2026-03-27 11:03 UTC
Replies (1)
-
@thgie@post.lurk.org 2026-03-27 11:54
That links up nicely with recent trends in sociology/psychology research and their interests in 'silicon personae' and 'synthetic data' (what a terminology, worthy a discussion on its own), e.g. creating survey data through GPT-backed simulated people. Besides audio-visual media and text, there is also a ongoing slopification of academic data-space. That process will most likely also intensify in the coming years. It just fits with the academic grind. Why ask people if you can prompt it? It's a such a defying assumption that everything there ever was is already contained in a LLM. https://www.cambridge.org/core/journals/political-analysis/article/synthetic-replacements-for-human-survey-data-the-perils-of-large-language-models/B92267DC26195C7F36E63EA04A47D2FE https://arxiv.org/html/2512.00461v1 @hito@tldr.nettime.org @bildoperationen@tldr.nettime.org