Remixing and blending tokens creates infinite generative options for creative prompting. Users can combing tokens from multiple sources and add additional tokens and parameters after their initial generation to coax the AI into providing exactly what they are looking for (or, alternatively, taking their thinking to entirely new places).

This leads to powerful capabilities, and a user doesn't need to know all that's happening below the scenes to take advantage of them.

  • Image generators allow users to combine different images, or introduce tokens from an image as an input source to your open ended prompt
  • Content editors allow users to direct the model to reproduce the original output with new parameters, like "shorter" or "more conversational"
  • Document parsers summarize the information in a website, document, or attachment and allow users to "chat" with the text
  • Chatbots store a history of information from a conversation that a user can retrieve or ask about even if the conversation has moved on
  • Users can regenerate previous responses using the original prompts, or adding new information

Remixing prompts might be the most human element of Artificial Intelligence today. If you have ever found yourself humming a familiar tune with new lyrics, or finding connections between books that seem to have nothing to do with each other, you know the power of remixing.

From a product and experience perspective, remixing can help resolve dead ends. When someone isn't getting the results they want out of a prompt, giving them the ability to share a reference point, or add new information helps them feel like they are back in control.

Make these features clear and obvious. They let users play with the model. The more you play with something, the better you understand it.

Details and variations

  • Before prompting, allow users to attach references to their prompt to guide the model and inject additional tokens
  • Consider allowing them to blend two references without additional prompt instructions as a discovery mechanism
  • After they have generated their initial response, give them sample actions to easily remix their prompt
  • Let users add or remove references that the AI is remixing against



Novel use cases
Asking a remote bot questions in an open ended field becomes tiresome. This pattern opens up new ways of prompting that many users have not been exposed to yet. What if you could analyze today's newspaper and uncover the most consistent themes - allowing you to ask the bot questions about current events. What if you could upload an old resume and your current LinkedIn, and ask the bot to give you a new way of defining yourself in a turbulent world. This is a space to be creative.

Make room for play
“The next big thing will start out looking like a toy” said Chris Dixon in 2010. With remixing, users don't have to get something perfectly right the first time. Each input leads them closer to where they are going (or, alternatively, let's them explore things they never thought to explore). The nuance here is to know your use case. In professional settings, conformity is critical. Design for that end (add parameters and clues to help shape better prompts). In consumer settings, giving people the tools to play might the best formula for retention and delight.

Potential risks

New inputs, new risk
Each new set of information introduces more things for the model to process, and more risk for chaotic results. Look for ways to make this predictable for users. Specific prompts like "make shorter" give users clues for which parameters you will explicitly introduce. Better yet, give them a way to see the improved prompt. Teach them to become better at writing inputs, so you can focus on making the outputs of your model exceptional.


Pithy suggestions from Grammarly on how to remix your writing
Notion places actions to remix text directly in the page menu
AI text generators like Notion also include inline prompts to remix or blend a generation
Midjourney and other image generators can extract tokens directly from different sources and don't require an additional prompt
Images can be remixed implicitly with prompts to generate similar images, which explores the tapestry to tokens extracted from the image
Midjourney's web interface offers similar options to remix the image into variations in subtle or strong ways
No items found.