I don't buy the argument that open-sourcing would allow citizens to better fight this. Let's be realistic. If the full model is published, for every one perso

I don't buy the argument that open-sourcing would allow citizens to better fight this. Let's be realistic. If the full model is published, for every one person who studies the model to understand how to fight false propaganda, there will be 5,000 who use it to shitposting for lulz. Also, as everyone's pointing out, the cat's out of the bag. Whatever OpenAI does, the same ability will be in the hands of state actors, and large corporations, and soon not-too-large corporations and dedicated individuals, probably in a few years. So basically whatever OpenAI does today doesn't really matter. We might as well think about totally different ways of building online communities when everyone has shitpointing sockpuppet engines powered by AWS/Azure DeepLearningToolKit Trial Version.