Would anyone want to experience a Long Form Q&A with the scrivener manual?

We wouldn’t have made it through over a decade of NiaDs if we stopped the nonsense!


Many days, nonsense is my raison d’être. If not for engaging in nonsense, creating chaos, and committing shenanigans, why else would I leave my warm, comfy bed?


Stephen King has clearly been hypnotising me for years, as I keep buying and reading his damn books. Who do I sue?

1 Like


As funny as always.


If you read the third word from page 3, the 6th from page 6, 9th from page 9 and so on, there is a cake recipe hidden in there…

But seriously, don’t you think that maybe you just love his stories?


Don’t you understand?
It has nothing to do with manipulating audiences to buy anything.
It is about encouraging a reader to continue reading the material they bought/obtained.
And in addition, encouraging the author to focus on the creation of the material when their attention might drift or develop a writer’s block.

What is wrong with you?

This is becoming really weird.

I like it.


If readers are being “hypnotically compelled” to keep reading and keep buying, how is that not manipulation?


There must be something to this manipulation/encouragement trick. How else to explain why I find myself compelled to keep returning to this thread? :man_shrugging:


Oh, a great many things.
(And I love every single one of them).


So every great writer is a minipulator in your opinion?
Well, whatever, I guess.
But they are not subliminally advertising as was originally accused.

Anyway, I haven’t heard back from anyone L&L as to whether they want to have me setup a demo to hold a conversation with the scrivener handbook,

Absolutely not.

You’re the one making the claim that hypnosis provides some level of compulsion beyond what merely great writing can achieve.


I know I can make people not like my next book.
The other way around is a tiny bit more complex.

(I am not Stephen King under an alias, btw. Saying just in case. :wink: )


I claim no such thing.
You really refuse to comprehend.

Great writing IS hypnosis.
(not all hypnosis is great writing)

Buy My Next Book


(I don’t quite master great writing yet. I have to resort to more brutal methods.)


Aha! I get what you’re trying to say now!

You don’t mean literally hypnotise readers, you mean to use techniques learnt from advanced story theory, psychology, heuristics and linguistics to improve the writing quality in such a way that it appeals more directly to a reader’s base desires.

I think you’re putting a lot of faith into the magical properties of AI. After all, even our own brains — the most sophisticated computers ever made — can’t do it well. I mean, our brains do a great job of rendering a realistic model of the world when responding to actual live data (ie seeing it!), but when left to generate it from scratch (eg, in dreams) our created reality is full of flaws (such as broken physics), inconsistencies (text that changes every time you read it) and absurdities (why is that giant bird talking to me in Spanish? ¡Ni siquiera hablo español!).

Visual hallucinations follow similar issues, and I’ve seen no reports from schizophrenics that the voices they hear tell them nice and helpful things (“GET OUT!!! The garden is lovely this time of year”).

…and in the written sense, creativity is not AI’s strong point. What we call AI is really just machine learning, which is derivative by definition. Hell, ask ChatGPT to tell you a joke and it will tell you one of the same 25 jokes in a loop. The kind of great writing you’re talking about rarely falls into that category.

Anyway, while I’m not an LL employee, my advice to you if you’re hoping to pitch the idea to corporate owners and the app’s technical design authority is to be ready to answer to these questions:

  • What would integrating an AI model directly into Scrivener add to the user experience over letting users use whatever external, dedicated AI model they choose?
  • What would LL’s advantage be to enable them to get an AI model built better/faster/cheaper than dedicated AI development companies?
  • How much would developing, integrating and maintaining this functionality at the leading edge of the market divert development efforts from core functionality?
  • Does this functionality work with the current design principles of the app, and therefore how much would need to be adjusted or redesigned to suit what is currently an edge-case function offer?
  • What would this do to app pricing, and how much of a market distorting impact would this price adjustment have?

Oh, yeah, and…

  • Does our customer base actually want this?

Now, I think my role here was to provide the nonsense, so to get us back on track… here is a picture of me with a horse:


I will let ChatGPT4 answer these points with no editing from me.

Integrating an AI model directly into Scrivener would enhance the user experience in several ways:

  1. Seamless Integration: By having the AI model directly in Scrivener, users would enjoy a seamless experience. They wouldn’t need to switch between different apps or platforms to leverage AI capabilities. This integration would simplify the writing process, making it faster and more efficient.

  2. Contextual Understanding: An integrated AI can be designed to have a deeper understanding of the context within a Scrivener project, which an external AI might lack. This can enable more accurate suggestions, content generation, and error detection.

  3. Optimization: The AI model can be optimized for writing and editing tasks that are specific to Scrivener, providing a more tailored and effective tool for users.

As for Literature and Latte’s (LL) advantage, the key lies in their understanding of their user base and the specific use cases that Scrivener addresses:

  1. Domain Expertise: LL has a deep understanding of writers’ needs and how they use Scrivener. This expertise can guide the development of an AI model that is highly relevant and beneficial for their users.

  2. User Data: If users consent, LL could leverage anonymized data from Scrivener usage to train the AI model, potentially resulting in a better and more tailored tool than what a generic AI company could offer.

  3. Partnerships: LL could potentially partner with an existing AI company to speed up development, leverage existing AI technologies, and reduce costs.

In terms of development efforts, it’s true that AI integration could divert resources from other areas. However, the potential benefits could outweigh the costs. The key would be to strike a balance, potentially by forming a dedicated AI team or partnering with an AI company, thus allowing the core team to continue focusing on Scrivener’s core functionality.

The integration of AI functionality should ideally align with Scrivener’s current design principles. There may be some adjustments or redesigns needed, but these should be minimal if the AI integration is well-planned. The aim should be to enhance the existing functions rather than reinvent the entire system.

Finally, regarding pricing, the addition of AI functionality could justify a price increase given the added value it provides. However, this would need to be carefully considered to avoid potentially alienating existing users. A tiered pricing model could be a solution, where users can choose to pay extra for the AI-enhanced version while still offering the base version at the current price. The market distorting impact could be limited if the new pricing is competitive and reflects the added value the AI brings.

So…? It is not about the manual.


There is no single “it”

There is a discussion track about adding AI to the manual (which is working very well by the way) and there is a separate discussion in response to some issues someone raised.

They are two separate discussions albeit with some overlaps.

Never ask a barber if you need a haircut, and never ask an AI if AIs are useful. ChatGPT’s ability to write persuasive marketing material is well-documented, as is the tenuous connection between that material and objective reality.

In particular, Scrivener does not currently collect any data from users. Not the content of their projects, not their usage of the application, nothing. People write all kinds of things in Scrivener, including material that is potentially illegal in a variety of places. So any proposal that relies on analyzing user data anywhere except on the user’s own machine is an immediate non-starter.