{"users":[{"id":16009,"username":"logankilpatrick","name":"Logan Kilpatrick","avatar_template":"/user_avatar/community.openai.com/logankilpatrick/{size}/14116_2.png","trust_level":1},{"id":748766,"username":"Daller","name":"Daller","avatar_template":"/user_avatar/community.openai.com/daller/{size}/436549_2.png","trust_level":3},{"id":1214162,"username":"LarisaHaster","name":"Larisa Haster ","avatar_template":"/user_avatar/community.openai.com/larisahaster/{size}/750520_2.png","flair_name":"Image_Enthusiasts","flair_url":"https://us1.discourse-cdn.com/openai1/original/4X/4/0/8/40887ee06450d296501a75ae7c3d51ff7b38f19c.webp","flair_group_id":56,"trust_level":3},{"id":632960,"username":"jeffvpace","name":"Vern","avatar_template":"/user_avatar/community.openai.com/jeffvpace/{size}/741732_2.png","trust_level":3},{"id":72385,"username":"VeitB","name":"","avatar_template":"/user_avatar/community.openai.com/veitb/{size}/712987_2.png","primary_group_name":"Community-Moderators","moderator":true,"trust_level":4},{"id":726999,"username":"Tina_ChiKa","name":"Thalhofer Bettina","avatar_template":"/user_avatar/community.openai.com/tina_chika/{size}/597547_2.png","trust_level":2},{"id":1213959,"username":"vittoroliveira","name":"Vittor Oliveira","avatar_template":"/user_avatar/community.openai.com/vittoroliveira/{size}/738410_2.png","trust_level":1},{"id":228799,"username":"_j","name":"Jay","avatar_template":"/user_avatar/community.openai.com/_j/{size}/719663_2.png","trust_level":3},{"id":1337172,"username":"emc2","name":"Elon Colello","avatar_template":"/user_avatar/community.openai.com/emc2/{size}/754624_2.png","trust_level":1},{"id":1093386,"username":"recursionrecursion1","name":"Recursionrecursion1","avatar_template":"/user_avatar/community.openai.com/recursionrecursion1/{size}/639099_2.png","trust_level":2},{"id":1079,"username":"PaulBellow","name":"Paul Bellow","avatar_template":"/user_avatar/community.openai.com/paulbellow/{size}/597962_2.png","primary_group_name":"Community-Moderators","flair_name":"Community-Moderators","flair_url":"shield-halved","flair_group_id":47,"moderator":true,"trust_level":4},{"id":412409,"username":"VivacityDesign","name":"alessandro demontis (ContainerMan)","avatar_template":"/user_avatar/community.openai.com/vivacitydesign/{size}/230492_2.png","trust_level":2},{"id":699820,"username":"mitchell_d00","name":"Mitchell D McPhetridge","avatar_template":"/user_avatar/community.openai.com/mitchell_d00/{size}/458995_2.png","trust_level":2},{"id":561234,"username":"kavitatipnis","name":"Kavita Tipnis Rasal","avatar_template":"/user_avatar/community.openai.com/kavitatipnis/{size}/313326_2.png","primary_group_name":"tl3","flair_name":"tl3","flair_group_id":49,"trust_level":3},{"id":464733,"username":"platypus","name":"Armin Catovic","avatar_template":"/user_avatar/community.openai.com/platypus/{size}/443080_2.png","trust_level":3},{"id":792,"username":"sergeliatko","name":"Serge Liatko","avatar_template":"/user_avatar/community.openai.com/sergeliatko/{size}/816_2.png","trust_level":3},{"id":1337325,"username":"danylokv","name":"danylokv","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/d/7c8e57/{size}.png","trust_level":1},{"id":1009035,"username":"windysoliloquy","name":"Espresso Bean","avatar_template":"/user_avatar/community.openai.com/windysoliloquy/{size}/738470_2.png","trust_level":2},{"id":1303096,"username":"erkin","name":"erkin","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/e/3ec8ea/{size}.png","trust_level":1},{"id":219175,"username":"aprendendo.next","name":"aprendendo","avatar_template":"/user_avatar/community.openai.com/aprendendo.next/{size}/721758_2.png","trust_level":3},{"id":704629,"username":"phyde1001","name":"Peter Hyde","avatar_template":"/user_avatar/community.openai.com/phyde1001/{size}/711957_2.png","trust_level":2},{"id":1198588,"username":"taras_shebek","name":"taras shebek","avatar_template":"/user_avatar/community.openai.com/taras_shebek/{size}/467426_2.png","trust_level":0},{"id":17132,"username":"merefield","name":"Robert Barrow","avatar_template":"/user_avatar/community.openai.com/merefield/{size}/14812_2.png","trust_level":3},{"id":40833,"username":"EricGT","name":"EricGT","avatar_template":"/user_avatar/community.openai.com/ericgt/{size}/20571_2.png","trust_level":4},{"id":1090974,"username":"Stephen_Braniff","name":"Stephen Braniff","avatar_template":"/user_avatar/community.openai.com/stephen_braniff/{size}/622431_2.png","trust_level":2},{"id":1326918,"username":"JozeRizel","name":"Joze Rizel","avatar_template":"/user_avatar/community.openai.com/jozerizel/{size}/748237_2.png","trust_level":0},{"id":1298105,"username":"Alex_Kondor1","name":"Alex Kondor","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/a/8edcca/{size}.png","trust_level":1},{"id":1335721,"username":"Bloxy","name":"Bloxy","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/b/77aa72/{size}.png","trust_level":0},{"id":1217949,"username":"Tyler_Suard","name":"Tyler Suard","avatar_template":"/user_avatar/community.openai.com/tyler_suard/{size}/691408_2.png","trust_level":0},{"id":1334478,"username":"Oscararce_Martinez","name":"Oscararce Martinez","avatar_template":"/user_avatar/community.openai.com/oscararce_martinez/{size}/752968_2.png","trust_level":0},{"id":11799,"username":"benlamm","name":"Ben Lamm","avatar_template":"/user_avatar/community.openai.com/benlamm/{size}/10046_2.png","trust_level":2},{"id":233,"username":"daveshapautomator","name":"David Shapiro","avatar_template":"/user_avatar/community.openai.com/daveshapautomator/{size}/8779_2.png","trust_level":2},{"id":10407,"username":"jhsmith12345","name":"Hanley Smith","avatar_template":"/user_avatar/community.openai.com/jhsmith12345/{size}/8785_2.png","trust_level":2},{"id":1271867,"username":"eddiesamule250","name":"Eddie Samule","avatar_template":"/user_avatar/community.openai.com/eddiesamule250/{size}/451560_2.png","trust_level":0},{"id":4078,"username":"Foxalabs","name":"Spencer Bentley","avatar_template":"/user_avatar/community.openai.com/foxalabs/{size}/738355_2.png","primary_group_name":"Community-Moderators","flair_name":"Community-Moderators","flair_url":"shield-halved","flair_group_id":47,"moderator":true,"trust_level":4},{"id":1275854,"username":"Blogger24","name":"Bloggerid24","avatar_template":"/user_avatar/community.openai.com/blogger24/{size}/512202_2.png","trust_level":0},{"id":509645,"username":"DysTopia","name":"Dys