Mon Stack IA ! Quel est le vôtre ?...

Value of the Information & Strength of the Argument

The video’s value lies in the practical, experience-based demonstration of how AI tools can be integrated into a creative and professional workflow. The creator clearly argues that tool effectiveness is context-dependent, not universal, and supports this with specific examples of his own tasks. He demonstrates a thoughtful approach to tool selection, explaining why he uses one tool over another based on its strengths (e.g., Perplexity for research, Claude for writing). The argumentation is logical and honest, including admissions of tools he recognizes are useful but does not personally adopt (e.g., NotebookLM). However, the content remains subjective and lacks empirical comparison or controlled testing, relying entirely on personal preference and anecdotal experience. The argument would be stronger with benchmarks or user studies, but as an expert opinion, it is well-reasoned and reasonably persuasive.

Scientific Rigor, Source Quality, Title Accuracy

The title accurately reflects the content, which is a personal account of the AI tools the creator uses, and it invites audience participation. Scientific rigor is low: the video offers no external citations, data, or systematic reviews; it is a purely opinion-based sharing of personal workflow. The only source mentioned is the creator’s own website (renaud-dekode.fr), which cannot be considered an authoritative reference. The content is internally consistent and the arguments are coherent, but the reliability is limited to subjective experience. The video does not claim to be scientific, so it is appropriate for its purpose, though viewers seeking evidence-based comparisons will be disappointed.

The title directly matches the content: the presenter shares his personal AI tool stack and invites viewers to share theirs, as described in the video.

6/10

The content is a subjective, experience-based overview of AI tools used by the creator in a professional and creative context. It offers clear, practical insights but lacks rigorous scientific validation or external citation, making its reliability moderate.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 00:30Introduction: the creator announces he will present his AI tool stack and emphasizes that tool choice depends on individual use cases.
• 02:30Discusses ChatGPT GPT-5 Thinking as his primary tool for analysis, brainstorming, and reflective tasks, and explains his one-shot prompting strategy.
• 05:00Introduces Perplexity as his main research and fact-checking tool, highlighting its usefulness for sourcing and its Comet browser mode.
• 07:00Mentions Feedly for information aggregation and AI-assisted sorting, and NotebookLM for learning (with the caveat that he rarely uses it).
• 09:00Explains his use of Claude Sonnet 4.5 for writing, spreadsheet work, and generating formatted documents, and Gemini 2.5 Pro for handling large contexts and Google integrations.
• 11:00Talks about development tools: Claude Code (web) and Gemini 2.5 Pro for command-line coding, and Hostinger Horizon for vibe coding web apps.
• 13:00Covers visual content creation: Midjourney for inspiration, Flux Kontext Pro for image generation, and Canva for layout composition.
• 15:00Describes video editing with CapCut and occasional use of Runway for AI video effects, plus Suno for music generation.
• 17:00Mentions local AI models via LM Studio and AnythingLLM, including GPT-OS 20B and Mistral Small, and his automation stack using Make and n8n.
• 20:00Concludes by inviting viewers to share their own AI stacks on his website, reiterating that usage determines the ideal toolset.

• Renaud Dékode's website — The creator refers viewers to his website for sharing their AI stacks and accessing additional content related to AI tools.