[RandomGerm4n]

I think it's immoral to withhold LLMs for “security reasons.” All LLMs are all trained on forum posts from real users, source code from free software, and books written by authors. All of these people should therefore also have access to them. Ideally, the models should all be open weight. But making the model accessible only to certain groups of people is even worse than if it were at least available for purchase as a service for everyone.