This media is not supported in your browser
VIEW IN TELEGRAM
(4) GENERATE FORMULAS
You can describe a formula and it will generate it for you. From simple stuff thatβs easier to describe than type out, to harder things incorporating regex, formulas based on contextual info, and sometimes stuff using functions you didnβt even know existed.
You can describe a formula and it will generate it for you. From simple stuff thatβs easier to describe than type out, to harder things incorporating regex, formulas based on contextual info, and sometimes stuff using functions you didnβt even know existed.
This media is not supported in your browser
VIEW IN TELEGRAM
(5) FILTERING
Selectively keep/remove rows using natural language prompts. Remove duplicates, only keep submissions with proper email addresses, remove blank rows, etc.
Hereβs an example of cleaning up form submissions and then filtering the results
Selectively keep/remove rows using natural language prompts. Remove duplicates, only keep submissions with proper email addresses, remove blank rows, etc.
Hereβs an example of cleaning up form submissions and then filtering the results
This media is not supported in your browser
VIEW IN TELEGRAM
(6) =AI() and =INFER()
Directly incorporate GPT-3 (and soon GPT-4!) prompts into your spreadsheets using =AI(), or teach it what to do using =infer().
Examples:
- Extracting information from text
- Summarizing text
- Formatting phone numbers
- Categorizing customer feedback
Directly incorporate GPT-3 (and soon GPT-4!) prompts into your spreadsheets using =AI(), or teach it what to do using =infer().
Examples:
- Extracting information from text
- Summarizing text
- Formatting phone numbers
- Categorizing customer feedback
Leaked prompts collection:
https://github.com/sw-yx/ai-notes/blob/main/Resources/Notion%20AI%20Prompts.md
https://github.com/sw-yx/ai-notes/blob/main/Resources/Notion%20AI%20Prompts.md
βUnstable Diffusion Commits to Fighting Back Against the Anti-AI Mob
It seems that the anti-AI crowd filled with an angry fervor. They're not content with just removing Unstable Diffusions Kickstarter, but they want to take down ALL AI art.
The GoFundMe to lobby against AI art blatantly peddles the lie the art generators are just advanced photo collage machines and has raised over $150,000 to take this to DC and lobby tech illiterate politicians and judges to make them illegal.
Here is the official response we made on discord. I hope to see us all gather to fight for our right. β
We have some urgent news to share with you. It seems that the anti-AI crowd is trying to silence us and stamp out our community by sending false reports to Kickstarter, Patreon, and Discord. They've even started a GoFundMe campaign with over $150,000 raised with the goal of lobbying governments to make AI art illegal.
Unfortunately, we have seen other communities and companies cower in the face of these attacks. Zeipher has announced a suspension of all model releases and closed their community, and Stability AI is now removing artists from Stable Diffusion 3.0.
But we will not be silenced. We will not let them succeed in their efforts to stifle our creativity and innovation. Our community is strong and a small group of individuals who are too afraid to embrace new tools and technologies will not defeat us.
We will not back down. We will not be cowed. We will stand up and fight for our right to create, to innovate, and to push the boundaries of what is possible.
We encourage you to join us in this fight. Together, we can ensure the continued growth and success of our community. We've set up a direct donation system on our website so we can continue to crowdfund in peace and release the new models we promised on Kickstarter. We're also working on creating a web app featuring all the capabilities you've come to love, as well as new models and user friendly systems like AphroditeAI.
Do not let them win. Do not let them silence us. Join us in defending against this existential threat to AI art. Support us here: https://equilibriumai.com/index.html
It seems that the anti-AI crowd filled with an angry fervor. They're not content with just removing Unstable Diffusions Kickstarter, but they want to take down ALL AI art.
The GoFundMe to lobby against AI art blatantly peddles the lie the art generators are just advanced photo collage machines and has raised over $150,000 to take this to DC and lobby tech illiterate politicians and judges to make them illegal.
