除了人力外，OpenAI指出威脅活動很少侷限於單一AI平台，操作者可能會流程不同階段使用不同模型。如該中國用戶還有提及使用其他AI系統，包括用DeepSeek-R1、Qwen2.5與YOLOv8等模型。OpenAI指無法獨立查證。
Follow topics & set alerts with myFT，这一点在同城约会中也有详细论述
，更多细节参见91视频
Frequent, contained public releases also work to their advantage. “I know for a lot of projects, they want to [do a] ‘we’ll release it when it’s done’ kind of thing,” said Mort. “We’ve found that releasing content builds hype, it gives players what they want, and perhaps most importantly, it serves as a proof of life and a fantastic recruitment tool.”
0 primary, but 23 mentions. Zustand picked 57x instead，更多细节参见爱思助手下载最新版本