Claude Code, and Codex with shared context files and agents. It popularizes the concept of owning your AI context on your local hard drive, contrasting with browser-based lock-in. The presentation of agents as a method to preserve context window and delegate tasks is particularly insightful.

Pour aller plus loin :

  • Model Context Protocol (MCP) — An open standard for connecting AI tools to data sources, enabling interoperability between terminal AI and external services. Highly relevant to the ecosystem shown.
  • Obsidian — A markdown note-taking app that stores notes as plain files, which can be used as the project context repository demonstrated in the video.
  • Git — The version control system used by the presenter to track changes in his project files, a key component of his workflow.
  • Ollama — A tool for running local LLMs, mentioned as an alternative for privacy-conscious users, aligning with the video’s discussion of local models.

165 words

Radar Profile

The radar profile shows strong scores across all dimensions, with quantitative information and practical relevance being particularly high, reflecting the video's comprehensive tutorial nature. The technical level is intermediate to advanced, suitable for viewers with some command-line experience. The reliability score is slightly lower due to anecdotal claims, but the overall profile indicates a high-quality, informative resource.

Reliability 7/10

💬 Très positif. Sur les 30 commentaires analysés, la majorité exprime une grande gratitude, qualifiant la vidéo de 'eye-opener' et 'most useful', certains mentionnant un changement immédiat dans leur façon de travailler. Plusieurs commentaires demandent un tutoriel complémentaire sur l'isolation de ces outils dans des conteneurs pour des raisons de sécurité. Quelques commentaires notent la sincérité du créateur, notamment la prière de fin, avec une réaction émotionnelle positive.