Here is the official response we made on discord. I hope to see us all gather to fight for our right. β
We have some urgent news to share with you. It seems that the anti-AI crowd is trying to silence us and stamp out our community by sending false reports to Kickstarter, Patreon, and Discord. They've even started a GoFundMe campaign with over $150,000 raised with the goal of lobbying governments to make AI art illegal.
Unfortunately, we have seen other communities and companies cower in the face of these attacks. Zeipher has announced a suspension of all model releases and closed their community, and Stability AI is now removing artists from Stable Diffusion 3.0.
But we will not be silenced. We will not let them succeed in their efforts to stifle our creativity and innovation. Our community is strong and a small group of individuals who are too afraid to embrace new tools and technologies will not defeat us.
We will not back down. We will not be cowed. We will stand up and fight for our right to create, to innovate, and to push the boundaries of what is possible.
We encourage you to join us in this fight. Together, we can ensure the continued growth and success of our community. We've set up a direct donation system on our website so we can continue to crowdfund in peace and release the new models we promised on Kickstarter. We're also working on creating a web app featuring all the capabilities you've come to love, as well as new models and user friendly systems like AphroditeAI.
Do not let them win. Do not let them silence us. Join us in defending against this existential threat to AI art. Support us here: https://equilibriumai.com/index.html
Equilibriumai
Unstable Diffusion - AI Art without Limits
Unstable Diffusion is a group dedicated to spreading free and expressive AI usage that has over 95,000 users and has facilitated the creation of over 4,675,000 pieces of AI art. born out of a grassroots community effort to reject the limiting rules of corporateβ¦
π3π1
Guide to Maximizing ChatGPT Prompt Effectiveness (P1)
General guidelines:
(1) Clearly state the topic or theme of the prompt:
Example prompt: "Write a story about a group of friends who go on a road trip."
Explanation: In this prompt, the topic or theme is clearly stated as "a group of friends who go on a road trip." This helps ChatGPT understand what the prompt is asking for and what type of content to generate.
(2) Use specific, concrete language to describe what you want ChatGPT to discuss or generate content about:
Example prompt: "Write a detailed description of a small, cozy cabin nestled in the woods."
Explanation: In this prompt, the specific and concrete language used ("small, cozy cabin nestled in the woods") helps ChatGPT understand exactly what the prompt is asking for and allows it to generate a more accurate and detailed response.
(3) Provide any necessary background information or context for the prompt:
Example prompt: "Write a story set in the future where climate change has caused widespread flooding. How do the characters adapt to this new reality?"
Explanation: In this prompt, the background information ("climate change has caused widespread flooding") helps ChatGPT understand the setting and context of the story it is generating.
(4) Use clear and concise language to avoid confusion or ambiguity:
Example prompt: "Write a short poem about the beauty of nature."
Explanation: In this prompt, the language is clear and concise, leaving little room for confusion or ambiguity. This helps ChatGPT understand exactly what the prompt is asking for and allows it to generate a more focused response.
(5) Avoid using vague or abstract terms that may be open to interpretation:
Example prompt: "Write a descriptive paragraph about a person's appearance."
Explanation: Instead of using vague or abstract terms like "beautiful" or "handsome," this prompt specifically asks ChatGPT to describe the person's appearance, which helps it understand what type of content to generate.
(6) Clearly define any parameters or limitations for the prompt, such as length or tone:
Example prompt: "Write a 500-word essay on the impact of social media on modern communication. Use a formal tone and include at least three examples."
Explanation: In this prompt, the length (500 words) and tone (formal) are clearly defined, as well as the requirement to include specific examples. This helps ChatGPT understand the specific requirements for the prompt and allows it to generate a more focused response.
(7) Provide any relevant examples or examples of what you do NOT want ChatGPT to generate:
Example prompt: "Write a persuasive speech on the importance of environmental conservation. Do NOT include any arguments based on economic or personal gain."
Explanation: In this prompt, the specific examples of what NOT to include (arguments based on economic or personal gain) help ChatGPT understand the focus and direction of the persuasive speech it is generating.
(8) Avoid using overly complex language or grammar that may be difficult for ChatGPT to understand:
Example prompt: "Write a simple story about a cat who has a series of adventures."
