We started to enter an era where design agency in architecture is shared. Even though it has been shared before between design professionals, now AI algorithms start to enter not only in the feasibility and the execution of the project but also in the design process Co-authorship between AI and designers is becoming the new normal. 
     With the recent developments of machine learning, in specific Linguistics-based Diffusion Models, language is increasingly repositioning itself as a central mode of representation and understanding of architecture. These models are trained with massive datasets like the LAION-5B archive of over 5 billion image-text pairs. This gives designers access to a vast pool of sorted data and enables them to work with language and images in a synthetic way. 
     This diploma project uses Linguistics-based Diffusion Models as a design tool and reconfigures the process of making space  by putting atmosphere and materiality at the start of the design process and develops three dimensionality out of it.