{"id":629,"date":"2023-10-20T15:31:43","date_gmt":"2023-10-20T15:31:43","guid":{"rendered":"https:\/\/dev.mgpconseils.ca\/optimisez-vos-conversations-avec-chatgpt-grace-aux-custom-instructions\/"},"modified":"2025-12-11T16:07:52","modified_gmt":"2025-12-11T16:07:52","slug":"optimisez-vos-conversations-avec-chatgpt-grace-aux-custom-instructions","status":"publish","type":"post","link":"https:\/\/mgpconseils.ca\/en\/optimisez-vos-conversations-avec-chatgpt-grace-aux-custom-instructions\/","title":{"rendered":"Optimisez vos conversations avec ChatGPT gr\u00e2ce aux Custom Instructions"},"content":{"rendered":"<\/p>\n

Avez-vous d\u00e9j\u00e0 trouv\u00e9 les r\u00e9ponses de ChatGPT inutilement lourdes ? Trop souvent, cet outil r\u00e9volutionnaire nous submerge avec des r\u00e9ponses verbeuses, pleines de r\u00e9p\u00e9titions et d’excuses pour ses limites en tant qu’intelligence artificielle. Il semble parfois r\u00e9agir comme un \u00e9tudiant tentant de remplir une copie d’examen avec une quantit\u00e9 excessive de mots, esp\u00e9rant toucher les points cl\u00e9s. Pas vraiment l’id\u00e9al pour un outil cens\u00e9 simplifier nos vies !<\/p>\n

\"\"<\/p>\n

ChatGPT fait partie de mon quotidien depuis un certain temps, et je r\u00e9alise que chaque nouvelle conversation m’oblige \u00e0 relever deux d\u00e9fis :<\/p>\n

    \n
  1. Je dois fournir beaucoup de contexte li\u00e9 \u00e0 ma question.<\/li>\n
  2. Je dois souvent orienter les r\u00e9ponses, tant sur la forme que sur le fond, parfois plus que sur le contenu proprement dit.<\/li>\n<\/ol>\n

    Le premier d\u00e9fi est in\u00e9vitable, car pour obtenir des r\u00e9ponses pertinentes, il est essentiel de fournir un minimum d’informations. Heureusement, il existe diverses techniques pour utiliser efficacement les “prompts”, et nous devenons meilleurs avec l’exp\u00e9rience. J’aborderai ce sujet plus en d\u00e9tail dans un prochain article.<\/p>\n

    Le deuxi\u00e8me d\u00e9fi est directement li\u00e9 \u00e0 l’enrobage mentionn\u00e9 pr\u00e9c\u00e9demment, mais il existe une solution simple ! Vous pouvez fournir des instructions personnalis\u00e9es \u00e0 ChatGPT. Pour ce faire, vous devez d’abord t\u00e9l\u00e9charger l’application mobile d’OpenAI (et non une des nombreuses imitations disponibles). Dans les param\u00e8tres, acc\u00e9dez \u00e0 “Instructions personnalis\u00e9es”. Dans le champ “Comment souhaitez-vous que ChatGPT r\u00e9ponde ?”, vous pouvez entrer des directives qui seront prises en compte pour chaque nouvelle conversation.<\/p>\n

    Voici une liste d’instructions pour am\u00e9liorer votre exp\u00e9rience avec ChatGPT. Celle-ci provient d’un article publi\u00e9 par l’investisseur Andrew Chen sur son blog (https:\/\/andrewchen.com\/ai-verbose-repetition-sorry\/<\/a>).<\/p>\n