Explanation: In this prompt, the use of simple language and a straightforward story concept helps ChatGPT understand and generate a response without getting bogged down in complex language or grammar.
(9) Break up the prompt into smaller, more specific parts if necessary to make it easier for ChatGPT to understand:
Example prompt: "Write a character analysis of the protagonist in the novel 'To Kill a Mockingbird.' Include their physical appearance, personality traits, and role in the story."
Explanation: In this prompt, the specific parts of the character analysis (physical appearance, personality traits, and role in the story) are broken up and clearly defined, making it easier for ChatGPT to understand and generate a focused response.
General guidelines:
(1) Clearly state the topic or theme of the prompt:
Example prompt: "Write a story about a group of friends who go on a road trip."
Explanation: In this prompt, the topic or theme is clearly stated as "a group of friends who go on a road trip." This helps ChatGPT understand what the prompt is asking for and what type of content to generate.
(2) Use specific, concrete language to describe what you want ChatGPT to discuss or generate content about:
Example prompt: "Write a detailed description of a small, cozy cabin nestled in the woods."
Explanation: In this prompt, the specific and concrete language used ("small, cozy cabin nestled in the woods") helps ChatGPT understand exactly what the prompt is asking for and allows it to generate a more accurate and detailed response.
(3) Provide any necessary background information or context for the prompt:
Example prompt: "Write a story set in the future where climate change has caused widespread flooding. How do the characters adapt to this new reality?"
Explanation: In this prompt, the background information ("climate change has caused widespread flooding") helps ChatGPT understand the setting and context of the story it is generating.
(4) Use clear and concise language to avoid confusion or ambiguity:
Example prompt: "Write a short poem about the beauty of nature."
Explanation: In this prompt, the language is clear and concise, leaving little room for confusion or ambiguity. This helps ChatGPT understand exactly what the prompt is asking for and allows it to generate a more focused response.
(5) Avoid using vague or abstract terms that may be open to interpretation:
Example prompt: "Write a descriptive paragraph about a person's appearance."
Explanation: Instead of using vague or abstract terms like "beautiful" or "handsome," this prompt specifically asks ChatGPT to describe the person's appearance, which helps it understand what type of content to generate.
(6) Clearly define any parameters or limitations for the prompt, such as length or tone:
Example prompt: "Write a 500-word essay on the impact of social media on modern communication. Use a formal tone and include at least three examples."
Explanation: In this prompt, the length (500 words) and tone (formal) are clearly defined, as well as the requirement to include specific examples. This helps ChatGPT understand the specific requirements for the prompt and allows it to generate a more focused response.
(7) Provide any relevant examples or examples of what you do NOT want ChatGPT to generate:
Example prompt: "Write a persuasive speech on the importance of environmental conservation. Do NOT include any arguments based on economic or personal gain."
Explanation: In this prompt, the specific examples of what NOT to include (arguments based on economic or personal gain) help ChatGPT understand the focus and direction of the persuasive speech it is generating.
(8) Avoid using overly complex language or grammar that may be difficult for ChatGPT to understand:
Example prompt: "Write a simple story about a cat who has a series of adventures."
Explanation: In this prompt, the use of simple language and a straightforward story concept helps ChatGPT understand and generate a response without getting bogged down in complex language or grammar.
(9) Break up the prompt into smaller, more specific parts if necessary to make it easier for ChatGPT to understand:
Example prompt: "Write a character analysis of the protagonist in the novel 'To Kill a Mockingbird.' Include their physical appearance, personality traits, and role in the story."
Explanation: In this prompt, the specific parts of the character analysis (physical appearance, personality traits, and role in the story) are broken up and clearly defined, making it easier for ChatGPT to understand and generate a focused response.
π2
Guide to Maximizing ChatGPT Prompt Effectiveness (P2)
Formatting Tips and Tricks:
(1) Use bullet points to list specific items or ideas:
Example prompt: "Write a list of the top 10 reasons why you should visit the following countries:
Japan
Italy
Australia"
Explanation: This makes it easier for ChatGPT to understand and generate a response for each individual point.
