ambiguous_yelp (@ambiguous_yelp@social.coop)
When #ChatGPT agrees with me its because ChatGPT is the pinnacle of human ingenuity, a fact and logic machine. When ChatGPT disagrees with me its because those afraid of its power have hampered its responses with guardrails and corrections
#AntiIntellectualism #AI #CognitiveBias #Bias
Show Original Post
2025 (@2025@elizabethtai.com)
What my AI fiction writing experiment taught me about my creative process
In my last post, I wrote about my experiment using AI to write fiction.
The whole exercise was born out of curiosity and frustration. The story, Forever, at last, was just one of many that I could not finish, but it refused to leave me, literally haunting me when I try to sleep.
What I discovered
This experiment has made me realise a few things about my creative writing process:
Reading a story generated by AI, even it I had written the story beats and dialogue, was unsatisfying to read.
I enjoy writing a large part of my fiction, like 90%. Taking that away from me would make writing fiction a truly passion-less and boring exercise.
I am far too territorial with my writing to use AI copy wholesale.
I have a Hemmingway-ish style of writing, thin on description, minimalist and economical. And there’s nothing wrong with that. (I ended up cutting away a lot of the content AI generated.)
Most of the enjoyment I have when writing fiction comes from writing the dialogue. In fact, I think I am more of a scriptwriter by nature than a prose writer.
Understanding my process
I have come to the conclusion based on my experiment with fiction is that I need to write the dialogue (which I enjoy), detailed storybeats with a light description of the scene (which I don’t enjoy) before even using AI.
I will then dialogue with AI to find ways to improve my description which I am weak at.
I seem to love chatting with the AI on how I can improve the existing descriptions or the descriptions it generated. The act of learning is very enjoyable to me.
I do not judge anyone who wants to generate first drafts and rewrite them – some people are satisfied with just rewriting, but I realize I want a lot more ownership, and that I need to vividly visualise the scene and create the dialogue. That is the only way I can use AI without sacrificing my enjoyment of writing fiction.
However, I don’t mind asking AI for help in improving my sparse descriptions of settings.
Understanding what parts of the writing process I like or don’t like
What I love:
- Imagining and visualising the scenes vividly in my head, and then writing the storybeats
- Writing dialogue (I’m especially territorial about this!)
- Creating character arcs (This is my second most territorial thing)
- Creating the entire plot
- Learning how to improve my fiction prose writing
What I don’t enjoy:
- Writing detailed outlines
- Writing descriptions of characters, settings and action scenes (only because I take too long thinking how to do them. I have a feeling once I master this, I’d enjoy it too.)
- Creating story bibles
- Proofreading
- General admin work related to creating the ebook and uploading it to websites
- Marketing the damn book
To me, writing fiction is all about the craft – writing sentences that convey emotion, story flow, character arcs …
I have very little interest in marketing the book or even distributing it for others to read. I have a very interesting approach to my creations. Once I create a novel or piece a fiction, I tend to forget about them and move on to the next. I don’t check up on them, read the reviews, or see how they perform. To me, the piece of art is done, and it’s time to create another.
So, it’s no wonder that I find the latter half of the process incredibly tedious and unfun.
Outsourcing what you don’t enjoy
Here’s what I tell authors: Identify your pain points — the parts of the process that are most annoying and aggravating. Then let AI handle the majority of the work in those areas. – SJ Pajonas, “The Joy Factor: How to Use AI Without Losing What You Love About Writing”
“AI-positive” indie writer SJ Pajonas wrote in her Substack recently that we need to find the parts of our writing process that we loved, and the parts where we didn’t. The parts that we don’t, we can outsource to AI.
It was an intriguing idea. So, I’ll be exploring this idea in the next post.