Topia","avatar_template":"/user_avatar/community.openai.com/dystopia/{size}/711735_2.png","flair_name":"Image_Enthusiasts","flair_url":"https://us1.discourse-cdn.com/openai1/original/4X/4/0/8/40887ee06450d296501a75ae7c3d51ff7b38f19c.webp","flair_group_id":56,"trust_level":2},{"id":1160834,"username":"Samuel_Jonsson","name":"Samuel Jonsson","avatar_template":"/user_avatar/community.openai.com/samuel_jonsson/{size}/661965_2.png","trust_level":1},{"id":1309291,"username":"T.Maillet","name":"","avatar_template":"/user_avatar/community.openai.com/t.maillet/{size}/741809_2.png","trust_level":2},{"id":1305684,"username":"Baseballguru","name":"","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/b/13edae/{size}.png","trust_level":1},{"id":1307427,"username":"justanotheruser","name":"","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/j/779978/{size}.png","trust_level":0},{"id":1324650,"username":"aisystemsdev","name":"","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/a/d2c977/{size}.png","trust_level":1},{"id":1185511,"username":"Vinny1","name":"","avatar_template":"/user_avatar/community.openai.com/vinny1/{size}/516760_2.png","trust_level":0},{"id":1216273,"username":"jukp","name":"Just Pedersen","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/j/bbce88/{size}.png","trust_level":0},{"id":1317878,"username":"ethanAI","name":"Ethan Ross","avatar_template":"/user_avatar/community.openai.com/ethanai/{size}/742977_2.png","trust_level":2},{"id":478475,"username":"aureliodavinci","name":"AV","avatar_template":"/user_avatar/community.openai.com/aureliodavinci/{size}/266076_2.png","trust_level":2},{"id":1322193,"username":"TIME3","name":"ROME","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/t/5fc32e/{size}.png","trust_level":0},{"id":834641,"username":"cyborgsdream333","name":"Cybernoid","avatar_template":"/user_avatar/community.openai.com/cyborgsdream333/{size}/755136_2.png","trust_level":2},{"id":1318263,"username":"isabella11","name":"Isabella","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/i/e47c2d/{size}.png","trust_level":1},{"id":1310407,"username":"keesan","name":"","avatar_template":"/user_avatar/community.openai.com/keesan/{size}/740295_2.png","trust_level":1},{"id":1317192,"username":"FATIMA21","name":"","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/f/4af34b/{size}.png","trust_level":0},{"id":617918,"username":"darthzini","name":"Sini Tikkala","avatar_template":"/user_avatar/community.openai.com/darthzini/{size}/343706_2.png","trust_level":0},{"id":1307570,"username":"lnobrega","name":"Luís Nóbrega","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/l/ea5d25/{size}.png","trust_level":2},{"id":1301342,"username":"Mark_G","name":"","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/m/e8c25b/{size}.png","primary_group_name":"ACN_T1_team","flair_name":"ACN_T1_team","flair_group_id":55,"moderator":true,"trust_level":3},{"id":269430,"username":"OpenAI_Support","name":"OpenAI Support","avatar_template":"/user_avatar/community.openai.com/openai_support/{size}/561845_2.png","admin":true,"trust_level":4},{"id":1309252,"username":"HanakoAbsed","name":"useless","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/h/8edcca/{size}.png","trust_level":1},{"id":1302247,"username":"rafa3","name":"Rafa","avatar_template":"/user_avatar/community.openai.com/rafa3/{size}/740294_2.png","trust_level":1},{"id":-1,"username":"system","name":"system","avatar_template":"https://us1.discourse-cdn.com/openai1/original/4X/8/7/c/87c2e81543bbb01aa9641aeef8fe14cab5e9590a.png","admin":true,"moderator":true,"trust_level":4},{"id":1293283,"username":"Balanced_Tiger","name":"Tiger","avatar_template":"/user_avatar/community.openai.com/balanced_tiger/{size}/731372_2.png","trust_level":1},{"id":1308414,"username":"Just_a_human","name":"","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/j/6de8d8/{size}.png","trust_level":1},{"id":1309855,"username":"Agentify.sh","name":"Agentify.sh","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/a/48db29/{size}.png","trust_level":2},{"id":1306030,"username":"HarshEntrepre","name":"Harsh Patel","avatar_template":"/user_avatar/community.openai.com/harshentrepre/{size}/738389_2.png","trust_level":0},{"id":940211,"username":"richard547","name":"Yaroslav","avatar_template":"/user_avatar/community.openai.com/richard547/{size}/624121_2.png","trust_level":2},{"id":1304941,"username":"jerryshah","name":"jerry","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/j/9dc877/{size}.png","trust_level":0},{"id":-3,"username":"discourse_ai_spam","name":"Discourse AI Spam Scanner","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/d/c68b51/{size}.png","admin":true,"trust_level":4}],"primary_groups":[{"id":47,"name":"Community-Moderators"},{"id":49,"name":"tl3"},{"id":55,"name":"ACN_T1_team"}],"flair_groups":[{"id":56,"name":"Image_Enthusiasts","flair_url":"https://us1.discourse-cdn.com/openai1/original/4X/4/0/8/40887ee06450d296501a75ae7c3d51ff7b38f19c.webp","flair_bg_color":"","flair_color":""},{"id":47,"name":"Community-Moderators","flair_url":"shield-halved","flair_bg_color":"","flair_color":""},{"id":49,"name":"tl3","flair_url":null,"flair_bg_color":"","flair_color":""},{"id":55,"name":"ACN_T1_team","flair_url":null,"flair_bg_color":"","flair_color":""}],"topic_list":{"can_create_topic":false,"more_topics_url":"/c/prompting/8?page=1","per_page":30,"top_tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"},{"id":10,"name":"gpt-4","slug":"gpt-4"},{"id":33,"name":"api","slug":"api"},{"id":15,"name":"gpt-35-turbo","slug":"gpt-35-turbo"},{"id":113,"name":"prompt","slug":"prompt"},{"id":176,"name":"prompt-engineering","slug":"prompt-engineering"},{"id":296,"name":"dalle3","slug":"dalle3"},{"id":355,"name":"assistants-api","slug":"assistants-api"},{"id":28,"name":"fine-tuning","slug":"fine-tuning"},{"id":42,"name":"image-generation","slug":"image-generation"},{"id":248,"name":"prompting","slug":"prompting"},{"id":20,"name":"plugin-development","slug":"plugin-development"},{"id":345,"name":"assistants","slug":"assistants"},{"id":41,"name":"dalle2","slug":"dalle2"},{"id":631,"name":"gpt-4o","slug":"gpt-4o"},{"id":199,"name":"function-calling","slug":"function-calling"},{"id":349,"name":"gpts","slug":"gpts"},{"id":314,"name":"dalle","slug":"dalle"},{"id":220,"name":"custom-instructions","slug":"custom-instructions"},{"id":268,"name":"rag","slug":"rag"},{"id":50,"name":"chatgpt-plugin","slug":"chatgpt-plugin"},{"id":31,"name":"bug","slug":"bug"},{"id":52,"name":"playground","slug":"playground"},{"id":312,"name":"gpt-4-vision","slug":"gpt-4-vision"},{"id":683,"name":"gpt-4o-mini","slug":"gpt-4o-mini"},{"id":6,"name":"codex","slug":"codex"},{"id":284,"name":"dall-e-3","slug":"dall-e-3"},{"id":7,"name":"embeddings","slug":"embeddings"},{"id":121,"name":"prompts-as-code","slug":"prompts-as-code"},{"id":344,"name":"custom-gpt","slug":"custom-gpt"},{"id":69,"name":"research","slug":"research"}],"topics":[{"fancy_title":"About