Indeed. For now, for example, if you try to connect to Wikipedia or other Signal or some other communication platforms from Beijing, then the Beijing authorities know that you’re making the connection and they can deny that connection if they don’t feel like it. But when it comes to the proliferation of safe and open AI models, it can run on your phone, it can run on your laptop. There’s no easy way to censor or even to monitor that you’re using this kind of tool. So the most important thing here is to make sure the tools are trustworthy, they’re repeatable, and they do not hallucinate in their social cultural context.