So recently, I introduced paid Subscriptions on my Substack (more options will come to access my products) and sent out a newsletter explaining why. Here, we will focus on the Triad of writing SensAI’s. A play on them being your writing, Sensei.
Straight after that, I worked on my SensAI GPTs and upgraded them to V 2.0’s. I’ve enhanced security to protect my GPTs and your data.
Any stories or story concepts you enter into my GPTs will not be stored or used in any way to contribute to any AI training. As I will demonstrate in the video, as you are logged in, it won’t even release or remember any information about the story you just finished using in any of the GPTs. It has no access to it to be able to give it to you. If you insist or trigger any other security protocols. Well, you will see what will happen.
As a writer myself, I take this very seriously. So much so that you can’t use these GPTs to write a story, rewrite your story or even give you a story concept. These GPTs are specifically crafted for a set purpose. I have spent close to a year working on these. Refining the process. Making them as accurate and realistic in the output as I could. The last bit was a challenge that I will explain.
How did this start? I read a medium article by a writer talking about the current review process for writers' works. Specifically on Amazon. We all have experienced or had a friend who has experienced unfair and ridiculous reviews. Worse, the troll reviews by people seeking to damage a writer’s launch date. They are so obvious, I’ve called on Amazon a few times to do something. Falls on deaf ears, of course. A blank, no name, no profile pic user, reviews immediately on launch, and a 1-star review with no context. It couldn’t be more obvious, and almost immediately, the algorithm has moved on from your launch, and sales are completely tanked. It’s all from some petty, jealous, hateful troll.
Of course, some writers feel they can go after legit reviewers who leave a 3-star review with context. Even a 1-star review by a reader, with context, is valid. Not everyone will like your work. Some will hate it. They have a right to review it.
Then there are the reviews like. The book was delivered late. 1-star. Or the delivery driver left the parcel in the rain, and the book was ruined. 1-star. These reviews are ridiculous and made by people who don’t understand the process, who is responsible for what and who the review affects. Things out of the author's control should not be part of the review of the book.
Then, some reviewers always give the same scores for every review they do. 3-star, or 5-star. No context. Then there’s the emotional factor and unpredictability of humans. Great writing gets readers thinking, stirs emotions, and pushes boundaries. Some readers will hate you, and some will love you. It’s the nature of the beast.
This writer was lamenting the unfair aspects of the review process and was sure there must be a better way, but couldn’t think of what that could be. That’s where I said,
”Hold my beer.”
I got to thinking and formulating my idea. Now, I immediately dismissed AI, replacing the review process. That’s just flipping the current issues with other issues. But what if AI reviews could offer a counterbalance to the review process? Giving readers the option of looking at human-based reviews and maybe thinking, “Oh, these reviews aren’t that great.” Unaware that maybe there’d been an attack by a group against the author. But they look at the review AI has given. Specifically, StorySensAI. Where they see the AI praise the technical aspects, the storytelling, and other 20 points, StorySensAI reviews from my instructions.
The challenge I had was AI doesn’t like to be overly critical of human users. So, for a while, I was getting frustrated with AI as it reviewed all my short stories and gave me scores of 4 to 5 out of 5 for all my writing, offering some critiques to improve. But hold up, if I am getting 5/5. How are you saying I can improve? I can’t. So you are lying to me. Sugarcoating the feedback in an attempt not to hurt my feelings?
Besides, I know my stories from thirty years ago are not worthy of 4 out of 5 stars. At best, a 3 out of 5. It took a lot of work to customise prompts and instructions. Testing and retesting. Finally, I got it, consistently too. The moment when I got the review, I was like. Hey, that score stings a little. There it is. Reality.
Now, StorySensAI can offer suggestions as to how I can improve my writing. It tells me what I’ve done well, StorySensAI can offer suggestions as well as what I could do better. Also tells me why it’s making those suggestions.
I was still having regular issues, though, of StorySensAI going off the rails at times and completely missing the entire plot of the story. Or making stuff up. That’s where SummarySensAI comes in. It does a 300-word summary of the story. You provide that to StorySensAI at the start, which enhances the review process, keeping StorySensAI from glitching out. Keeps StorySensAI focused on the story review at hand. So to speak.
SummarySensAI is also useful for summarising your story for any marketing you want to do for your story.
Finally, we have ConceptSensAI. This is a stand-alone GPT that will take a concept for a story you might have and will provide feedback on various aspects. Originality, target audience, story length, plus more.
I will now demonstrate in the following video how they work and the feedback you get. This is also just the beginning. Already developing genre-specific, as well as expanding the capabilities beyond short stories. Which StorySensAI can already do. It is suggested that you post your story in 5000-word chunks when prompted. You can probably get away with up to maybe 7000 word chunks. For the total story length, I’ve pushed StorySensAI with over 15000 words. Entering into Novella criteria.
Currently the beta launch is only available to my paid substack susbscribers. https://jamesdarmstrong.substack.com/
Website coming soon.
Cheers
JD