the Prompting category","id":36,"title":"About the Prompting category","slug":"about-the-prompting-category","posts_count":1,"reply_count":5,"highest_post_number":1,"image_url":null,"created_at":"2021-03-10T20:40:28.668Z","last_posted_at":"2021-03-10T20:40:28.674Z","bumped":true,"bumped_at":"2023-12-17T14:37:02.048Z","archetype":"regular","unseen":false,"pinned":true,"unpinned":null,"excerpt":"Learn more about prompting by sharing best practices, your favorite prompts, and more! \nPlease keep all prompts appropriate and within OpenAI’s TOS.","visible":true,"closed":true,"archived":false,"bookmarked":null,"liked":null,"tags":[],"tags_descriptions":{},"views":33822,"like_count":77,"has_summary":false,"last_poster_username":"logankilpatrick","category_id":8,"op_like_count":77,"pinned_globally":false,"featured_link":null,"ai_topic_gist":null,"has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest single","description":"Original Poster, Most Recent Poster","user_id":16009,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Collection of GPT-image-generator 2.0 issues, bugs, and work-around tips (check first post)","id":1379535,"title":"Collection of GPT-image-generator 2.0 issues, bugs, and work-around tips (check first post)","slug":"collection-of-gpt-image-generator-2-0-issues-bugs-and-work-around-tips-check-first-post","posts_count":344,"reply_count":270,"highest_post_number":356,"image_url":null,"created_at":"2026-04-22T17:03:11.369Z","last_posted_at":"2026-08-07T17:49:23.781Z","bumped":true,"bumped_at":"2026-08-07T17:49:23.781Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":42,"name":"image-generation","slug":"image-generation"},{"id":1133,"name":"gpt-image-2","slug":"gpt-image-2"}],"tags_descriptions":{},"views":20398,"like_count":707,"has_summary":true,"last_poster_username":"Daller","category_id":8,"op_like_count":12,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"OpenAI’s image team is reportedly working on GPT Image 2.0’s recurring noise-pattern issue, while community members warn that affected images need filtering from future training data to prevent the artifact from spreading.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":748766,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Frequent Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":72385,"primary_group_id":47,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":726999,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Having trouble getting transparent backgrounds in ChatGPT images","id":1380143,"title":"Having trouble getting transparent backgrounds in ChatGPT images","slug":"having-trouble-getting-transparent-backgrounds-in-chatgpt-images","posts_count":101,"reply_count":56,"highest_post_number":104,"image_url":null,"created_at":"2026-05-01T13:50:43.184Z","last_posted_at":"2026-08-03T01:45:38.966Z","bumped":true,"bumped_at":"2026-08-03T01:45:38.966Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"},{"id":42,"name":"image-generation","slug":"image-generation"}],"tags_descriptions":{},"views":11870,"like_count":170,"has_summary":true,"last_poster_username":"recursionrecursion1","category_id":8,"op_like_count":4,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Users report inconsistent alpha transparency and cropping in generated images, prompting calls for built-in manual editing; file inspection suggests some outputs already contain alpha, while JPEG cannot, and WebP or PNG plus careful conversion workflows are recommended.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1213959,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Frequent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":1337172,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1093386,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"AI prompts for sharper fiction editing for writers LLM writing + NLP","id":1363351,"title":"AI prompts for sharper fiction editing for writers LLM writing + NLP","slug":"ai-prompts-for-sharper-fiction-editing-for-writers-llm-writing-nlp","posts_count":27,"reply_count":12,"highest_post_number":27,"image_url":"https://us1.discourse-cdn.com/openai1/optimized/4X/d/2/4/d24036d163c5fe2e7b1661dc805bb0e171d21320_2_1024x608.png","created_at":"2025-10-21T18:56:42.638Z","last_posted_at":"2026-08-02T22:36:09.928Z","bumped":true,"bumped_at":"2026-08-02T22:36:09.928Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":273,"name":"ai-writing","slug":"ai-writing"}],"tags_descriptions":{},"views":6147,"like_count":102,"has_summary":false,"last_poster_username":"mitchell_d00","category_id":8,"op_like_count":14,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"PaulBellow’s collection of detailed AI prompts for fiction editing, including persona-based critique, story analysis, and weak-language detection, has gained positive community feedback and now includes a related YouTube resource.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1079,"primary_group_id":47,"flair_group_id":47},{"extras":null,"description":"Frequent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Frequent Poster","user_id":412409,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":699820,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Prompt Regression Testing - API Usage","id":1119299,"title":"Prompt Regression Testing - API Usage","slug":"prompt-regression-testing-api-usage","posts_count":12,"reply_count":8,"highest_post_number":12,"image_url":null,"created_at":"2025-02-13T00:57:28.468Z","last_posted_at":"2026-08-02T09:35:38.981Z","bumped":true,"bumped_at":"2026-08-02T09:35:38.981Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":33,"name":"api","slug":"api"},{"id":176,"name":"prompt-engineering","slug":"prompt-engineering"}],"tags_descriptions":{},"views":1370,"like_count":12,"has_summary":false,"last_poster_username":"danylokv","category_id":8,"op_like_count":2,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Prompt regressions remain difficult to prevent despite fixed context and models; practitioners pin dated model versions and manually run small criteria-based canary sets, while seeking simple scheduled live-endpoint monitoring with baseline scoring and Slack alerts.