Under the leadership of Harvard University and MIT, in partnership with teams from OpenAI and Google DeepMind, the MatrAIx system has been unveiled. This innovative system is capable of generating 8.3 billion AI-driven agents to replicate global human behavior, employing 1,290 dimensions for modeling purposes. The overall alignment between the behaviors of these AI agents and those of real humans is impressively high, reaching 91.5%. Nevertheless, this alignment varies across different scenarios: it climbs to between 92% and 96% in questionnaire and chat scenarios, while it stands at 83% in App manipulation tests. Presently, a carefully curated 'core set of personalities', numbering in the millions, has been made available on the Hugging Face platform. Concurrently, the code for the related 8B personality model has also been open-sourced on GitHub. This ultra-large-scale simulation system is heralded as a novel testing ground for social science and product research endeavors. However, the question of whether AI agents can fully substitute real human samples remains open and necessitates further cross-validation.