Prompt Engineering

What is AI Prompt Engineering for Text-to-Art?

​

Text-to-art prompt engineering is the process of creating prompts that are clear, concise, and informative enough to help large language models (LLMs) generate high-quality images. LLMs are trained on massive datasets of text and code, and they can be used for a variety of tasks, such as generating text, translating languages, and writing different kinds of creative content. However, LLMs are not perfect, and they can sometimes generate images that are inaccurate, misleading, or simply not very good. This is where text-to-art prompt engineering comes in. There are new Prompt startup companies specilizing in this area.

​

How to write good text-to-art prompts:

​

• The prompt should be clear and concise, so that the LLM knows exactly what you want it to do. Avoid using jargon or technical terms that the LLM may not understand.

•  The more specific you are, the better the LLM will be able to generate an image that meets your expectations. For example, if you want the LLM to generate a painting of a cat, be sure to specify the breed of cat, the pose, and the background.

• The prompt should provide the LLM with enough information to generate an image that is accurate and relevant. For example, if you want the LLM to generate a painting of a cat, be sure to provide the LLM with information about cats, such as their anatomy, their behavior, and their environment.

• When you are writing a prompt, be sure to use keywords that are relevant to the image you want the LLM to generate. For example, if you want the LLM to generate a painting of a cat, you might use keywords like "cat," "painting," "breed," "pose," and "background."

•  If you can, provide the LLM with examples of the type of image you want it to generate. For example, if you want the LLM to generate a painting of a cat, you might provide the LLM with examples of paintings of cats that you like.

•  If you want the LLM to avoid generating certain types of images, be sure to use negative language in the prompt. For example, if you don't want the LLM to generate images that are violent, you might say something like "Do not generate images that are violent."

​

Try this Midjourney prompt: "middle-aged man confronting a monster that is a hellish ghoulish monstrosity frightening and fierce."

​

What is AI Prompt Engineering for Writing?

​

AI writing prompt engineering is the process of creating prompts that are clear, concise, and informative enough to help large language models (LLMs) generate high-quality text. LLMs are trained on massive datasets of text and code, and they can be used for a variety of tasks, such as generating text, translating languages, and writing different kinds of creative content. However, LLMs are not perfect, and they can sometimes generate text that is inaccurate, misleading, or simply not very good. This is where AI writing prompt engineering comes in.

​

Tips for AI writing prompts:

​

• The prompt should be clear and concise, so that the LLM knows exactly what you want it to do. Avoid using jargon or technical terms that the LLM may not understand.

• The more specific you are, the better the LLM will be able to generate text that meets your expectations. For example, if you want the LLM to generate a poem, be sure to specify the type of poem (e.g., sonnet, haiku, free verse) and the topic.

• The prompt should provide the LLM with enough information to generate text that is accurate and relevant. For example, if you want the LLM to generate a blog post about a specific topic, be sure to provide the LLM with information about the topic, such as the main points you want to make and the sources you used.

• When you are writing a prompt, be sure to use keywords that are relevant to the topic you want the LLM to generate text about. 

• If you can, provide the LLM with examples of the type of text you want it to generate. For example, if you want the LLM to generate a poem, you might provide the LLM with examples of poems that you like.

• Use negative language. If you want the LLM to avoid generating certain types of text, be sure to use negative language in the prompt. For example, if you don't want the LLM to generate text that is offensive, you might say something like "Do not generate text that is offensive."

​

Try this ChatGPT 4 prompt: "You are an expert novelist in the style of Paddy Chayefsky. Outline a modern-day thriller-suspense spy novel with three story arcs. The overarching arc is the world settings of political tension caused by Russia, China, N. Korea, and Iran banding together to threaten Western democracies. In particular, Russia is threatening NATO, China is threatening Taiwan, N. Korea is threatening S. Korea, and Iran is threatening Israeli. The second story arc is about the personal conflict the main character, Dylan Wyatt, a CIA counterintelligence agent, has with Blake Thomas, an FBI counterintelligence agent, in stopping enemy espionage while they compete for the affections of New York Time reporter Joyce Davenport who covers politics. The third story arc is Dylan’s self-doubt caused by a prior event where he was wounded trying to stop terrorists who successfully exploded a bomb on a train. Dylan’s best friend is Air Force Captain Ryan Hawk, who is deployed to NATO and is training Ukrainian pilots. The outline should cover twenty chapters with interweaving plot and subplot events and include six main characters in the West and four main enemy characters in the rival countries. Dylan has exceptional computer skills and physical combat training.