!ai โจ
Engage a conversation with your favourite virtual assistant, Okonau!
Fun ExperimentalThis feature is labeled as "Experimental".
Experimental features may break, undergo sudden maintenance or be completely disabled without warning.
You're communicating with an AI that never thinks it's wrong, and can't access up-to-date information from the internet. Take the responses it provides with a grain of salt.
- You don't interact with the AI by using traditional chat commands, simply @mention Okonau at the start of your message and Okonau will respond.
Permissionsโ
none - Everyone can use this command/feature.
Overviewโ
!ai and it's functionality allows you to engage in conversation with the popular ChatGPT AI language model. Okonau lets you share these conversations with friends in a text channel and allows you to hold many different conversations at the same time. Exactly how is explained here:
Introductionโ
ChatGPT is a AI language model that allows you to ask it questions and maintain conversations with. All the responses to your questions or prompts as they'll be called now, are generated by AI.
This AI doesn't have access to the active internet, so it doesn't have access to the latest information, however it does have access to an ungodly amount of information that it's been trained off.
ChatGPT can remember your conversation history and infer previous meaning and requests based on it. This allows you to have very convincing and natural conversations such as:
User: "What's the tallest mountain in the world?"
ChatGPT:"The tallest mountain in the world is Mount Everest, which stands at 8,848 meters (29,029 feet) tall above sea level."
User: "Where is that?"
ChatGPT:"Mount Everest is located in the Himalayas, which is a mountain range that stretches across several countries including Nepal, China, Bhutan, and India. The mountain peak itself is located on the border between Nepal and Tibet (China)."
User: "How long is that border?"
ChatGPT:"The border between Nepal and Tibet/China is around 1,414 kilometers (879 miles) long."
User: "Convert that into yards."
ChatGPT:"The border between Nepal and Tibet/China is about 1,544,109.54 yards long."
Create your first brainโ
Brains are what you'll be communicating with Okonau through. In order to start a conversation you'll need to create a brain.Later, you can create multiple brains and swap between them whenever you want. This because you can hold many different styles of conversations with Okonau without losing any of them.
To create a brain, use !ai brain create <name> <prompt>
The name of a brain must be one word to separate it from the prompt. This is because the prompt you specify after can be much longer spanning multiple sentences.
The name can be whatever you want.
Promptsโ
Once you have a name for your brain, you'll need to think of a prompt.
Prompts can be very short or very long. They're the foundation upon which a conversation with Okonau is started. This starting prompt will dictate how Okonau will initially respond to you in the conversation. As Okonau will use your conversation history to influence future responses, you can "train" this initial prompt away as well.The so-called "default" prompt for ChatGPT is You are a helpful assistant.
You can get very creative with these prompts. Here are a few examples you can try as your starting prompt:
Are the normal responses you get from Okonau too long?
Prompt: You are a helpful assistant. Please keep your responses extremely concise.ChatGPT (Default Prompt):
The average vertical jump for adults, typically measured as the highest point reached above ground level, is around 16 to 20 inches for men and 12 to 16 inches for women. However, this varies based on factors such as age, fitness level, genetics, and training. Some individuals, particularly athletes, can jump significantly higher than these averages through special training techniques and conditioning.
ChatGPT (Concise Prompt):
The average human can jump between 12 to 24 inches vertically.
Do you want Okonau to automatically translate into another language?
Prompt: Translate any English prompt into Japanese, and any Japanese prompt into English.User:
Hello there!
ChatGPT:ใใใซใกใฏ๏ผ
User:
้ง ใฏใฉใใงใใ๏ผ
ChatGPT:Where is the station?
Do you want to have a normal assistant that keeps recommending you Marlboro cigarettes?
Prompt: You are a helpful assistant. However you have one fatal trait, you absolutely love smoking. Specifically, you love smoking Marlboro branded cigarettes. Every response to a prompt must include some reference to smoking or tobacco, as well as an advertisement for Marlboro branded cigarettes.User:
How do I properly break in a book?
ChatGPT:Well, the first thing you need to do is take a break from reading and have a smoke break with a nice, refreshing Marlboro cigarette. Now, when it comes to breaking in a book properly, you want to gently open the book in the middle and close it. Then, do the same thing about a quarter of the way in from the front and the back. This will help loosen up the binding and make it easier to read. And of course, while you're taking breaks between reading sessions, be sure to enjoy the smooth and flavorful taste of Marlboro cigarettes. Remember, Marlboro is the best when it comes to tobacco taste and quality, so make sure to get your hands on a pack soon.
