# created ```dataview LIST WITHOUT ID file.link FROM -"windsurf" WHERE file.cday = date(this.file.name) AND !startswith(file.folder, "windsurf") SORT file.cday ASC ``` # modified ```dataview LIST WITHOUT ID file.link FROM -"windsurf" WHERE file.mday = date(this.file.name) AND !startswith(file.folder, "windsurf") SORT file.mday ASC ``` --- # [[diary and thoughts]] #thought #diary --- # Notes on Monika's talk 'functionalist aesthetic theories' do not adequatly account for appreciating [[designed objects]] *as* mass designed products. ### The Raw Material: A Cultural Compost - [[The training]] process begins with the "matter" of the AI environment: the vast corpus of cultural data. - This dataset acts as a layer of cultural 'compost' or organic matter. - It is a rich but unstructured collection of past human creativity. - As Millière notes, this data exists in a high-dimensional space where most configurations are just noise. ### The Process: An Algorithmic Mycelium - [[The training]] algorithm functions like a vast _mycelial network_ (the network of fungi in soil) spreading through this cultural compost. - This algorithmic network decomposes the data. - It breaks down complex cultural artefacts into their fundamental patterns and relationships. - The process of _manifold learning_ is the network discovering the most efficient pathways for this decomposition. - It learns the underlying geometric structure where meaningful information is concentrated. ### The Result: A Fertile [[Latent Space]] - The final, trained model is not a dead archive, but a new, fertile medium: the 'soil' of the _latent space_. - This soil is an emergent substance, qualitatively different from the original compost. - Its structure is defined by the pathways of the learned manifold. - It is an autonomous, [[generative environment]] from which endless novel outputs (new 'growths') can emerge, a process that is irreversible.