22 minutes pour bien utiliser ChatGPT en 2026 (Tuto Complet)

22 minutes pour bien utiliser ChatGPT en 2026 (Tuto Complet)

by Ludo Salenne

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 Ludo Salenne 👥 266K 📅 January 8, 2026 ⏱ 22 min 👁 24K 📄 tutorial 🧭 2026-08-02
Available in: English (current) Français

Keywords

ChatGPTpromptRCTCARTELGPTs

Summary

This tutorial by Ludo Salenne aims to teach viewers how to use ChatGPT effectively in 2026. The video begins by highlighting a common mistake: treating ChatGPT like a friend on WhatsApp, which leads to generic responses. The author explains that ChatGPT is a predictive model, and the quality of output depends on the precision of the input, comparing it to GPS coordinates. He introduces two prompt methods: the simple RCT (Role, Context, Task) and the advanced CARTEL (Context, Audience, Role, Task, Examples, Deliverable). He demonstrates these methods with a practical example of creating a job posting for a community manager. To save time, he recommends using custom GPTs to automate prompt creation, and he offers a free guide and a paid formation. He also discusses the importance of customizing ChatGPT’s settings, such as tone and style, to get more professional responses. The video covers additional features like image generation, video generation, and the new Agent and Apps capabilities. It concludes with a recommended workflow for using ChatGPT in 2026, emphasizing the importance of clear communication and leveraging advanced features.

179 words

Critical Evaluation

The video provides a solid, practical introduction to using ChatGPT effectively. The author’s main argument is that users should treat ChatGPT as a tool that requires precise instructions, not as a conversational partner. This is a valuable insight, as many users indeed get suboptimal results due to vague prompts. The RCT and CARTEL methods are structured and easy to follow, and the example of creating a job posting illustrates their application well. The author also addresses the common concern about time investment by suggesting the use of custom GPTs to automate prompt creation, which is a clever solution.

The scientific rigor is moderate. The author explains that ChatGPT predicts words based on input, which is accurate, and he uses the GPS analogy to convey the importance of precise input. However, he does not delve into the underlying technology or cite external research. The video is more of an expert opinion based on practical experience than a scientifically rigorous analysis. The author does mention a video about how ChatGPT affects the brain, but he does not provide scientific evidence in this tutorial.

The quality of sources is limited. The video includes links to the author’s own products (a paid formation and a free guide) and to other tutorials on his channel. While these are relevant, they are not independent sources. The author does not cite any academic papers or official documentation from OpenAI, which would have strengthened the credibility.

The adéquation between title and content is good: the video delivers a comprehensive tutorial on using ChatGPT, covering prompt engineering, settings, and advanced features. The title accurately reflects the content.

The video contains a promotional segment for the author’s formation and guide, which is clearly marked as such. This does not detract from the educational value, but it is worth noting.

Overall, the video is a useful resource for beginners and intermediate users who want to improve their ChatGPT interactions. The methods are practical and likely to yield better results. However, for a more rigorous understanding, viewers should consult official documentation and academic sources.

342 words

Title / Content Match

The title accurately reflects the content: a 22-minute comprehensive tutorial on using ChatGPT effectively in 2026.

Quality & Reliability

7/10

The video is a practical tutorial based on the author's experience, offering structured prompt methods (RCT, CARTEL) and practical tips. It includes some promotional content for the author's own products, but the core advice is actionable and generally aligns with common best practices for using ChatGPT. The scientific basis is limited, but the methods are reasonable and widely applicable.

Chapters

Cited Sources

Concurring Sources

Dissenting Sources

  • No direct discordant sources found — The video's advice is generally consistent with common practices, but it lacks scientific citations. Some might argue that the CARTEL method is not empirically validated, but it is a reasonable heuristic.

Contribution & Novelties

The video offers a structured approach to prompt engineering with the RCT and CARTEL methods, which are practical and easy to apply. It also emphasizes the importance of customizing ChatGPT settings and using custom GPTs to automate prompt creation, which can significantly improve efficiency. The author provides a free guide and a paid formation for further learning.

Pour aller plus loin :

111 words

Radar Profile

The radar profile shows high scores in quantity of information and quality of information, indicating a content-rich tutorial. The technical level is moderate, suitable for beginners, and the reliability is good but not exceptional due to the lack of external sources.

Reliability 7/10