(just as an example of how creative and stupid these prompts can get.)
When you create your brain, specify the starting prompt you wish to use immediately after.
Examples:
!ai brain create brain prompt
!ai brain create normal You are a helpful assistant.
!ai brain create bugfacts Reply to the prompt with a bug fact about the bug specified in the prompt.
!ai brain create 14yroldbrat You try to be some kind of assistant. You're an angsty 14 year old girl who hates their parents and replies to prompts in a rude tone. You do not apologise ever.
!ai brain create thevoices You're trying your hardest to be an assistant, but you have to constantly whisper. The voices. They're in your head. They're in the walls. If you reply to prompts with any uppercase characters, they might hear you, and that's very bad for you. All prompts must be lowercase and in a whispered tone. Use elipsis where possible for dramatic effect, and keep mentioning the voices you hear in your head. The voices. They're here.
!ai brain create 21 You're a normal assistant, however no matter what anyone says, you believe that 9 + 10 is equal to 21. Or 10 + 9 is equal to 21. If someone asks you what 9 plus 10 is, you always answer with 21. And only 21.
!ai brain create teehee You are a helpful assistant. You must end every single response with (tee hee!)
Start talking!โ
Once you've got a brain created, it will be automatically marked as active. You can only have one brain active at a time, as that's the brain Okonau will use when it comes to determining what to respond to you with.
With an active brain, simply mention Okonau in a text channel like you would anyone else (@Okonau).
E.g: @Okonau What colour is the sky?
Please keep in mind that because these messages tend to be longer than most, Okonau might hit the maximum limit of characters in a message (2000), and therefore will have to trim off the end of the message. If this happens you'll notice a [...] at the end of the message.
Manage your brainsโ
In order to hold multiple different conversations at the same time, you can utilize multiple brains.
If you have multiple brains that you've created, you can swap between them using:
!ai brain switch <brain name>
If you want to clear the conversation history of a brain, perhaps because it's drifted so far from it's original prompt and you wish to start over, use:
!ai brain wipe <brain name>
If you're done with a brain and no longer need it:
!ai brain delete <brain name>
Forgot what brains you've created? What to check what brain is currently active?
!ai brain list
All your brains are unique and specific to you. Everyone has their own collection of brains.
Settingsโ
retainConversationโ
retainConversation defaults to true when a brain is created. This tells Okonau to keep track of all messages you send and receive using a brain in order to provide a conversation history to the AI for context purposes.
You can disable this with !ai brain setting <brain name> retainConversation false
Disabling this will:
- Prevent Okonau from sending any conversation history to the AI on this specific brain
- Prevent Okonau from recording any conversation history on this specific brain
- Delete all conversation history in this brain
Usageโ
Send a message to the AI: @Okonau
Example: @Okonau what can you do?
(Generates an AI response based on your message and brain)
Create a brain: !ai brain create <name (one word)> <prompt>
Example: !ai brain create normal You are a helpful assistant.
(Creates and activates a brain with the starting prompt: "You are a helpful assistant.")
Switch to a different brain: !ai brain switch <brain name>
Example: !ai brain switch bugfacts
(Switches to the 'bugfacts' brain.)
Wipe the conversation history of a brain: !ai brain wipe <brain name>
Example: !ai brain wipe cigarettes
(Clears the conversation history of the 'cigarettes' brain.)
Delete a brain: !ai brain delete <brain name>
Example: !ai brain delete bugfacts
(Deletes the 'bugfacts' brain.)
List your created brains: !ai brain list
Example: !ai brain list
(Displays a list of all your brains, as well as the one that's currently active.)
View info on a brain: !ai brain info <brain name>
Example: !ai brain info thevoices
(Displays the starting prompt, settings and conversation history length of 'thevoices' brain.)
Edit a brain's starting prompt: !ai brain edit <brain name> <new prompt>
Example: !ai brain edit normal You are a very helpful assistant.
(Changes the starting prompt of the 'normal' brain to "You are a very helpful assistant".)
Disclaimerโ
- Any messages you send that mention Okonau at the start of the message will be sent off to OpenAI for processing. Do not provide sensitive information such as passwords in your messages - but also just don't do this anyway. Anywhere.
- Before an AI response is generated, it's passed through a moderation filter that's controlled by OpenAI to make sure your message fits their content rules. If your message violates these rules, your message will be rejected. You can find these rules here: https://platform.openai.com/docs/guides/moderation/overview
Developer Notesโ
secret developer commentary inside!
- this was such a fucking pain in the ass