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":561234,"primary_group_id":49,"flair_group_id":49},{"extras":null,"description":"Frequent Poster","user_id":464733,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":792,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1337325,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Prompt Forge App v5.1 was bugged- critical flaws fixed in next release","id":1378339,"title":"Prompt Forge App v5.1 was bugged- critical flaws fixed in next release","slug":"prompt-forge-app-v5-1-was-bugged-critical-flaws-fixed-in-next-release","posts_count":44,"reply_count":23,"highest_post_number":46,"image_url":"https://us1.discourse-cdn.com/openai1/original/4X/9/2/5/925ad1f449645fcf4f730b675a169dcb75ba5b7e.jpeg","created_at":"2026-04-01T23:04:52.552Z","last_posted_at":"2026-08-01T20:36:49.540Z","bumped":true,"bumped_at":"2026-08-01T20:36:49.540Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"},{"id":33,"name":"api","slug":"api"},{"id":20,"name":"plugin-development","slug":"plugin-development"},{"id":30,"name":"project","slug":"project"},{"id":42,"name":"image-generation","slug":"image-generation"}],"tags_descriptions":{},"views":1788,"like_count":69,"has_summary":false,"last_poster_username":"PaulBellow","category_id":8,"op_like_count":7,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Prompt Forge’s developer reported fixing critical flaws and planning an updated release with reduced free prompts, while a recent commenter asks whether the full version was ever released.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1009035,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":72385,"primary_group_id":47,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":1303096,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1079,"primary_group_id":47,"flair_group_id":47}]},{"fancy_title":"Requesting one text element without additional text","id":1388677,"title":"Requesting one text element without additional text","slug":"requesting-one-text-element-without-additional-text","posts_count":11,"reply_count":2,"highest_post_number":11,"image_url":"https://us1.discourse-cdn.com/openai1/optimized/4X/0/b/6/0b6177eda8d32cac0b84053a383170b5bbdabef9_2_1024x768.jpeg","created_at":"2026-08-01T13:23:57.917Z","last_posted_at":"2026-08-01T15:43:58.783Z","bumped":true,"bumped_at":"2026-08-01T15:43:58.783Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":42,"name":"image-generation","slug":"image-generation"},{"id":829,"name":"gallery","slug":"gallery"},{"id":890,"name":"imagegen","slug":"imagegen"}],"tags_descriptions":{},"views":123,"like_count":34,"has_summary":false,"last_poster_username":"PaulBellow","category_id":8,"op_like_count":5,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"To generate an image with one requested text element and no extras, users recommend explicitly specifying “one sign only,” defining the exact wording, and prohibiting all other typography; detailed prompt preprocessing and occasional post-editing may improve reliability.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":72385,"primary_group_id":47,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":219175,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":704629,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":699820,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1079,"primary_group_id":47,"flair_group_id":47}]},{"fancy_title":"Memory-First Conversational Architecture as an Alternative to Long Context Windows","id":1375640,"title":"Memory-First Conversational Architecture as an Alternative to Long Context Windows","slug":"memory-first-conversational-architecture-as-an-alternative-to-long-context-windows","posts_count":10,"reply_count":5,"highest_post_number":10,"image_url":null,"created_at":"2026-03-04T08:22:28.168Z","last_posted_at":"2026-07-31T06:09:13.157Z","bumped":true,"bumped_at":"2026-07-31T06:09:13.157Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":33,"name":"api","slug":"api"}],"tags_descriptions":{},"views":1078,"like_count":3,"has_summary":false,"last_poster_username":"Stephen_Braniff","category_id":8,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Persistent memory using retrieval, multicall, RAG, and orchestration frameworks could support long-horizon conversations by dynamically shifting context, while ChatGPT’s limited long-form memory may reflect market demand rather than technical constraints.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1198588,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":17132,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":40833,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1090974,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Different answers for same prompt?","id":1385774,"title":"Different answers for same prompt?","slug":"different-answers-for-same-prompt","posts_count":12,"reply_count":1,"highest_post_number":12,"image_url":null,"created_at":"2026-07-05T11:28:37.767Z","last_posted_at":"2026-07-25T08:15:55.224Z","bumped":true,"bumped_at":"2026-07-25T08:15:55.224Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"}],"tags_descriptions":{},"views":585,"like_count":11,"has_summary":false,"last_poster_username":"Bloxy","category_id":8,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"ChatGPT responses vary because of built-in randomness and model updates; users can improve consistency by saving successful prompts, adding detailed context and personas, and specifying clear response constraints.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1326918,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Recent Poster","user_id":1298105,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1335721,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Ambiguous prompts can trigger unexpected refusals in GPT-5.6 Sol","id":1387780,"title":"Ambiguous prompts can trigger unexpected refusals in GPT-5.6 Sol","slug":"ambiguous-prompts-can-trigger-unexpected-refusals-in-gpt-5-6-sol","posts_count":4,"reply_count":0,"highest_post_number":4,"image_url":"https://us1.discourse-cdn.com/openai1/original/4X/1/e/1/1e10b7fc3a68ac668067e57f59fbd0d5dc3d29b4.png","created_at":"2026-07-22T04:16:50.067Z","last_posted_at":"2026-07-24T21:15:05.456Z","bumped":true,"bumped_at":"2026-07-24T21:15:05.456Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":176,"name":"prompt-engineering","slug":"prompt-engineering"}],"tags_descriptions":{},"views":238,"like_count":7,"has_summary":false,"last_poster_username":"Tyler_Suard","category_id":8,"op_like_count":4,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Users report that GPT-5.6 Sol sometimes misinterprets harmless but context-dependent prompts as policy violations, while others attribute refusals to ambiguity and recommend making requests explicit, actionable, and specifying the desired output format.