Sublime

by deploying alternative news feed algorithms on in-silico social media platforms, where large language model (LLM) agents that mimic human social media users interact with one another, we can explore and test the impact of these alternative algorithms on macro-level social outcomes, such as conflicts and polarization.[385]