“Mid Journey is an independent research lab that has released its AI-based, text-to-image tool that can blow your mind.”
prompt: “An Indian temple located in South Korea with architecture design by Gaudi in 4K resolution with Cyberpunk aesthetic.”
The internet has witnessed the rise of AI tools in the last few years, it now seems impossible to ignore them, for designers these AI tools could be extremely important and beneficial to be incorporated in their design workflows.
Mid Journey is research exploring the potential of Artificial Intelligence that is not only taking my Instagram feed by storm but also the internet in general. This post is about the potential of ‘Mid Journey’ and AI generally in the process of Architecture Design and conceptualisation.
“It’s unfounded that AI can replace humans, especially as designers.” – Patrick Hebron
MidJourney granted access to the world through discord, where everyone could sign up for the beta version to explore the possibilities with their crafted algorithm that allowed users to translate information from a descriptive text into an image, and the results are surprisingly good.
The algorithm analyses the written text, understands the framing of the sentence, learns the objective of the user, reads through the descriptive qualities of the target objective, and associates imagery with these attributes in variable proportions to generate four images as options. On the generation of the first set of four images, the user then has options to upscale and render a higher resolution and/or create another set of four variations for a certain option from the earlier iteration. In the process of creating these iterations, the user further trains the algorithm fine-tuning the desired output. Interestingly, the written prompt can be both vague and descriptive. The aspects not mentioned in the prompt allow MidJourney AI to interpret the missing information and generate images regardless.
Mid Journey is a great way to curate a style scape for a project, helping both the designer and the client in visualising the made choices in a semi-realistic way, aiding informed decision-making at the conceptual stage, and at the max setting an aesthetic pallet for the design assignment.
AI’s ability to interpolate three-dimensional interpretation of the written prompt allows the user to generate iterations which possess spatial qualities, hence the ability to do so, elevates its use case for visualising images with three-dimensional attributes. The possibility of the huge variety of iterations allows a high standard of curation at the conceptual stage, expanding the possibilities of imagination roadblocks, that are often experienced in the creative process.