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":1217949,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Why won’t image 7 control the pose?","id":1387960,"title":"Why won't image 7 control the pose?","slug":"why-wont-image-7-control-the-pose","posts_count":2,"reply_count":0,"highest_post_number":2,"image_url":null,"created_at":"2026-07-23T11:26:30.073Z","last_posted_at":"2026-07-23T12:49:48.514Z","bumped":true,"bumped_at":"2026-07-23T12:49:48.514Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"}],"tags_descriptions":{},"views":118,"like_count":3,"has_summary":false,"last_poster_username":"LarisaHaster","category_id":8,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Image 7’s pose is likely being ignored because the overlong, contradictory prompt dilutes priorities; compare the reference and result, simplify to pose and essential identities first, then add outfits, vehicle, and background through separate edits.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1334478,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56}]},{"fancy_title":"Using openai to solve Wordle","id":18529,"title":"Using openai to solve Wordle","slug":"using-openai-to-solve-wordle","posts_count":14,"reply_count":7,"highest_post_number":14,"image_url":null,"created_at":"2022-06-03T14:42:05.743Z","last_posted_at":"2026-07-22T17:49:21.444Z","bumped":true,"bumped_at":"2026-07-22T17:49:21.444Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[],"tags_descriptions":{},"views":2504,"like_count":9,"has_summary":false,"last_poster_username":"benlamm","category_id":8,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Structured, mandatory constraints were recommended to stop the model from proposing invalid Wordle guesses, while the original poster’s brief response suggested the commenter’s bot still needed improvement.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":11799,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":233,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":10407,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":1271867,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":4078,"primary_group_id":47,"flair_group_id":47}]},{"fancy_title":"Designing system prompts for culturally specific educational content using OpenAI models","id":1373400,"title":"Designing system prompts for culturally specific educational content using OpenAI models","slug":"designing-system-prompts-for-culturally-specific-educational-content-using-openai-models","posts_count":2,"reply_count":0,"highest_post_number":2,"image_url":null,"created_at":"2026-02-05T09:13:56.442Z","last_posted_at":"2026-07-20T23:33:44.320Z","bumped":true,"bumped_at":"2026-07-20T23:33:44.320Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":10,"name":"gpt-4","slug":"gpt-4"},{"id":33,"name":"api","slug":"api"},{"id":355,"name":"assistants-api","slug":"assistants-api"}],"tags_descriptions":{},"views":158,"like_count":0,"has_summary":false,"last_poster_username":"Alex_Kondor1","category_id":8,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"For culturally specific educational content, ground models with trusted facts instead of relying on recall, and instruct them to acknowledge regional or traditional variation when uncertain to reduce hallucinations while preserving beginner-friendly explanations.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1275854,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1298105,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Testing a Sharp-Tongued AI Persona — Looking for Prompt Tweaks","id":1369265,"title":"Testing a Sharp-Tongued AI Persona — Looking for Prompt Tweaks","slug":"testing-a-sharp-tongued-ai-persona-looking-for-prompt-tweaks","posts_count":193,"reply_count":138,"highest_post_number":196,"image_url":"https://us1.discourse-cdn.com/openai1/optimized/4X/e/9/0/e900415d74bb2e609108ba09f1470b6533cf8520_2_1024x532.jpeg","created_at":"2025-12-14T15:58:39.938Z","last_posted_at":"2026-07-17T16:17:59.536Z","bumped":true,"bumped_at":"2026-07-17T16:17:59.536Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"visibility_reason_id":4,"tags":[{"id":33,"name":"api","slug":"api"},{"id":52,"name":"playground","slug":"playground"},{"id":596,"name":"persona","slug":"persona"}],"tags_descriptions":{},"views":5698,"like_count":254,"has_summary":true,"last_poster_username":"LarisaHaster","category_id":8,"op_like_count":11,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A sharp, sarcastic AI persona prompt is being refined for consistent, controlled behavior across models, and the latest participant expressed support by following and starring the project, prompting the author’s thanks.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Frequent Poster","user_id":1009035,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":40833,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":699820,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":509645,"primary_group_id":null,"flair_group_id":56}]},{"fancy_title":"How should a “prompt engineer” prompt be updated for GPT-5.5?","id":1384583,"title":"How should a “prompt engineer” prompt be updated for GPT-5.5?","slug":"how-should-a-prompt-engineer-prompt-be-updated-for-gpt-5-5","posts_count":13,"reply_count":5,"highest_post_number":14,"image_url":null,"created_at":"2026-06-24T10:32:52.512Z","last_posted_at":"2026-06-30T18:40:34.421Z","bumped":true,"bumped_at":"2026-06-30T18:40:34.421Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"},{"id":496,"name":"gpt5","slug":"gpt5"}],"tags_descriptions":{},"views":2741,"like_count":22,"has_summary":false,"last_poster_username":"Baseballguru","category_id":8,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Participants suggest newer models benefit from outcome-first, natural-language guidance with clear goals, context, constraints, success criteria, selective questions, lightweight self-checks, and iterative testing, while many now let GPT-5.5 generate detailed prompts for Codex or other tools.