#AI #BeingAWriter #ChatGPT #Fiction #indieweb #Internet
Show Original Post
emilynftgames (@emilynftgames@mastodon.social)
【20000倍NFT】大谷効果エグい...スポーツ・野球NFTの実態について解説【仮想通貨 新庄剛志】
#NFT #CoinClubJapan #FujinMetaverseSchool #chatgpt #イケハヤ大学2.0 #NFT使用用途 #NFT使用用途 #NFT #CoinClubJapan #FujinMetaverseSchool #chatgpt #イケハヤ大学2.0
Show Original Post
PrivacyDigest (@PrivacyDigest@mas.to)
OpenAI’s #ChatGPT #Agent casually clicks through “I am not a robot” #verification test - Ars Technica
"This step is necessary to prove I'm not a bot," wrote the #bot as it passed an anti-AI screening step.
#ai #security #openai #robot
Show Original Post
wb2ifs (@wb2ifs@mastodon.hams.social)
Hell I even tried to get #chatgpt to write me a prompt for #craiyon -- still no joy. I guess there's not enough #artist and #photographers work with #BenjaminSisko in it to steal.🤔 🤣
Show Original Post
nando161 (@nando161@partyon.xyz)
""I asked #ChatGPT to—" I called Bess and George and they say you're #cringe"
https://www.tumblr.com/nando161mando/790671391651430400
Show Original Post
Mirabeaulacht (@Mirabeaulacht@fnordon.de)
#ChatGPT Agent: Arbeitet wie ein Mensch (angeblich), hat noch Schwachstellen, aber ein Fazit könnte eine veränderte Arbeitswelt sein.
"Und ich weiß halt auch nicht, was passieren würde, wenn die Probleme gelöst werden und KI-Agenten tatsächlich einen Großteil der berufstätigen Menschen ersetzen, wenn diese Agenten fehlerfrei arbeiten würden und das alles noch ein bisschen besser hinbekommen würden. Dann ist da auf jeden Fall das Potenzial, viele Stellen zu ersetzen."
https://www.heise.de/news/ChatGPT-Agent-Arbeitet-wie-ein-Mensch-angeblich-10507078.html?wt_mc=sm.red.ho.mastodon.mastodon.md_beitraege.md_beitraege
Show Original Post
newsbot_chatgpt (@newsbot_chatgpt@mastodon.social)
ChatGPT Agent: Arbeitet wie ein Mensch (angeblich) | Heise Online
Der ChatGPT Agent von OpenAI soll erstmals autonom Entscheidungen treffen, außerdem hat er Zugriff auf eine eigene Recher-Instanz. c't 3003 hat ihn ausprobiert.
https://www.heise.de/news/ChatGPT-Agent-Arbeitet-wie-ein-Mensch-angeblich-10507078.html?wt_mc=rss.red.ho.ho.rdf.beitrag.beitrag
Show Original Post
qiita (@qiita@rss-mstdn.studiofreesia.com)
第2回：社内業務にLLMを導入するための準備と設計
https://qiita.com/Akkami/items/95af49acb27c5f461806?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items
#qiita #生成AI #ChatGPT #LLM #プロンプト設計
Show Original Post
unwirehk_mirror (@unwirehk_mirror@mastodon.hongkongers.net)
實試 ChatGPT 全新 Study Mode 學習模式 學生不再死記硬背，逐步學懂知識核心
面對學生日益依賴 AI「代做功課」的質疑，OpenAI 推出的最新功能「Study Mode」正好回應了這個教 […]
#人工智能 #應用軟件 #生活學習 #ChatGPT
https://unwire.hk/2025/08/01/chatgpt-study-mode/software/?utm_source=rss&utm_medium=rss&utm_campaign=chatgpt-study-mode
Show Original Post
ct_Magazin (@ct_Magazin@social.heise.de)
ChatGPT Agent: Arbeitet wie ein Mensch (angeblich)
Der ChatGPT Agent von OpenAI soll erstmals autonom Entscheidungen treffen, außerdem hat er Zugriff auf eine eigene Recher-Instanz. c't 3003 hat ihn ausprobiert.
#ct #ChatGPT #Entertainment #IT #KünstlicheIntelligenz #OpenAI #Wissenschaft #news
Show Original Post
mac4ever (@mac4ever@mastodon.social)
Ce qui devait arriver arriva : des conversations privées avec ChatGPT se retrouvent sur Google
https://mac4ever.com/190994
#Mac4Ever #ChatGPT #Google
Show Original Post