The science of simulated users.
Questions, methods, and findings from the MatrAIx community—covering persona grounding, simulated users, interactive environments, evaluation, and identity drift.
PersonalAgentEval
The best personal agent adapts to different users and preferences across domains and contexts. Seven commercial personal agents, 152 tasks, simulated users with full digital lives, and where today's agents still fail.
MatrAIx Core Team
Simulating the World with 8.3 Billion Personas
The MatrAIx technical report in plain language: a persona population, four kinds of studies, 1,010 reusable tasks, and 18,189 trials, with the demo video.
MatrAIx Core Team
AgentPersonaBench: Benchmarking Persona-Driven User Simulation
When language models act as simulated users, do latent personas govern what they do? Evaluating 20 frontier models across 2,460 everyday tasks in surveys, chats, web, and desktop apps.
MatrAIx Core Team
Why Human Simulation Is the Next Step for AI Agents
Why AI needs human simulation: evidence, evaluation gaps, population models, individual digital twins, and a research agenda.
Xiaomin Li, Yuexing Hao, and Core Team
Research
Have research or insights to share? We welcome community contributions.
Submit a post →