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1160834,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1309291,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":792,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1305684,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Tasking ChatGPT with collecting product URLs from a CSV BOM","id":1380396,"title":"Tasking ChatGPT with collecting product URLs from a CSV BOM","slug":"tasking-chatgpt-with-collecting-product-urls-from-a-csv-bom","posts_count":9,"reply_count":4,"highest_post_number":9,"image_url":null,"created_at":"2026-05-06T13:28:13.214Z","last_posted_at":"2026-06-28T17:41:06.377Z","bumped":true,"bumped_at":"2026-06-28T17:41:06.377Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"}],"tags_descriptions":{},"views":274,"like_count":8,"has_summary":false,"last_poster_username":"aisystemsdev","category_id":8,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"ChatGPT alone is deemed unreliable for enriching BOM files with verified product and datasheet URLs; participants recommend a structured, auditable workflow or dedicated Python/Codex-based tool with batching, evidence, reconciliation, and persistent logic.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1307427,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Recent Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1324650,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Same prompt, different results","id":1355886,"title":"Same prompt, different results","slug":"same-prompt-different-results","posts_count":10,"reply_count":3,"highest_post_number":10,"image_url":null,"created_at":"2025-09-02T17:41:55.129Z","last_posted_at":"2026-06-23T12:14:42.806Z","bumped":true,"bumped_at":"2026-06-23T12:14:42.806Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":296,"name":"dalle3","slug":"dalle3"}],"tags_descriptions":{},"views":1311,"like_count":5,"has_summary":false,"last_poster_username":"ethanAI","category_id":8,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"ChatGPT’s web interface may refine or reinterpret image prompts before generation, while the API sends the raw prompt, so matching results requires explicitly specifying composition, lighting, style, camera angle, and other visual details.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1185511,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1216273,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":1009035,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1317878,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Image model stuck on the same style","id":1357399,"title":"Image model stuck on the same style","slug":"image-model-stuck-on-the-same-style","posts_count":27,"reply_count":17,"highest_post_number":27,"image_url":null,"created_at":"2025-09-10T16:13:38.075Z","last_posted_at":"2026-06-22T18:41:54.889Z","bumped":true,"bumped_at":"2026-06-22T18:41:54.889Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":42,"name":"image-generation","slug":"image-generation"}],"tags_descriptions":{},"views":3720,"like_count":22,"has_summary":false,"last_poster_username":"DysTopia","category_id":8,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A user reports image generations ignoring prompt changes and showing the wrong illuminated face side; a suggested fix is adding custom instructions so image outputs remain independent unless explicitly asked to reference prior images.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":478475,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":1009035,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":509645,"primary_group_id":null,"flair_group_id":56}]},{"fancy_title":"Reducing AI-sounding output in a mostly stateless multimodal commentary bot","id":1384232,"title":"Reducing AI-sounding output in a mostly stateless multimodal commentary bot","slug":"reducing-ai-sounding-output-in-a-mostly-stateless-multimodal-commentary-bot","posts_count":2,"reply_count":0,"highest_post_number":2,"image_url":null,"created_at":"2026-06-21T01:04:02.116Z","last_posted_at":"2026-06-22T15:38:32.288Z","bumped":true,"bumped_at":"2026-06-22T15:38:32.288Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":355,"name":"assistants-api","slug":"assistants-api"}],"tags_descriptions":{},"views":135,"like_count":2,"has_summary":false,"last_poster_username":"sergeliatko","category_id":8,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A commenter recommends reducing AI-like commentary by explicitly defining human style, building human and neutral sample pairs, fine-tuning separate neutralizer and humanizer models, then generating, neutralizing, and re-humanizing each response.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1322193,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":792,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Do AI Prompts Get Lost In Translation?","id":1379795,"title":"Do AI Prompts Get Lost In Translation?","slug":"do-ai-prompts-get-lost-in-translation","posts_count":5,"reply_count":0,"highest_post_number":5,"image_url":null,"created_at":"2026-04-26T03:46:49.670Z","last_posted_at":"2026-06-10T18:33:12.760Z","bumped":true,"bumped_at":"2026-06-10T18:33:12.760Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":true,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":176,"name":"prompt-engineering","slug":"prompt-engineering"}],"tags_descriptions":{},"views":407,"like_count":6,"has_summary":false,"last_poster_username":"keesan","category_id":8,"op_like_count":3,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Participants discuss how multilingual prompting may alter results, with recent comments emphasizing prompt sensitivity in conversation, uncertainty about model behavior and language-training quality, and concerns that long-context instructions are often ignored without stronger execution guardrails.","has_accepted_answer":true,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Accepted Answer","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Frequent Poster","user_id":834641,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1318263,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1310407,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"ERROR in PRISM , KINDLY HELP! urgent","id":1382862,"title":"ERROR in PRISM , KINDLY HELP! urgent","slug":"error-in-prism-kindly-help-urgent","posts_count":2,"reply_count":0,"highest_post_number":2,"image_url":null,"created_at":"2026-06-06T11:47:49.197Z","last_posted_at":"2026-06-06T12:55:59.453Z","bumped":true,"bumped_at":"2026-06-06T12:55:59.453Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":122,"name":"error","slug":"error"},{"id":1139,"name":"prism","slug":"prism"}],"tags_descriptions":{},"views":309,"like_count":3,"has_summary":false,"last_poster_username":"VeitB","category_id":8,"op_like_count":2,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A university assistant professor cannot access PRISM after a failed email verification/sign-up attempt, receiving a revoked OAuth token 401 error, and a community member has escalated the issue to the team for review.