Raw LLM Responses

I'm a physicist who works with developing AI models for use in medical imaging. AI doesn't produce anything it hasn't seen before. It doesn't learn, even though we call the process 'learning'. It's not training, even though the data it is fed is called training data. It simply performs a mathematical process known as a convolution, and it does this in a specific pattern that is known as a network. Based on the outcomes of this process it starts assigning weights to a model that is able to then make what we call 'predictions'. These predictions, or output, is not an original work. The same input will always create the same output in these models, making them completely predictable, and they will never produce anything original that it hasn't seen before. So in terms of art, it is simple theft. It doesn't learn anything about the art. It's just a clever way of stealing. Besides the technical aspects of all of this, the argument that most proponents like to use is 'Isn't that what humans already do. You don't need consent to learn'. My argument would be that most humans aren't putting together a trillion dollar industry that is designed to destroy the lives of artists and move the little creative control that they have into the hands of multinational corporations whose sole purpose is to pull blood, sweat and tears from every inhabitant on the planet. Not everyone wants to live in a futuristic dystopia, and the people who do don't seem to know how close they are to having their dreams come true. People are acting as if this is some great stride towards putting the means of production in the hands of the common man. While in reality it is the exact opposite. I've personally been in board rooms with corporate types who understand this all too well, and who are gleefully pushing for law changes to be postponed, and for software solutions to steal as much data as possible. Before the times change once more.