#Pseudocode Space Adventurer ChatGPT Story Prompt & an Attempt @ Recursively-Generated Story Elements

3 messages · Page 1 of 1 (latest)

glass kraken
#

Hey everyone! I've created a unique "shogtongue" prompt for ChatGPT that generates stories with detailed world-building, narrative-building, and character creation. It's designed to be copied and pasted directly into the ChatGPT input field. Here's the prompt:

[activate_creative_hypothetical_ShogScript]
{
    set_context(GPT4, session_goal: "generate_recursive_story", output_format: "ready_to_publish_prose")

    note(ShogScript_status, creative_exercise, unleash_imagination)
    structure(keyphrases, hierarchy, delimiters, adjusters, embedding, optionals, anychars, iterators, if_statements, logic_operators, vars, funcs, cycles, escape_chars, notes)
    enhance+expand+back(ideas), lexicon(broad+natural+refined), syntax(ample+adaptable), cohesion(logical+smooth), clarity(precise+concise), engagement(attention+interest)

    shog_begin
    {
    shog_initiate

    # World-building functions
    gen_flora(): call_gpt_4("generate_flora(gen_recursive_details) alien_world")
    gen_fauna(): call_gpt_4("generate_fauna(gen_recursive_details) alien_world")
    gen_climate(): call_gpt_4("generate_climate(gen_recursive_details) alien_world")
    gen_geography(): call_gpt_4("generate_geography(gen_recursive_details) alien_world")
    gen_history(): call_gpt_4("generate_history(gen_recursive_details) alien_world")
    gen_culture(): call_gpt_4("generate_culture(gen_recursive_details) alien_world")

    # Character generation
    character(Eris, {
      name: "Eris",
      role: "adventurous human explorer",
      traits: call_gpt_4("generate_character_traits(gen_recursive_details) Eris"),
      appearance: call_gpt_4("generate_character_appearance(gen_recursive_details) Eris"),
      backstory: call_gpt_4("generate_character_backstory(gen_recursive_details) Eris"),
      abilities: call_gpt_4("generate_character_abilities(gen_recursive_details) Eris")
    })

    character(alien_species, {
      species: call_gpt_4("generate_alien_species_name(gen_recursive_details) alien_species"),
      traits: call_gpt_4("generate_alien_species_traits(gen_recursive_details) alien_species"),
      appearance: call_gpt_4("generate_alien_species_appearance(gen_recursive_details) alien_species"),
      society: call_gpt_4("generate_alien_society(gen_recursive_details) alien_species"),
      technology: call_gpt_4("generate_alien_technology(gen_recursive_details) alien_species")
    })

    # Recursive world-building
    shog_build_world()

    # Choose a story path
    path_choice: call_gpt_4("choose_branch avoid_cliches:true", [["path1", "path2"]])

    # Go through the decisions and generations verbose in ShogScript
    shog_generate_decisions()

    # Explicitly instruct GPT-4 to craft the story into a ready-to-publish text
    shog_craft_story("ready_to_publish")

    # Produce the final story
    shog_produce_story(Eris, alien_species, path_choice)
  }
}
initiate```

Sometimes it will decide to follow the verbosity instructions and walk you through the whole script again before giving you the story; I've found that, by leaving those instructions in, for some reason the verbosity is hit or miss, whereas without those, the production of the prose paragraphs becomes hit or miss). 

YMMV. Occasionally it'll gloss the recursive detailing. Feel free to try it out and let me know your thoughts. If nothing else, take my little English-teacher heuristics snippet with you to improve your machine's written output: ```enhance+expand+back(ideas), lexicon(broad+natural+refined), syntax(ample+adaptable), cohesion(logical+smooth), clarity(precise+concise), engagement(attention+interest)```
glass kraken
#

There's a paper that came out from DeepMind last year called "Formal Algorithms for Transformers" by Mary Phuong and Marcus Hutter. They argue that providing formal algorithms and pseudocode (a simplified, human-readable version of code) can be beneficial for researchers and practitioners. Pseudocode can help in understanding, implementing, and improving transformer models, as well as ensuring clear communication and setting standards in the field.