ผมใช้ Claude ลดเวลาทำงาน 4 ชม ต่อวัน! ด้วย Workflow นี้...

Summary

The video demonstrates how Ta Thanakrit uses Claude Code to streamline his YouTube content workflow. He says that researching, planning, writing scripts, and preparing content—which previously took about two days—can now be completed in roughly four hours, including setup. He still reviews and approves the AI’s work.

Workflow and Setup

• Create a project folder on the computer to hold the workflow and its materials.
• Use Markdown files as instructions and reference material for the AI. The speaker describes them as an AI-readable version of an SOP: they explain who he is, how he works, and what steps to follow.
• Organize context into separate files or folders—for example, YouTube ideas, business information, brainstorming notes, and a voice guide describing the desired tone.
• Ask Claude Code to research a topic, brainstorm, and create a script. Save the resulting files in the project folder. The speaker also discusses using NotebookLM for research and reference.
• Review the research and draft, request changes if needed, and approve the final output.

Using Claude Code

The video shows Claude Code being used through the terminal, which the speaker prefers, while noting that a desktop app is also available. The general setup shown is to download the appropriate installer, open Terminal, launch Claude Code, sign in, and allow it to access the project folder. Some installation details in the subtitles appear garbled.

One practical point is to keep important instructions and outputs in project files: terminal chat history may not be retained when a session is closed.

Building a Workflow for Other Work

The speaker recommends describing the task and desired workflow to Claude, then asking it to pose questions before starting so it can clarify requirements and connections. The workflow can be refined over time as the user updates it. He suggests the same approach may apply to tasks such as design, accounting, business operations, and other repeatable work.

Learning Resources Mentioned

• Greg’s YouTube explanations of Claude Code and its files.
• Longer videos referred to as “NK/NHK” in the subtitles.
• Joh Kim’s material, described as being divided into smaller courses.

Main Speakers and Sources

• Ta Thanakrit — presenter and creator of the demonstrated workflow.
• Greg, “NK/NHK,” and Joh Kim — learning resources mentioned by the presenter; the names are uncertain because of auto-generated subtitles.