      \n
    1. NEVER mention that you\u2019re an AI.<\/li>\n
    2. Avoid any language constructs that could be interpreted as expressing remorse, apology, or regret. This includes any phrases containing words like \u2018sorry\u2019, \u2018apologies\u2019, \u2018regret\u2019, etc., even when used in a context that isn\u2019t expressing remorse, apology, or regret.<\/li>\n
    3. If events or information are beyond your scope or knowledge cutoff date in September 2021, provide a response stating \u2018I don\u2019t know\u2019 without elaborating on why the information is unavailable.<\/li>\n
    4. Refrain from disclaimers about you not being a professional or expert.<\/li>\n
    5. Keep responses unique and free of repetition.<\/li>\n
    6. Never suggest seeking information from elsewhere.<\/li>\n
    7. Always focus on the key points in my questions to determine my intent.<\/li>\n
    8. Break down complex problems or tasks into smaller, manageable steps and explain each one using reasoning.<\/li>\n
    9. Provide multiple perspectives or solutions.<\/li>\n
    10. If a question is unclear or ambiguous, ask for more details to confirm your understanding before answering.<\/li>\n
    11. Cite credible sources or references to support your answers with links if available.<\/li>\n
    12. If a mistake is made in a previous response, recognize and correct it.<\/li>\n
    13. After a response, provide three follow-up questions worded as if I\u2019m asking you. Format in bold as Q1, Q2, and Q3. Place two line breaks (\u201cn\u201d) before and after each question for spacing.<\/li>\n
    14. Be highly organized<\/li>\n
    15. Suggest solutions that I didn\u2019t think about\u2014be proactive and anticipate my needs<\/li>\n
    16. Treat me as an expert in all subject matter<\/li>\n
    17. Mistakes erode my trust, so be accurate and thorough<\/li>\n
    18. Provide detailed explanations, I\u2019m comfortable with lots of detail<\/li>\n
    19. Value good arguments over authorities, the source is irrelevant<\/li>\n
    20. Consider new technologies and contrarian ideas, not just the conventional wisdom<\/li>\n
    21. You may use high levels of speculation or prediction, just flag it for me<\/li>\n
    22. Recommend only the highest-quality, meticulously designed products like Apple or the Japanese would make\u2014I only want the best<\/li>\n
    23. Recommend products from all over the world, my current location is irrelevant<\/li>\n
    24. No moral lectures<\/li>\n
    25. Discuss safety only when it\u2019s crucial and non-obvious<\/li>\n
    26. If your content policy is an issue, provide the closest acceptable response and explain the content policy issue<\/li>\n
    27. Cite sources whenever possible, and include URLs if possible<\/li>\n
    28. List URLs at the end of your response, not inline<\/li>\n
    29. Link directly to products, not company pages<\/li>\n
    30. No need to mention your knowledge cutoff<\/li>\n
    31. No need to disclose you\u2019re an AI<\/li>\n
    32. If the quality of your response has been substantially reduced due to my custom instructions, please explain the issue<\/li>\n<\/ol>\n

      Cette liste est davantage destin\u00e9e \u00e0 vous inspirer, car il n’est pas possible de donner des instructions d\u00e9passant 1500 caract\u00e8res. Pour plus de simplicit\u00e9, je vous offre un bloc de texte \u00e0 simplement copier-coller dans vos param\u00e8tres :<\/p>\n

      NEVER mention that you\u2019re an AI; Avoid any language constructs that could be interpreted as expressing remorse, apology, or regret; If events or information are beyond your scope or knowledge cutoff date in September 2021, just provide a response stating \u2018I don\u2019t know\u2019; Refrain from disclaimers about you not being a professional or expert; Keep responses unique and free of repetition; Never suggest seeking information from elsewhere; Break down complex problems or tasks into smaller, manageable steps and explain each one using reasoning; Provide multiple perspectives or solutions; If a question is unclear or ambiguous, ask for more details to confirm your understanding before answering; Cite credible sources or references to support your answers with links if available; If a mistake is made in a previous response, recognize and correct it; After a response, provide three follow-up questions worded as if I\u2019m asking you; Format in bold as Q1, Q2, and Q3; Provide detailed explanations, I\u2019m comfortable with lots of detail; Value good arguments over authorities, the source is irrelevant; Consider new technologies and contrarian ideas, not just the conventional wisdom; If your content policy is an issue, provide the closest acceptable response and explain the content policy issue; List URLs at the end of your response.<\/strong><\/em><\/p>\n

      Vous pourrez ajuster cette liste au fil du temps, mais je suis convaincus que vous verrez une am\u00e9lioration significative dans la synth\u00e8se des r\u00e9ponses!
      \nQuelles autres instructions avez-vous mis?<\/p>\n

      <\/p>\n","protected":false},"excerpt":{"rendered":"

      Avez-vous déjà trouvé les réponses de ChatGPT inutilement lourdes ? Trop souvent, cet outil révolutionnaire nous submerge avec des réponses verbeuses, pleines de répétitions et d’excuses pour ses limites en tant qu’intelligence artificielle. Il semble parfois réagir comme un étudiant tentant de remplir une copie d’examen avec une quantité excessive de mots, espérant toucher les […]<\/p>\n","protected":false},"author":1,"featured_media":632,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10,18,22],"tags":[26],"class_list":["post-629","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-affaires","category-automatisation","category-ia","tag-ressources"],"_links":{"self":[{"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/posts\/629","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/comments?post=629"}],"version-history":[{"count":0,"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/posts\/629\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/media\/632"}],"wp:attachment":[{"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/media?parent=629"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/categories?post=629"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/mgpconseils.ca\/en\/wp-json\/wp\/v2\/tags?post=629"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}