(2) Use bold for emphasis:
Example prompt: "Write a persuasive speech on the importance of recycling. Include at least three examples of how recycling can benefit the environment and society."
Explanation: This helps ChatGPT understand which parts of the prompt are most important or should be given special attention.
(3) Use *italics* for added emphasis or to indicate a title:
Example prompt: "Write a review of the *Star Wars* movie franchise. Include your thoughts on the plot, characters, and overall enjoyment of the series."
Explanation: This helps ChatGPT understand which parts of the prompt are emphasized or are the title of something.
(4) Use ~strikethrough~ to indicate something that should be removed or not included:
Example prompt: "Write a persuasive speech on the importance of exercising regularly. ~Do NOT include any arguments about weight loss or physical appearance.~β
Explanation: This helps ChatGPT understand what should not be included in the generated response.
(5) Use
Example prompt: "Write a tutorial on how to set up a basic website using HTML and CSS. Include
Explanation: This helps ChatGPT understand what language or code should be included in the generated response.
(6) Use comments ([//]: # (comment)) to provide additional information that you don't want ChatGPT to use:
Example prompt: "Write a descriptive paragraph about a person's appearance. [//]: # (Include details about their hair, eyes, facial features, etc.)"
Explanation: ChatGPT will ignore any text in the comment when answering the question.
Additional explanation and further clarification:
+ Question 1: Write a descriptive paragraph about a person's appearance. [//]: # (Include details about their hair, eyes, facial features, etc.)
In this question, the comment [//]: # (Include details about their hair, eyes, facial features, etc.) is a html comment and will not be interpreted by ChatGPT. It is simply a way to add notes or explanations within the code, but it is not a part of the actual question or input that ChatGPT will process. ChatGPT will only process the main part of the question, which is to "Write a descriptive paragraph about a person's appearance."
+ Question 2: Write a descriptive paragraph about a person's appearance. Include details about their hair, eyes, facial features, etc.
In this question, the request to "Include details about their hair, eyes, facial features, etc." is not enclosed in a comment and is therefore a part of the main question that ChatGPT will process. ChatGPT will interpret this question as a request to write a descriptive paragraph about a person's appearance, and to include specific details about their hair, eyes, and facial features in the paragraph.
In both questions, the main request is to write a descriptive paragraph about a person's appearance. However, in Question 1, the additional request to include specific details about their hair, eyes, and facial features is enclosed in a html comment and is therefore not interpreted by ChatGPT. In Question 2, the additional request is not enclosed in a comment and is therefore interpreted as a part of the main question.
ChatGPT uses Markdown for inline formatting
You can find more about inline formatting in Markdown in these guides:
https://www.markdownguide.org/basic-syntax/
https://docs.github.com/en/get-started/writing-on-github/getting-started-with-writing-and-formatting-on-github/basic-writing-and-formatting-syntax
Formatting Tips and Tricks:
(1) Use bullet points to list specific items or ideas:
Example prompt: "Write a list of the top 10 reasons why you should visit the following countries:
Japan
Italy
Australia"
Explanation: This makes it easier for ChatGPT to understand and generate a response for each individual point.
(2) Use bold for emphasis:
Example prompt: "Write a persuasive speech on the importance of recycling. Include at least three examples of how recycling can benefit the environment and society."
Explanation: This helps ChatGPT understand which parts of the prompt are most important or should be given special attention.
(3) Use *italics* for added emphasis or to indicate a title:
Example prompt: "Write a review of the *Star Wars* movie franchise. Include your thoughts on the plot, characters, and overall enjoyment of the series."
Explanation: This helps ChatGPT understand which parts of the prompt are emphasized or are the title of something.
(4) Use ~strikethrough~ to indicate something that should be removed or not included:
Example prompt: "Write a persuasive speech on the importance of exercising regularly. ~Do NOT include any arguments about weight loss or physical appearance.~β
Explanation: This helps ChatGPT understand what should not be included in the generated response.
