HERE IS THE ULTIMATE GUIDE FOR

[autumn ivy]

The last one is a list of more examples you can think about using!

... Although, Chat doesn't have Research papers...

[autumn ivy]

This was the start of how I refined the process in the very first image, with a few more steps, reached that outline of a general process.

[misty frost]

That's actually useful and very well organized.

Thanks

[sand panther]

Please tell me I can download the text for this?

[autumn ivy]

I will work on a .Txt version in a bit

[red stag]

This is great work and thanks for sharing. Have you noticed that one of the referenced books doesn't exist "Information Fatigue Syndrome" by Mark Levy.

[autumn ivy]

I don't see that one in my list?

[red stag]

Do I need to run this prompt before the prompt in the first image?

[autumn ivy]

that is inresting, but it didn't use that book in the pull.

step 1-5 is exactly what I ran.

In the very last image.

[red stag]

Do you need to run step 1 - 5 every single time before the main prompt (prompt in the first image)

I also wouldn't worry too much about the non-existent book. That would be too easy for us 🙂

[autumn ivy]

Sorry, I couldn't figure out how to save as a .txt on a Mac

[tardy girder]

Whats pretty interesting is if you ask this way it can provide quotes from its resources and cite them correctly

[autumn ivy]

That is because it was trained on this stuff, it just couldn't rewrite every word of it. The liks are also hit or miss sense it's old data and those are most subject to Bit Decay.

[manic thicket]

Thank you for the great work 🙏

[woven pebble]

Use picture to text converter. You'll need to convert one picture at a time but it's quite easy

aahh I don't know if what I just said was relevant so ignore if it's not

[runic helm]

@autumn ivy Incredible work, thank you. I can’t wait to try this for a grant writing project I’m working on.

I love the use of ‘extrapolate on this…” what a great way to modify its output.

[autumn ivy]

This tool is a sortof categorical exploration tool. The idea is to use the Lexicon if you have a new idea without words, to give it shape. You can use the Bufdha bot to form that new idea for the Lexicon. Then the Story Writer Obie can be used to construct a story, but formed around your goals and tasks and not a real story. It is more of an outline shaping tool to bring some structure to your idea. Imagination and categories.

[runic helm]

Yes, your last sentence is exactly how I see it as well. The way I explained ChatGPT to a friend the other day was like this “It can articulate the complexities in your head with a little coaching. You have to be strategically psychological about your input in order to shape the desired output.”

I’ll have to read more into Bufdha & Obie. Since I started I’ve been testing how to use natural language to get the AI to learn my tones, personal info/history and goals to have it create on-brand copy, scripts, biographies and more in my voice.

Can you point to some more resources or threads where I can learn more?

[autumn ivy]

Unfortunately this is a sort of discovery process that doesn't have a lot of words to describe it. Even AI is nonsense like "the internet" it is just massive and hard to talk about because of this. This learning I have done is via exploring Chat. Sometimes using references that it knows, but mostly just asking it to list simple categories and keywords or examples. I can then trim out what I don't want, and rewrite without those keywords or categories. The goal is to simply bring shape your ideas with it. It is not psychological... unless you mean by an examination of your own psychology via this process of working out ideas into fully formed thoughts.

The Bots just sort of give Chat the exact structure I want for that final result really.

[dim condor]

@autumn ivy I'm using your methods as a way to learn how I can help write content for real estate professionals who have Investor clients and I've run into a snag. When I get to Step 4 and Step 5, I get pushback from ChatGPT. I've asked it to research articles and find best practices for success in real estate investing. Any suggestions on how I can move forward?

[autumn ivy]

You may want to practice the Natural Language method first before using Prime Directives to control format with small token space memory. Do some research on Zero Shot, One Shot, and Multi Shot prompting, and the general format of a Natural Language Program, or a "prompt."

When using this method, it has been nerfed by updates, chat won't let you put the input into more than one nested container any more. Nesting means like ("[this is nested 3 times]"), and a container is [] or "" or () or {} used to define context uniquely; for example I use [] to define the Prime Directive variable conditions to expressly define how I want a session to handle my syntax, but I tend to use () or "" to define what is Context to my main directive request for chat to preform. Instead do directive variables like [format: list of 5 in simple summary format.] [context: tips for investing in real estate.]. You also have to be simple about your directive verb choices and context samples by using sound logic for higher complexity goals; each input should contain a SINGLE context base and a FORMAT output, OR you can do this type of Comparison Operation IF there are correlations in the datasets in the format you are requesting (like, compare and contrast solution A to solution B by including examples for each and the strengths and weaknesses of each.) The best results I have found so far is figuring out how to apply the subject of, say a book on this subject of building a profitable portfolio using some algorithmic model (for example, complex problem solving using TRIZ method) and applying this analysis format to generate more relevant output.