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1317192,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":72385,"primary_group_id":47,"flair_group_id":null}]},{"fancy_title":"DALLE3 and gpt-image-1 Prompt Tips and Tricks Thread","id":498040,"title":"DALLE3 and gpt-image-1 Prompt Tips and Tricks Thread","slug":"dalle3-and-gpt-image-1-prompt-tips-and-tricks-thread","posts_count":132,"reply_count":81,"highest_post_number":138,"image_url":null,"created_at":"2023-11-12T23:10:52.727Z","last_posted_at":"2026-05-29T01:12:40.834Z","bumped":true,"bumped_at":"2026-05-29T01:12:40.834Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":296,"name":"dalle3","slug":"dalle3"},{"id":314,"name":"dalle","slug":"dalle"},{"id":829,"name":"gallery","slug":"gallery"}],"tags_descriptions":{},"views":111140,"like_count":202,"has_summary":true,"last_poster_username":"_j","category_id":8,"op_like_count":38,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A previously difficult prompt challenge—creating a Yuan-Ti-like humanoid with two snake arms—is reported as solved in gpt-image-2 by simply asking clearly, reducing the need for special prompting tricks.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1079,"primary_group_id":47,"flair_group_id":47},{"extras":null,"description":"Frequent Poster","user_id":748766,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":617918,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Prompting Guide - GPT real time mini","id":1381758,"title":"Prompting Guide - GPT real time mini","slug":"prompting-guide-gpt-real-time-mini","posts_count":6,"reply_count":3,"highest_post_number":6,"image_url":null,"created_at":"2026-05-25T20:35:27.121Z","last_posted_at":"2026-05-27T09:59:20.387Z","bumped":true,"bumped_at":"2026-05-27T09:59:20.387Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":true,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":732,"name":"realtime","slug":"realtime"},{"id":1009,"name":"gpt-realtime","slug":"gpt-realtime"}],"tags_descriptions":{},"views":521,"like_count":10,"has_summary":false,"last_poster_username":"lnobrega","category_id":8,"op_like_count":2,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"OpenAI Support confirmed there is no dedicated gpt-realtime-mini prompting guide yet; users should adapt the newer realtime guide with shorter prompts, simpler tool instructions, and separate testing for interruptions and tool-calling, while guidance is treated as a feature request.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":1307570,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Recent Poster","user_id":1301342,"primary_group_id":55,"flair_group_id":55},{"extras":null,"description":"Recent Poster","user_id":269430,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Natural-language requests vs. long structured prompts: what actually improves performance?","id":1381299,"title":"Natural-language requests vs. long structured prompts: what actually improves performance?","slug":"natural-language-requests-vs-long-structured-prompts-what-actually-improves-performance","posts_count":12,"reply_count":11,"highest_post_number":20,"image_url":null,"created_at":"2026-05-19T12:47:36.055Z","last_posted_at":"2026-05-21T09:44:34.119Z","bumped":true,"bumped_at":"2026-05-21T09:44:34.119Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":true,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"}],"tags_descriptions":{},"views":1035,"like_count":23,"has_summary":false,"last_poster_username":"HanakoAbsed","category_id":8,"op_like_count":2,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Participants agreed that effective prompting depends on task, model, and workflow, with the poster concluding that iterative dialogue suits exploration while structured templates are better for execution, consistency, and constraints.","has_accepted_answer":true,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":1309252,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster, Accepted Answer","user_id":1214162,"primary_group_id":null,"flair_group_id":56},{"extras":null,"description":"Frequent Poster","user_id":40833,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":1302247,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":-1,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"A use-case example for data transfer between LLM chat threads and maintaining architectural continuity across systems","id":1381023,"title":"A use-case example for data transfer between LLM chat threads and maintaining architectural continuity across systems","slug":"a-use-case-example-for-data-transfer-between-llm-chat-threads-and-maintaining-architectural-continuity-across-systems","posts_count":1,"reply_count":0,"highest_post_number":1,"image_url":null,"created_at":"2026-05-15T19:06:35.040Z","last_posted_at":"2026-05-15T19:06:35.134Z","bumped":true,"bumped_at":"2026-05-15T19:06:35.134Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"},{"id":526,"name":"prompt-system","slug":"prompt-system"}],"tags_descriptions":{},"views":121,"like_count":0,"has_summary":false,"last_poster_username":"Balanced_Tiger","category_id":8,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Balanced_Tiger explains that LLM thread outputs must be shaped for their destination: broad audits aid human understanding, while structured extraction preserves validated, reusable context for continuity across layered collaboration systems.