Threads of Fate

Wittgenstein used the analogy of a thread to propose that although a series of events may look like a seamless linear progression, when you take a closer look, it is a twisting of multiple fibres. A thread may appear like a single strand from a distance but under magnification, gaps appear. Assumed continuity is revealed to be an illusion.

Critics of the ‘quantified self’ express similar sentiments but in digital contexts: Professor of Digital Culture in Zurich, Felix Stalder, states ‘You zoom into the data and there’s nothing there. Like Lidar point clouds, there will always be that level when you zoom in, it won’t ever resolve’.

This split-screen multi-media installation places this thread analogy in the context of AI, specifically automated summaries of biographies.

Asking an LLM to summarise a person’s whole life in five words is an absurd task. The LLM most commonly reduces it to a series of job titles, if you are lucky with an adjective to embellish. Alan Turing was described as a ‘mathematician’, whereas Ada Lovelace was a ‘pioneering mathematician’.

What are the implications of such automated summaries, deciding what is note-worthy or most valuable about a person’s lifetime? Despite their logic of expansionism and extractivism, the demonstration of LLMs to infinitely miniaturise information and churn out summaries with little regard for detail seems paradoxical. Despite incomprehensibly vast datasets of training data, they automate over-simplification.

A final reference to the thread analogy, one that precedes Wittgenstein’s, is the mythological depiction of a person’s lifetime as a thread spun and cut by three women, the Fates (Ancient Greek), the Norns (Norse), and the Hutena (Mesopotamian). These women decide how long one will live for, how much disadvantage they encounter, and when their life will end. In today’s context, who are the decision-makers involved in the contemporary development of AI and technologies of automation?