In computer science, ‘classification’ is when you propagate already existing distinctions (e.g. you already have a database of flower images and you want to recognize new instances of the same kinds of flowers) and ‘clustering’ is when you are making the distinction from scratch entirely based on new data (e.g. you want to distinguish unknown alien species from a new planet). While classification is a technique for propagating existing knowledge, clustering is a form of computing/thinking that is built on the original unknowing. This is why classification is a ‘supervised’ process whereas clustering is an ‘unsupervised’ process .
The supervised classification process propagates its knowledge with a negation (even hatred) of ignorance (dismissing ‘stupidities’ and being ‘smart’ instead) whereas the unsupervised clustering process acknowledges the symptom of its ignorance (which is really shared and social) and engages with it. Classification aims to defeat the current ignorance in order to declare the power of knowledge, whereas clustering aims to process the present ‘nonknowledge’ in order to turn ‘crass ignorance’ into ‘learned-trained ignorance’ (quoted concepts from Lacan, Écrits). Classification seeks to decrypt everything with the master key of knowledge (which is in fact a fetish and an imposture as well as shaky) whereas clustering has to decipher the messy articulations of desire in arbitrary signifiers. The deciphering of clusters is in touch with the unconscious insofar as it ‘knows that it doesn’t know’ many things .
Despite its strengths, clustering is a significantly weaker force than classification because symptoms of ignorance are systematically sacrificed for the fetishes of knowledge. This is commonly known as the patriarchal sacrifice of the feminine embodiments of clusters (called ‘truth’) for the sake of the masculine authorities of classes (called ‘reality’). Classification builds on the ‘eternalism’ of the Master-Signifier whereas clustering engages with the ‘mobilism’ of the objet petit a (quoted concepts from Bard & Söderqvist, Syntheism). Classification alienates the subject by identifying it with fetishes, whereas clustering achieves the subject’s separation by identifying it with symptoms .
This distinction lets me propose an alternative to Marxist class struggle: Cluster struggle .
Marxist class struggle is about propagating the knowledge of the material conditions of the capitalist society in order to promote ‘class consciousness’: The great magnitude of the workers’ exploitation is re-read as a strong evidence of their combined class power instead of their individual weakness and defeat. This new evidence presented by Marxism supports the economic and political embodiments of the trade union and the communist party. Such evidence-based embodiments constitute the feminine aspect of Marxism. Then in the next step, the Communist Party is fetishised as a self-standing authority and regarded as the proof of an independent power that can even function as the communist State of a country. Such proof-based symbolic authorizations constitute the masculine aspect of Marxism .
Fully operative ‘class consciousness’ is only associated with the self-standing masculine aspect of Marxism, and the evidence-based embodiments are subjected to evaluations based on their contributions to the proof-based authorities. But the neoliberal historical conditions have rendered the forming of ‘class consciousness’ impossible by feminizing the proletariat into the precariat . This causes workers to re-focus their attention on the specificities of their particular conditions, which leads them to a regression from ‘class struggle’ which is mainly a masculine domain to ‘cluster struggle’ which is a more feminine domain . This is not a political or moral regression and it’s not an acceptance of total defeat. It’s a regression only in the Freudian sense of getting back to basics, a regression from ‘class consciousness’ to the ‘cluster unconscious’. So prioritizing ‘cluster struggle’ over ‘class struggle’ is not a cause for pride or glory, but it’s a necessary job that has to be done in the present conditions.
 My PhD was about certain clustering algorithms that can guess the number of clusters present in the data. See my science page.
 See “Masculine and Feminine: Truth, Reality and Semblances”, “Desire to Repair and Desire to Decompose”, “Authorization and Embodiment in Fetish and Symptom”, “Syntheism desribes a generic process of authorization”
 Fashionable terms like ‘intersectionality’ attempt to capture this feminine cluster struggle but fall flat and turn into a Superego moralism of beautiful souls currently known as ‘cancel culture’, which is what Mark Fisher once called the ‘Vampire Castle’.