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest single","description":"Original Poster, Most Recent Poster","user_id":1293283,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Image generation: Limb placement for different walking steps","id":1380627,"title":"Image generation: Limb placement for different walking steps","slug":"image-generation-limb-placement-for-different-walking-steps","posts_count":4,"reply_count":1,"highest_post_number":4,"image_url":"https://us1.discourse-cdn.com/openai1/optimized/4X/5/5/2/552864d99d1f525ad0ce8ee49c94b930fe924833_2_1024x682.jpeg","created_at":"2026-05-10T10:48:45.568Z","last_posted_at":"2026-05-15T16:44:18.837Z","bumped":true,"bumped_at":"2026-05-15T16:44:18.837Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":true,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":42,"name":"image-generation","slug":"image-generation"}],"tags_descriptions":{},"views":231,"like_count":10,"has_summary":false,"last_poster_username":"Agentify.sh","category_id":8,"op_like_count":2,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"The user seeks reliable AI-generated platformer walk sprites with distinct limb poses, and responders suggest using community sprite-sheet workflows, stick-figure guides, or posed 3D mannequins before applying the desired style with an image model.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1308414,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":72385,"primary_group_id":47,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":1301342,"primary_group_id":55,"flair_group_id":55},{"extras":"latest","description":"Most Recent Poster","user_id":1309855,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Prompting vs Structure - A Boundary Test","id":1379094,"title":"Prompting vs Structure - A Boundary Test","slug":"prompting-vs-structure-a-boundary-test","posts_count":2,"reply_count":0,"highest_post_number":2,"image_url":"https://us1.discourse-cdn.com/openai1/original/4X/0/6/f/06f088e834f21e0fd418d40fde19312704819dde.jpeg","created_at":"2026-04-16T08:55:09.167Z","last_posted_at":"2026-05-05T11:00:14.217Z","bumped":true,"bumped_at":"2026-05-05T11:00:14.217Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":true,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"},{"id":42,"name":"image-generation","slug":"image-generation"},{"id":1133,"name":"gpt-image-2","slug":"gpt-image-2"}],"tags_descriptions":{},"views":315,"like_count":8,"has_summary":false,"last_poster_username":"EricGT","category_id":8,"op_like_count":5,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Tina_ChiKa’s image-model test argues prompts improve realistic appearance but not physical consistency, while EricGT suggests an “underdrawings” approach from HackerNews as a possible way to add the missing structural guidance.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":726999,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":1301342,"primary_group_id":55,"flair_group_id":55},{"extras":"latest","description":"Most Recent Poster","user_id":40833,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Prompt for automation solution image","id":1380179,"title":"Prompt for automation solution image","slug":"prompt-for-automation-solution-image","posts_count":11,"reply_count":6,"highest_post_number":11,"image_url":null,"created_at":"2026-05-02T03:35:32.542Z","last_posted_at":"2026-05-04T03:05:32.733Z","bumped":true,"bumped_at":"2026-05-04T03:05:32.733Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":true,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"chatgpt","slug":"chatgpt"}],"tags_descriptions":{},"views":297300,"like_count":8,"has_summary":false,"last_poster_username":"windysoliloquy","category_id":8,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"HarshEntrepre is refining prompts for an automation shelf robot image, with others advising more precise mechanical descriptors, single-module technical views, and explicit placement after a suggested prompt still produced an oversized robot and misplaced medicines.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1306030,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":940211,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":632960,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":1301342,"primary_group_id":55,"flair_group_id":55},{"extras":"latest","description":"Most Recent Poster","user_id":1009035,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Reverse engineering the Fameviso prompt (How they bypassing AI detectors?)","id":1380105,"title":"Reverse engineering the Fameviso prompt (How they bypassing AI detectors?)","slug":"reverse-engineering-the-fameviso-prompt-how-they-bypassing-ai-detectors","posts_count":3,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2026-04-30T12:44:12.714Z","last_posted_at":"2026-04-30T13:31:07.466Z","bumped":true,"bumped_at":"2026-04-30T13:31:07.466Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":355,"name":"assistants-api","slug":"assistants-api"}],"tags_descriptions":{},"views":316,"like_count":2,"has_summary":false,"last_poster_username":"LarisaHaster","category_id":8,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A new API user seeks more human-sounding casual social posts, while replies warn that undisclosed AI impersonation violates terms and suggest improving tone through clear developer instructions, input-output examples, and persona prompt resources.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1303096,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":228799,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":1214162,"primary_group_id":null,"flair_group_id":56}]},{"fancy_title":"DIEP Flap Surgery Recovery Prompts | TXDIEPFLAP","id":1379877,"title":"DIEP Flap Surgery Recovery Prompts | TXDIEPFLAP","slug":"diep-flap-surgery-recovery-prompts-txdiepflap","posts_count":2,"reply_count":0,"highest_post_number":4,"image_url":null,"created_at":"2026-04-27T13:34:33.654Z","last_posted_at":"2026-04-28T16:07:16.131Z","bumped":true,"bumped_at":"2026-04-28T16:07:16.131Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"visibility_reason_id":1,"tags":[],"tags_descriptions":{},"views":101,"like_count":1,"has_summary":false,"last_poster_username":"OpenAI_Support","category_id":8,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A user seeking consistent AI outputs for post-surgical recovery guidance was advised to use fixed response sections, explicit detail requirements, reference examples, repeated constraints, system instructions, and clear limits to general information rather than personalized medical advice.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":1304941,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":-3,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Recent Poster","user_id":-1,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":269430,"primary_group_id":null,"flair_group_id":null}]}]}}