J'ai utilisé ChatGPT tous les jours pendant 1 an — Voici ce que j'ai appris

J'ai utilisé ChatGPT tous les jours pendant 1 an — Voici ce que j'ai appris

🎙 Ludo Salenne 👥 267K 📅 November 21, 2023 ⏱ 29 min 👁 85K 📄 expert opinion 🧭 2026-08-21
Available in: English (current) Français

Keywords

ChatGPTprompt engineeringAI biasGPTsautomation

Summary

In this video, Ludo Salenne shares his experience of using ChatGPT daily for a year, highlighting key lessons for effective use. He emphasizes the importance of avoiding bias in prompts, as language models predict words based on probabilities rather than reasoning. He demonstrates how adding irrelevant information can skew results, using examples like a power outage and a fabricated text. He then explains the significance of crafting high-quality prompts, introducing a framework of role, context, task, and desired characteristics. He warns against relying on pre-made solutions like plugins and GPTs, showing a deliberately poorly configured GPT that produces nonsensical content. He advocates for learning to do tasks yourself and understanding the fundamentals of AI. He also discusses the importance of breaking complex tasks into smaller steps (prompt chaining) for better results, and mentions that beta features may occasionally bug, advising persistence. The video includes a promotional segment for his training course.

151 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides valuable, practical insights based on the creator’s extensive experience with ChatGPT. The argumentation is solid, supported by concrete demonstrations and examples that illustrate the points effectively. The advice on avoiding bias, crafting detailed prompts, and chaining tasks is actionable and well-explained. The warning about the potential dangers of poorly configured GPTs is particularly relevant. However, the argumentation is largely anecdotal and lacks scientific rigor, relying on personal experience rather than empirical studies.

Scientific Rigor, Source Quality, Title Accuracy

The video does not cite external scientific sources, but it references the creator’s own tutorials and training course. The title accurately reflects the content. The demonstrations are clear and serve as practical evidence. The promotional segment for the training course is transparent but may introduce bias. Overall, the scientific rigor is moderate, typical of an expert opinion video rather than a research-based presentation.

153 words

Title / Content Match

The title accurately reflects the content: a personal retrospective of one year of daily ChatGPT use, sharing lessons learned and practical tips.

Quality & Reliability

7/10

The video offers practical, experience-based advice on using ChatGPT, with concrete examples and demonstrations. However, it lacks rigorous scientific sourcing and includes promotional content for the creator's own training course, which may introduce bias.

Chapters

Cited Sources

Concurring Sources

Dissenting Sources

  • AI hallucination — The video demonstrates AI hallucination but does not explicitly discuss the concept, which is a known limitation of language models.

Contribution & Novelties

The video offers a practical, experience-based perspective on using ChatGPT effectively, emphasizing the importance of prompt engineering and caution with pre-built tools. It provides concrete examples of common pitfalls and demonstrates a structured approach to crafting prompts. The warning about the potential dangers of poorly configured GPTs is a valuable addition to the discourse.

Pour aller plus loin :

  • Prompt engineering — Overview of techniques for designing effective prompts.
  • Language model — Explanation of how language models predict text.
  • OpenAI — Official site for ChatGPT and GPT models, providing documentation and updates.

92 words

Radar Profile

The radar profile shows high scores in information quantity and technical level, reflecting the video's practical depth. Quality and reliability are moderate, consistent with an opinion-based presentation. The overall balance indicates a useful but not scientifically rigorous resource.

Reliability 6/10