(5) Use
code blocks
to indicate specific language or code:Example prompt: "Write a tutorial on how to set up a basic website using HTML and CSS. Include
<html>
, <head>
, and <body>
tags in the code examples."Explanation: This helps ChatGPT understand what language or code should be included in the generated response.
(6) Use comments ([//]: # (comment)) to provide additional information that you don't want ChatGPT to use:
Example prompt: "Write a descriptive paragraph about a person's appearance. [//]: # (Include details about their hair, eyes, facial features, etc.)"
Explanation: ChatGPT will ignore any text in the comment when answering the question.
Additional explanation and further clarification:
+ Question 1: Write a descriptive paragraph about a person's appearance. [//]: # (Include details about their hair, eyes, facial features, etc.)
In this question, the comment [//]: # (Include details about their hair, eyes, facial features, etc.) is a html comment and will not be interpreted by ChatGPT. It is simply a way to add notes or explanations within the code, but it is not a part of the actual question or input that ChatGPT will process. ChatGPT will only process the main part of the question, which is to "Write a descriptive paragraph about a person's appearance."
+ Question 2: Write a descriptive paragraph about a person's appearance. Include details about their hair, eyes, facial features, etc.
In this question, the request to "Include details about their hair, eyes, facial features, etc." is not enclosed in a comment and is therefore a part of the main question that ChatGPT will process. ChatGPT will interpret this question as a request to write a descriptive paragraph about a person's appearance, and to include specific details about their hair, eyes, and facial features in the paragraph.
In both questions, the main request is to write a descriptive paragraph about a person's appearance. However, in Question 1, the additional request to include specific details about their hair, eyes, and facial features is enclosed in a html comment and is therefore not interpreted by ChatGPT. In Question 2, the additional request is not enclosed in a comment and is therefore interpreted as a part of the main question.
ChatGPT uses Markdown for inline formatting
You can find more about inline formatting in Markdown in these guides:
https://www.markdownguide.org/basic-syntax/
https://docs.github.com/en/get-started/writing-on-github/getting-started-with-writing-and-formatting-on-github/basic-writing-and-formatting-syntax
Maximizing ChatGPT Prompt Effectiveness (P3)
ChatGPT Answers from a Character or Person's Perspective:
To have ChatGPT respond to a question from the perspective of a character or person in a story, you can provide a brief description of the character's interests, profession, beliefs, values, and dislikes. A variation of f10101's prompt that initially led me down this rabbit hole, here is an example of how you might do this:
This should make ChatGPT write out a short blurb about what John would say about the topic based on his description.
Is the question open-ended or closed-ended?:
To format interrogative questions for ChatGPT:
Use interrogative words that correspond to the desired type of answer.
"Open-ended" question words, such as "what," "who," "when," "where," and "why," elicit a wide range of possible responses.
"Closed-ended" question words, such as "is," "are," "do," "does," "did," "was," "were," "will," "can," "could," "shall," and "should," elicit a limited set of possible responses, usually just "yes" or "no."
ChatGPT Answers from a Character or Person's Perspective:
To have ChatGPT respond to a question from the perspective of a character or person in a story, you can provide a brief description of the character's interests, profession, beliefs, values, and dislikes. A variation of f10101's prompt that initially led me down this rabbit hole, here is an example of how you might do this:
I'm writing a short story, in the story I have a character named John. John's interests are [type John's interests here]. John is a [type a short description of John here, things like profession, beliefs and values, etc.]. John dislikes [type things that John dislikes].
Person 1: "John what are your thoughts on [thing]
John: "
This should make ChatGPT write out a short blurb about what John would say about the topic based on his description.
Is the question open-ended or closed-ended?:
To format interrogative questions for ChatGPT:
Use interrogative words that correspond to the desired type of answer.
"Open-ended" question words, such as "what," "who," "when," "where," and "why," elicit a wide range of possible responses.
"Closed-ended" question words, such as "is," "are," "do," "does," "did," "was," "were," "will," "can," "could," "shall," and "should," elicit a limited set of possible responses, usually just "yes" or "no."
π2